WO2021087273A1

WO2021087273A1 - Generation of genome-wide crispr rna libraries using crispr adaptation in bacteria

Info

Publication number: WO2021087273A1
Application number: PCT/US2020/058232
Authority: WO
Inventors: Saeed TAVAZOIE; Wenyan Jiang
Original assignee: The Trustees Of Columbia University In The City Of New York
Priority date: 2019-11-01
Filing date: 2020-10-30
Publication date: 2021-05-06

Abstract

Described herein is a method of producing a CRISPR RNA (crRNA) library using the CRISPR-Cas adaptation machinery in bacteria. In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library. In some embodiments, the bacterial cells constitutively express dCas9, tracrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters.

Description

GENERATION OF GENOME-WIDE CRISPR RNA LIBRARIES USING CRISPR ADAPTATION IN BACTERIA [0001] This application claims the benefit of and priority to U.S. Provisional Patent Application No.62/929,205, filed on November 1, 2019, the content of which is hereby incorporated by reference in its entirety. [0002] All patents, patent applications and publications cited herein are hereby incorporated by reference in their entirety. The disclosures of these publications in their entireties are hereby incorporated by reference into this application. [0003] This patent disclosure contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure as it appears in the U.S. Patent and Trademark Office patent file or records, but otherwise reserves any and all copyright rights. GOVERNMENT SUPPORT [0004] This invention was made with government support under AI077562 awarded by the National Institutes of Health. The government has certain rights to the invention. SEQUENCE LISTING [0005] The instant application contains a Sequence Listing which has been submitted electronically in ASCII format and is hereby incorporated by reference in its entirety. Said ASCII copy, created on October 30, 2020, is named 19240_1169WO1_SL.txt and is 29,240 bytes in size. BACKGROUND OF THE INVENTION [0006] CRISPR-Cas9 systems revolutionized genome editing by targeting specific DNA sequences using complimentary guide RNA. Large guide RNA libraries allow for genome-wide CRISPR screens enable systematic interrogation of gene function. Pooled oligonucleotide synthesis is almost exclusively used for the generation of the gRNA or crRNA libraries; however, custom gRNA or crRNA library generation is expensive, time-consuming, and labor-intensive. SUMMARY OF THE INVENTION [0007] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); a trans-activating crRNA sequence (tracrRNA); CRISPR-associated endonuclease Cas1 (Cas1); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR-associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

; and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding dCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding dCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding dCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; and c) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library.

[0008] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: a hyperactive variant of CRISPR-associated endonuclease Cas9 (hCas9); a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); or a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); /racrRNA; Casl;

Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising nucleotide sequence at least 80% identical to

1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, /racrRNA, Casl, Cas2,

Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding the CRISPR-associated endonuclease, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; c) introducing genomic DNA into the bacterial cells; and d) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library. [0009] In some embodiments, the bacterial cells constitutively express dCas9, tracrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the bacterial cells constitutively express the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding dCas9, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid. [0010] In some embodiments, the method further comprises producing competent bacterial cells before introducing the genomic DNA. In some embodiments, the genomic DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is sheared before introducing the genomic DNA. In some embodiments, the genomic DNA is introduced into the bacterial cells using electroporation. [0011] In some embodiments, in step c) a nucleic acid sequence from the bacterial cell is integrated into the CRISPR array. In some embodiments, a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array. In some embodiments, transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed. In some embodiments, the CRISPR-associated endonuclease is hCas9. In some embodiments, the CRISPR- associated endonuclease is dCas9. In some embodiments, the CRISPR-associated endonuclease is hCas9. In some embodiments, the CRISPR-associated endonuclease is a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9). In some embodiments, the method further comprises isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. [0012] In some embodiments, the method further comprises generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic.

[0013] In some embodiments, the method further comprises isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

[0014] In some embodiments, the method further comprises subcloning the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence into a plasmid.

[0015] In some embodiments, the method further comprises providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a Cas9 lacking endonuclease activity (dCas9) and a /racrRNA, wherein the one or more nucleic acid sequences encoding dCas9 and /racrRNA is under the control of one or more inducible promoters; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array; inducing the expression of the said at least one nucleic acid sequence encoding dCas9 and /racrRNA that is under the control of the one or more inducible promoters; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a crRNA library. In some embodiments, the bacterial cells of the population of bacterial cells and the bacterial cells of the second population of bacterial cells are different species.

[0016] In some embodiments, the method further comprises generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic. In some embodiments, the method further comprises isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

[0017] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: hdCas9; /racrRNA; Casl; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

ID NO: 1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding hdCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding hdCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) isolating a single colony from the population of bacterial cells; c) incubating the single colony of bacterial cells to produce a population of clonal bacterial cells; d) inducing the expression of the said at least one nucleic acid sequence encoding hdCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; e) introducing genomic DNA into the bacterial cells of the population of clonal bacterial cells; and f) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library.

[0018] In some embodiments, the bacterial cells constitutively express hdCas9, /racrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding hdCas9, /racrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the method further comprises producing competent bacterial cells before introducing the genomic DNA. In some embodiments, the genomic

DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is sheared before introducing the genomic DNA. In some embodiments, the genomic DNA is introduced into the bacterial cells using electroporation. In some embodiments, a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array. In some embodiments, the method further comprises isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

[0019] In some embodiments, the method further comprises generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic.

[0020] In some embodiments, the method further comprises isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

[0021] In some embodiments, the method further comprises providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a hyperactive variant of

CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a /racrRNA; Casl;

Cas2; and Csn2; wherein at least one of the one or more nucleic acid sequences encoding the hdCas9, //mrRNA, Casl, Cas2, and Csn2 is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the hdCas9, /racrRNA, Casl, Cas2, and Csn2 that are not under the control of the one or more inducible promoters are constiutively expressed; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence; inducing the expression of the said at least one nucleic acid sequence encoding the hdCas9, /racrRNA, Casl,

Cas2, and Csn2 that is under the control of the one or more inducible promoters; and introducing genomic DNA into the bacterial cells of the second population of bacterial cells; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a dual crRNA library.

[0022] In some embodiments, the method further comprises generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the dual crRNA library with a test compound. In some embodiments, the test compound is an antibiotic. In some embodiments, the method further comprises isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

[0023] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing two populations of bacterial cells wherein the bacterial cells of each population comprise one or more nucleic acid sequences encoding: a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a trans-activating crRNA sequence (//mv RNA); CRISPR- associated endonuclease Casl (Casl); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR- associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to ); and (ii) a canonical

sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein the one or more nucleic acid sequences encoding hdCas9, //mv RNA, and the CRISPR array are constitutively expressed and are present on a first plasmid and wherein the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 are under the control of one or more inducible promoters and are present on a second plasmid, wherein the second plasmid further comprises an origin of replication and a nucleic acid sequence encoding resistance to an antibiotic wherein the origin of replication and nucleic acid sequence encoding resistance to an antibiotic are different between the first and second populations of bacterial cells; b) inducing in both populations of bacterial cells the expression of the said at least one nucleic acid sequence encoding hdCas9, /racrRNA, Casl, Cas2, Csn2, and the CRISPR array that is under the control of one or more inducible promoters; c) mixing the bacterial cells of both populations of bacterial cells to generate a third population of bacterial cells; d) introducing genomic DNA into the bacterial cells of the third population of bacterial cells; and e) incubating the bacterial cells of step d), thereby producing a population of bacterial cells comprising a crRNA library.

[0024] In some embodiments, the bacterial cells of the population of bacterial cells are Gram positive. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus. In some embodiments, the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA). In some embodiments, the bacterial cells of the population of bacterial cells are Escherichia coli. In some embodiments, the bacterial cells of the population of bacterial cells are pathogenic. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the genomic DNA is of Escherichia coli. In some embodiments, bacterial cells of the population of bacterial cells are Staphylococcus aureus and the bacterial cells of the second population of bacterial cells are Escherichia coli.

[0025] In some embodiments, the /racrRNA is a minimal /racrRNA sequence required for targeting. In some embodiments, said minimal /racrRNA sequence comprises 89 nucleotides.

[0026] In certain aspects, the invention provides a population of bacterial cells capable of producing a CRISPR RNA (crRNA) library, said bacterial cells of the population of bacterial cells comprising one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; racrRNA; Casl; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to ; and (ii) a canonical

sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding hdCas9, tracrRNA, Casl, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding hdCas9, racrRNA, Casl, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constitutively expressed.

[0027] In some embodiments, the bacterial cells constitutively express the CRISPR-associated endonuclease, /racrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding the CRISPR- associated endonuclease, /racrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the CRISPR array further comprises a nucleic acid sequence integrated from the bacterial cell.

[0028] In some embodiments, the CRISPR array further comprises a nucleic acid sequence integrated from an exogenous genomic DNA. In some embodiments, transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed. In some embodiments, the CRISPR-associated endonuclease is hCas9. In some embodiments, the CRISPR- associated endonuclease is dCas9. In some embodiments, the CRISPR-associated endonuclease is hdCas9.

[0029] In some embodiments, the exogenous genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the bacterial cells of the population of bacterial cells are Gram-positive. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus. In some embodiments, the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA). In some embodiments, the bacterial cells of the population of bacterial cells are Escherichia coli. In some embodiments, the bacterial cells of the population of bacterial cells are pathogenic. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the exogenous genomic DNA is of Escherichia coli.

[0030] In some embodiments, the /racrRNA is a minimal /racrRNA sequence required for targeting. In some embodiments, said minimal /racrRNA sequence comprises 89 nucleotides.

[0031] In some embodiment, the crRNA library covers at least 80%, 81%, 82%, 83%, 84%,

85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome. In some embodiments, the crRNA library covers at least 80%, 81%, 82%, 83%,

84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome. In some embodiments, the method further comprises isolating said crRNA library.

[0032] In certain aspects, the invention provides a first plasmid comprising a nucleic acid sequence encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; /racrRNA; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to ; and (ii) a

canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; operably linked to a constitutive promoter sequence; and a second plasmid comprising a nucleic acid sequence encoding: Cas1; Cas2; and Csn2; operably linked to an inducible promoter sequence. [0033] In some embodiments the repeat sequence comprises a nucleotide sequence at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identical to

ID NO: 1). In some embodiments the repeat sequence comprises

In some embodiments the repeat sequence consists of

In some embodiments the repeat sequence comprises equivalent sequences to

( Q ) that can be duplicated after each adaptation event. BRIEF DESCRIPTION OF FIGURES [0034] The patent or application file contains at least one drawing originally executed in color. To conform to the requirements for PCT patent applications, many of the figures presented herein are black and white representations of images originally created in color. [0035] FIGS.1A-F show generation of genome-wide gRNA libraries by CRISPR-Cas adaptation. (A) A hyperactive CRISPR-Cas adaptation machinery consists of the 89-nt tracrRNA, hdCas9, Cas1, Cas2 and Csn2. New spacers are integrated into the empty CRISPR array denoted as “R”. (B) Generation of a diverse gRNA library via hyperactive CRISPR-Cas adaptation. (C) Number of reads and location of all 129,856 spacers matching the S. aureus RN4220 genome obtained from deep sequencing of the gRNA library generated in (B). Three gap regions correspond to prophages present in the NCBI reference genome but missing in RN4220. (D) Number of spacers mapped to each of all 2,666 annotated genes in RN4220 versus gene length. The percentage of genes covered by one or more (1+) and three or more (3+) spacers are shown. (E) Number of reads and location of all 462,382 spacers matching the E. coli MG1655 genome obtained from deep sequencing. lacI is abnormally enriched due to an additional presence of the gene in helper plasmid, pCCC. (F) Same as (D) except the genome is MG1655. “gRNA” in the figure refers to crRNA. [0036] FIGS.2A-D show that polarized CRISPR adaptation reveals historical contingency in the acquisition of increasing antibiotic resistance. (A) Schematic showing the sequential construction of a dual-gRNA (2G) library to study the historical contingency of genetic interactions leading to increased antibiotic resistance in bacteria. Gent: gentamicin. (B) Scatter plot showing all gRNA pairs detected in the 2G library. For each gene, the fraction of gRNAs targeting it among the single-gRNA library and number of genes paired with it in the 2G library are shown. (C) Polarized CRISPR adaptation. Spacers are numbered to indicate the order of adaptation. (D) The ratio of the second spacer and the first spacer (Spc2/Spcl) can be used as an indicator for historical contingency between genes along an evolutionary trajectory. Fitness of genes at 9 hours post-selection in gentamicin (1 ug/mL) was measured using Z-score. Using gRNA pairs containing qoxA as an example, Spc2/Spcl negatively correlated with Z-scores, indicating that historical contingency can be revealed by the ratio between the two spacers. Non-essential and essential genes were shown in purple and gray, respectively. “gRNA” in the figure refers to crRNA. “2G” in the figure refers to dual-spacer libraries also referred to as “2S” libraries.

[0037] FIGS. 3A-C show generation of genome-wide gRNA libraries by CRISPR-Cas adaptation. (A) An engineered hyperactive CRISPR-Cas adaptation machinery consists of the 89-nt /racrRNA, hdCas9 (a hyperactive variant of dCas9), Casl, Cas2 and Csn2. New spacers are integrated into the empty CRISPR array denoted as “R”. (B) Generation of a diverse gRNA library via hyperactive CRISPR-Cas adaptation in S. aureus. (C) Iteration of CRISPR adaptation generates a dual-gRNA (2G) library. “gRNA” in the figure refers to crRNA. “2G” in the figure refers to dual spacer libraries also referred to as “2S” libraries.

[0038] FIGS. 4A-F show generation of genome-wide gRNA libraries by CRISPR-Cas adaptation in S. aureus. (A) After CRISPR adaptation described in Figure 13, PCR with enrichment primers are performed and adaptation frequency is estimated by band intensity on 2% agarose gel.

(B) Spacer origins in library-2 (Lib-2). (C) Percentage of spacers with correct PAMs (i.e., NGG). (D) Among all spacers matching the chromosome, 99% were 30- or 31-nt, which are the length of canonical S. pyogenes CRISPR spacers. (E) Sequencing reveals 126,954 spacers matching the S. aureus RN4220 genome. This represents a 93% coverage of all possible targets as the genome harbors 136,928 PAMs. Three gap regions correspond to prophages present in the NCBI reference genome but missing in RN4220. (F) Number of spacers mapped to each of all 2,666 annotated genes in RN4220 versus gene length. The percentage of genes covered by one or more (1+) and three or more (3+) spacers are shown. “gRNA” in the figure refers to crRNA.

[0039] FIGS. 5A-F show Generation of genome-wide gRNA libraries by CRISPR-Cas adaptation in for other bacteria. (A) The system is the same as in Figure 13 except hdcas9 is replaced by hcas9. (B) A gRNA library targeting the genome of interest is generated in S. aureus and sub cloned into other bacteria. (C) Spacer origins of two gRNA libraries targeting E. coli MG1655 made in S. aureus. (D) Sequencing reveals 455,011 spacers matching the E. coli MG1655 genome. (E)

Number of spacers mapped to each of all 4,498 annotated genes in MG1655 versus gene length. All genes are covered by three or more (3+) spacers. (F) Alternatively, CRISPR adaptation machinery could be expressed in E. coli MG1655. Moderate adaptation events are observed when competent cells are electroporated with either genomic DNA (gDNA) or water (H2O) control. Adaptation is not detected in overnight culture (O/N) expressing the adaptation machinery. “gRNA” in the figure refers to crRNA.

[0040] FIGS. 6A-E shows genome-wide gRNA libraries identify known and novel pathways of aminoglycoside sensitivity. (A) S. aureus gRNA libraries are treated with sub-lethal concentration of gentamicin. (B) Distribution of Z-scores of genes after gentamicin (1 pg/mL) treatment. (C) Z-scores of all individual gRNAs targeting representative non-essential genes from triplicates (lib-1 through lib-3). Purple and orange dotted lines indicate the mean Z-scores of all antisense and sense gRNAs, respectively. (D) Same as (C) except the genes are essential. (E) Off-target effect is examined. “gRNA” in the figure refers to crRNA. Figure 6E discloses SEQ ID NOS 3, 5, 7, 4, 6, 8 and 9, respectively, in order of appearance.

[0041] FIGS. 7A-D show additional hits and validation of gentamicin sensitivity. (A) S. aureus gRNA libraries are treated with gentamicin (1 pg/mL) or grown in plain media . Scatter plot showing all genes with positive Z-scores (mean-Z) and P < 0.05 (Mann-Whitney U). (B) Same as (A) except dots represent all genes with negative Z-scores (mean-Z). (C) Potentiation of gentamicin by TOFA, an FAS inhibitors as shown by growth (OD600 measured at 18 hr), [gentamicin]: 0, 0.125, 0.25 and 0.5 pg/mL; [TOFA]: 0, 0.39, 0.78 and 1.56 pg/mL. (D) MICs of gentamicin of S. aureus cells harboring top enriched and depleted spacers. Triplicates were performed.

[0042] FIGS. 8A-E show that “One-vs-all” libraries identify genetic interactions that strengthen antibiotic resistance. (A) Fitness landscape showing microbial adaptation to extreme environment

(e.g., gentamicin) through sequential accumulation of mutations. (B) Cells harboring a universal qoxA-targeting spacer were subjected to CRISPR adaptation, generating a comprehensive “qoxA-vs- all” library. Distribution of log2 of fold-change of genes after treatment with gentamicin (4.0 pg/mL) is shown. (C) 38 dual-gRNAs targeting qoxA and gene X that are enriched at least 27-fold after selection in (B). log2FC(X) and log2FC(qoxA + X) are measured from either the single-gRNA or

“qoxA-vs-all” libraries subjected to gentamicin (4 pg/mL). Epistasis (e) between qoxA and gene X is estimated using log2FCs as a proxy for fitness. (D) Genetic interactions and epistasis. (E)

Correlation between the log2FC of gRNAs (either alone or with qoxA) underlined in (C) and their fitness measured by pairwise competition. “gRNA” in the figure refers to crRNA.

[0043] FIGS. 9A-C show construction of a dual-gRNA (2G) library by iterating CRISPR adaptation. (A) Schematic showing the sequential construction of a dual-gRNA (2G) library. (B)

Scatter plot showing all gRNA pairs detected in the 2G library. For each gene, the fraction of gRNAs targeting it among the single-gRNA library and number of genes paired with it in the 2G library are shown. (C) 2G library recapitulates hits identified in “one-vs-all” libraries. Venn diagrams shows that top enriched gRNA pairs in the “qoxA-vs-all” library and “ndh-vs-all” library are recapitulated by the 2G library. “gRNA” in the figure refers to crRNA. “2G” in the figure refers to dual-spacer libraries also referred to as “2S” libraries [0044] FIGS.10A-E show that polarized CRISPR adaptation reveals historical contingency in the acquisition of increasing antibiotic resistance. (A) 38 dual-gRNAs targeting qoxA and gene X that are enriched at least 27-fold after selection in Figure 8B. log2FC(X) is measured in low [gentamicin] (1 µg/mL) after 9 hours. log2FC(X) and log2FC(qoxA + X) are also measured from either the single-gRNA or “qoxA-vs-all” libraries subjected to high [gentamicin] (4 µg/mL). (B) Fitness of S. aureus RN4220 cells harboring single- and dual-gRNAs targeting qoxA and/or cydA in gentamicin (1.0 µg/mL) measured by pairwise competition. (C) A common competitor strain harboring a qoxA-targeting spacer but different resistance marker is used. (D) Polarized CRISPR adaptation. Spacers are numbered to indicate the order of adaptation. (E) Spc2/Spc1 ratios of gRNA pairs containing qoxA in the 2G library inversely correlated with Z-scores (gent1, 9 hr). “2G” in the figure refers to dual-spacer libraries also referred to as “2S” libraries [0045] FIG.11 shows a schematic of the bacterial fatty acid synthesis. [0046] FIG.12 shows that 5-(tetradecyloxy)-2-furoic acid (TOFA), a FAS inhibitor potentiated the effect of gentamicin as shown by bacterial growth (OD₆₀₀ measured at 18 h) and that Cerulenin ^{potentiates the effect of gentamicin as shown by bacterial growth (OD} ₆₀₀ ^{measured at 18 h).} [0047] FIGS.13A-E show generation of genome-wide crRNA libraries by CRISPR-Cas adaptation. (A) A hyperactive CRISPR-Cas adaptation machinery consists of the 89-nt tracrRNA, hdCas9, Cas1, Cas2 and Csn2. New spacers are integrated into the empty CRISPR array denoted as “R”. To generate a diverse crRNA library, sheared genomic DNA is electroporated into competent S. aureus cells harboring the adaptation machinery. (B) A crRNA library was generated by electroporating S. aureus RN4220 genomic DNA as described in (A). Number of reads and location of all 129,856 sequenced spacers matching the genome are shown. Three gap regions correspond to prophages present in the NCBI reference genome but missing in RN4220. (C) Number of spacers mapped to each of all 2,666 annotated genes in S. aureus RN4220 versus gene length. (D) A crRNA library was generated by electroporating E. coli MG1655 genomic DNA as described in Figure 22A. Number of reads and location of all 462,382 sequenced spacers matching the genome are shown. lacI was preferentially enriched due to an additional presence of the gene in helper plasmid, pCCC (Figure 22A). (E) Number of spacers mapped to each of all 4,498 annotated genes in E. coli MG1655 versus gene length. [0048] FIGS.14A-I show that genome-wide crRNA libraries identify known and novel pathways of aminoglycoside sensitivity. (A) S. aureus RN4220 cells with crRNA libraries generated by CRISPR adaptation were treated with sub-lethal concentration of gentamicin (1 µg/mL). (B) Distribution of Z-scores (P₉₅-Z) for each gene after gentamicin (1 µg/mL) treatment. Novel genes and pathways are underlined. (C) Z-scores of all individual crRNAs targeting qoxA from triplicates (lib-1 through lib-3) after treatment with gentamicin for 18 h. Purple and orange dotted lines indicate the mean Z-scores of all antisense and sense crRNAs, respectively. (D) Same as (C) except the gene is mvaD. (E) Scatter plot of all genes with positive Z-scores (Mean-Z) and P < 0.05 (Mann-Whitney U), measured at 4.5-hour post selection in gentamicin (1 µg/mL). Color codes are the same as those in (B). (F) Same as (E) except dots represent all genes with negative Z-scores (Mean-Z) and P < 0.05 (Mann-Whitney U). Genes involved in the fatty acid synthesis (FAS) pathway are underlined. (G) MICs (measured in triplicates) of gentamicin for S. aureus cells harboring representative top enriched and depleted spacers. (H) Vanadyl sulfate (VS), a mevalonate inhibitor antagonized the effect of gentamicin as shown by bacterial growth (OD₆₀₀ measured at 10 h). [gentamicin]: 0, 0.5, 1 and 2 µg/mL; [VS]: 0, 125, 250 and 500 µg/mL. See also Figure 34. (I) 5-(tetradecyloxy)-2-furoic acid (TOFA), a FAS inhibitor potentiated the effect of gentamicin as shown by bacterial growth (OD₆₀₀ measured at 18 h). [gentamicin]: 0, 0.125, 0.25 and 0.5 µg/mL; [TOFA]: 0, 0.39, 0.78 and 1.56 µg/mL. See also Figure 32. [0049] FIGS.15A-B show that “One-vs-all” libraries identify genetic interactions that strengthen antibiotic resistance. (A) S. aureus RN4220 cells harboring a universal qoxA-targeting spacer were subjected to CRISPR adaptation, generating a comprehensive “qoxA-vs-all” dual- perturbation library. The library was treated with high concentration of gentamicin (4.0 µg/mL) and the distribution of the 95^th percentile of log₂ of fold-change (P₉₅-log₂FC) for each gene after treatment is shown. (B) The top 45 most enriched dual-spacers targeting qoxA and gene X identified in (A) are shown. For every gene X, its log₂FCs measured from single spacer libraries subjected to 1 µg/mL or 4 µg/mL gentamicin for 18 h, as well as its log₂FC from “qoxA-vs-all” library subjected to 4 µg/mL gentamicin for 18 h are shown. Epistasis (ɛ) between qoxA and gene X in 4 µg/mL gentamicin was estimated using log₂FC as a proxy for fitness. [0050] FIGS.16A-E show that dual-spacer (2S) libraries capture pairwise genetic perturbations that strengthen antibiotic resistance. (A) Schematic showing the sequential construction of a dual- spacer (2S) library. (B) Scatter plot showing all spacer pairs detected in 2S library at a gene-gene level. For each gene, the fraction of spacers targeting it among the single-spacer (1S) library (x-axis) and number of genes connected to it in 2S library (y-axis) are shown. (C) Related to (B), histogram showing the distribution of number of connections each gene made in the 2S library. (D) Comparison of the top enriched hits (i.e., spacer pairs) containing qoxA in 2S library and “qoxA-vs-all” library generated in Figure 15A. Histograms showing the distributions of the log₂FC values (P₉₅-log₂FC) for each gene after gentamicin (4 µg/mL) treatment for both libraries. Venn diagram showing the top enriched hits identified from both libraries. (E) Distribution of the number of unique crRNAs targeting genes in the pre-selected 2S library. These genes were either identified as top hits in both libraries (red, 23 genes), or in “qoxA-vs- all” library alone (purple, 22 genes). Color codes match the Venn diagram in (D). A Mann-Whitney U test was performed between the two groups of genes and the P-value is shown. [0051] FIGS.17A-B show that polarized CRISPR adaptation reveals historical contingency in the acquisition of increasing antibiotic resistance. (A) Schematic showing polarized CRISPR adaptation in which new spacers are always acquired at the leader-proximal end of the array. (B) Spc2/Spc1 ratios of spacer pairs containing qoxA in 2S library inversely correlated with Z-scores of the non-qoxA gene measured in 1 µg/mL gentamicin at 9 h (P₉₅-Z, gent 1, 9 h), the point at which the second spacer was adapted. Distribution of P₉₅-Zs is also shown. [0052] FIG.18 shows genome-wide CRISPR libraries of existing and current studies and their gene coverage. Compilation of CRISPR libraries targeting a variety of prokaryotic and eukaryotic genomes rom previous (Bassett et al., 2015; Cui et al., 2018; Gilbert et al., 2014; Lee et al., 2019; Sanson et al., 2018; Sidik et al., 2016) and current studies. The number of total gRNAs (or crRNAs), total genes and average number of gRNAs (or crRNAs) per gene are shown. [0053] FIGS.19A-E show generation of crRNA libraries with hdCas9 in S. aureus. (A) Schematic of generating genome-wide crRNA libraries to be directly used in S. aureus. In their native context, while Cas1, Cas2 and Csn2 are involved in CRISPR adaptation, Cas9 and tracrRNA engage in both the adaptation and targeting process. Plasmid pTHR (aka. pWJ402) contains tracr, hdcas9 (hyper dead Cas9, D10A, I473F, H840A), an empty CRISPR array (R) and a chloramphenicol resistance marker (cat). Plasmid pCCC (aka. pWJ418) contains cas1, cas2 and csn2 under the IPTG-inducible promoter, pSpac. pCCC also contains a tetracycline resistance marker (tet). Overnight cells harboring these two plasmids were re-grown and CRISPR adaptation was induced, followed by electroporation of genomic DNA (gDNA). See “Generation of single-spacer (1S) libraries by CRISPR-Cas adaptation in S. aureus with hdCas9” in Methods for the detailed protocol of generating the rRNA libraries. (B) Left: Spacer adaptation detected in 2% agarose gel. Plasmids extracted from cells post CRISPR adaptation were subjected to PCR with enrichment primers. Upper (-163 bp) and lower (97 bp) bands corresponded to the adapted and non-adapted CRISPR arrays, respectively. crRNA libraries were made in RN4220 and wild-type S. aureus strains (Newman, TB4 and MW2). Right: Reference agarose gel showing PCR performed on mock samples prepared with known ratios of adapted and non-adapted CRISPR arrays. (C) Deep sequencing revealed spacer origin of six crRNA libraries generated in S. aureus RN4220. Spacers were derived from either the host chromosome or two helper plasmids. More chromosomal pacers were generated when electrocompetent cells were made at room temperature (Lib-2, Lib-3, Lib-5 and Lib-6), as opposed to at 4°C (Lib-1 and Lib-4). (D) Among all spacers matching the chromosome, 87%-90% had the correct NGG PAMs. Library-2 Lib-2) is shown as a representative. (E) Among all spacers matching the chromosome, 99% were 30- or 31-nt, which are the length of canonical S. pyogenes CRISPR spacers. Library-2 (Lib-2) is shown as a representative.

[0054] FIGS. 20A-C show that the majority of spacers were derived from helper plasmids when CRISPR adaptation was constitutively expressed. (A) S. aureus RN4220 cells harboring a constitutively expressed hyperactive CRISPR adaptation machinery. Plasmid pTHR (aka. pWJ402) contains tracr, hd cas9 (hyper dead Cas9, D10A, I473F, H840A), an empty CRISPR array (R) and a chloramphenicol resistance marker (cat). Plasmid pCCC aka. pAVl 12B) contains casl, cas2 and csn2 under a strong constitutive pTet promoter. pCCC also contains a tetracycline resistance marker (; tet ). (B) Overnight cells harboring these two plasmids in (A) were re-grown, followed by electroporation of S. aureus genomic DNA (gDNA). Spacer adaptation was detected in both overnight (O/N) cells and ells electroporated with gDNA even using non-enrichment primers (W1307 and L401). The percentage of cells with an adapted spacer is shown. (C) Deep sequencing revealed spacer origin of adapted overnight (O/N) cells and cells electroporated with gDNA.

[0055] FIGS. 21A-B show that internal DNA was the preferred substrate for CRISPR adaptation. (A) S. aureus RN4220 cells harboring a hyperactive CRISPR adaptation machinery were electroporated with A. coli MG1655 genomic DNA. Deep sequencing revealed the spacer origin of adapted cells. (B) Similar to Figure 13B, number of reads and location of all spacers matching the S. aureus RN4220 genome are shown.

[0056] FIGS. 22A-B show generation of crRNA libraries targeting organisms of interest with hCas9 in S. aureus. (A) Schematic of generating genome-wide crRNA libraries to be sub-cloned into other organisms of interest. In their native context, while Casl, Cas2 and Csn2 are involved in

CRISPR adaptation, Cas9 and IracrRN A engage in both the adaptation and targeting process.

Plasmid pTHR (aka. pWJ411) contains tracr, h cas9 (hyper Cas9, 1473F), an empty CRISPR array

(R) and a chloramphenicol resistance marker (cat). Plasmid pCCC contains casl, cas2 and csn2 under the IPTG-inducible promoter, pSpac. Two different pCCCs were constructed. pWJ418 carried a pT181 origin and an tetracycline resistance marker (tet); pWJ420 carried a pE194 origin and an erythromycin resistance marker (erm). Overnight cells harboring these pTHR and pCCC were re- grown and CRISPR adaptation was induced, followed by electroporation of genomic DNA (gDNA) of interest. See Generation of single-spacer (1S) libraries by CRISPR-Cas adaptation in S. aureus with hCas9” in Methods for the detailed protocol of generating and sub-cloning the crRNA libraries. (B) Deep sequencing revealed spacer origin of two E. coli crRNA libraries generated in S. aureus RN4220 before sub-cloning. The pCCC had either a pT181 origin (pWJ418) or a pE194 origin pWJ420). [0057] FIG.23 shows generation of an E. coli crRNA library with dCas9 in E. coli. E. coli MG1655 cells harboring the S. pyogenes CRISPR-Cas adaptation machinery were electroporated with its own genomic DNA. Plasmids extracted from cells post CRISPR adaptation were subjected to PCR with enrichment primers. Upper and lower bands on a 2% agarose gel corresponded to the adapted and non-adapted CRISPR arrays, respectively. Competent cells were prepared by re- inoculating overnight cells and growing to exponential phase (Methods). Moderate adaptation events were observed when competent cells were electroporated with either genomic DNA (gDNA) or water (H₂O) control. Adaptation was not detected in overnight culture (O/N). [0058] FIGS.24A-C show long-period treatment of crRNA libraries in gentamicin. (A) S. aureus RN4220 cells with crRNA libraries generated by CRISPR adaptation were treated with gentamicin (1 µg/mL) for 18 hours (Figure 14A). Distribution of Z-scores (Mean-Z) for each gene are shown (as opposed to P₉₅-Z in Figure 14B). (B) Scatter plot showing the correlation between the ^{means of P} ₉₅ ^{-Z and High} _2/3 ^{-Z of genes after 1S libraries were treated with gentamicin (1 µg/mL)} for 18 hours in triplicates. (C) Scatter plot showing the correlation between the means of P₉₅-Z and P₉₅-log₂FC of enriched genes after 1S libraries were treated with gentamicin (1 µg/mL) for 18 hours in triplicates. [0059] FIGS.25A-D show Z-scores of all individual crRNAs targeting representative non- essential genes enriched in gentamicin (1.0 µg/mL) after 18-hour and 4.5-hour treatment. (A) Z- scores of all individual crRNAs targeting the qox operon (qoxABCD) 18-hour and 4.5-hour post- election in gentamicin (1.0 µg/mL).18-hour experiments were done in triplicates (lib-1 through lib- 3).4.5-hour experiments were either treated with gentamicin or grown in plain media. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. For 18-hour experiments, horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs, respectively. For 4.5-hour experiments, horizontal purple and orange solid lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample treated with gentamicin, and horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs from the sample grown in plain media. (B-D) Same as (A) except the genes are ctaABM , mnhABCDEFG or ndh.

[0060] FIGS. 26A-B show genes and pathways in S. aureus enriched after gentamicin treatment. (A) Schematic showing genes and pathways in which crRNAs were highly enriched after exposure to gentamicin. Pathways with black edges were enriched in low [gentamicin] (i.e., 1 pg/mL) and sometimes high [gentamicin] (i.e., 4 pg/mL). Pathways with red edges were only enriched in high gentamicin] (i.e., 4 pg/mL) when a second gene was co-repressed (i.e., identified in 2S or individual one-vs-all” libraries). (B) Other genes and pathways in which crRNAs were highly enriched. Similar to (A), red edges indicate enrichment only in high [gentamicin] when a second gene was co repressed. Ovals indicate genes/pathways not previously implicated in aminoglycoside resistance or not annotated.

[0061] FIGS. 27A-B show short-period treatment of crRNA libraries in gentamicin. (A) A S. aureus RN4220 crRNA library generated by CRISPR adaptation was treated with gentamicin 1.0 pg/mL) or grown in plain medium for 4.5 hours. In order to detect genes that were significantly depleted in gentamicin but not plain medium, the Z-score of individual crRNAs targeting each gene measured in gentamicin and plain medium were subjected to a Mann-Whitney U test (Methods). (B) Distribution of Z-scores (Mean-Z) for each gene 4.5-hour post-selection in gentamicin (1 pg/mL). Many genes enriched after 4.5-hour selection were also seen after 18 hours (Figure 14B), with the exception of genes encoding the large and small ribosomal subunits (e.g, rplP and rplC ), rRNAs ( e.g. , R0001 - R0005 ) and the iron-sulfur cluster (e.g, sufBC VAV).

[0062] FIGS. 28A-C show Z-scores of all individual crRNAs targeting representative essential genes enriched in gentamicin (1.0 pg/mL) after 18-hour and 4.5-hour treatment. (A) Z-scores of all individual crRNAs targeting the mva operon (invaK IDK2) 18-hour and 4.5-hour post-selection in gentamicin (1.0 pg/mL). 18-hour experiments were done in triplicates (lib-1 through ib-3). 4.5-hour experiments were either treated with gentamicin or grown in plain media. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. For 18-hour experiments, horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs, respectively. For 4.5-hour experiments, horizontal purple and orange solid lines indicate the mean Z- scores of all antisense and sense crRNAs from the sample treated with gentamicin, and horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs from the sample grown in plain media. (B) Same as (A) except the genes are a part of the operon encoding ribosomal genes (rplP - rplC). (C) Z-scores of all individual crRNAs targeting rplB after 18-hour post-selection in gentamicin, as a blow-up of (B). Black arrows showing antisense crRNAs that could not be cloned. [0063] FIGS.29A-C show computational simulations of low-diversity libraries. (A) 100 simulations of low-diversity crRNA libraries. In each simulation, 10 random crRNAs per gene rom the comprehensive crRNA library generated by CRISPR adaptation were selected and P₉₅-Zs were calculated. Heat map showing the mean of P₉₅-Zs (from triplicates, 18-hour treatment) of the top 30 enriched hits identified experimentally from the comprehensive library, as well as those from the 100 simulated low-diversity libraries. Black edge indicates P₉₅-Z is greater than 1.96 (i.e., P < 0.05). Enriched hits that were missing in more than 50% of the simulations are labeled. (B) Similar to (A) except heat map showing the -log(P) value (Mann-Whitney U test) of the 11 genes Figure 14F) that were significantly depleted in gentamicin (1.0 µg/mL) after 4.5-hour treatment. All these hits were missing in more than 50% of the simulations. (C) Number of crRNAs per gene increases the statistical power of discovery. Scatter plot showing the correlation between the number of unique crRNAs targeting each of the 11 gene that were significantly depleted in gentamicin (Figure 14F) and their -log(P) value of the Mann-Whitney U test (Methods). [0064] FIGS.30A-G show Z-scores of all individual crRNAs targeting representative genes depleted in gentamicin (1.0 µg/mL) after 4.5-hour treatment. (A) Z-scores of all individual crRNAs targeting accD treated with gentamicin (1.0 µg/mL) or grown in plain media for 4.5 hours. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Horizontal purple and orange solid lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample treated with gentamicin, and horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample grown in plain media. The number of crRNAs targeting the gene (N) and the P value of the Mann-Whitney U test (P) are also shown. (B-G) Same as (A) except the genes are accA, accC, fabZ, atpB, atpF or ftzH, respectively. [0065] FIGS.31A-E show Z-scores of all individual crRNAs targeting representative genes containing non-neutral off-target sites. (A) Z-scores of all individual crRNAs targeting SAOUHSC_0125718-hour post-selection in gentamicin (1.0 µg/mL) (triplicates). Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense crRNAs, respectively. The crRNA with abnormally high Z-score is circled and the base-pairing between the spacer and the non-neutral off-target site is shown. A gene (e.g., qoxA, qoxC and mnhA) is defined as non-neutral if repression of it by CRISPRi increases cell’s fitness in gentamicin (Figure 14B). (B-E) Same as (A) except the genes are SAOUHSC_00628, SAOUHSC_03037, SAOUHSC_03016 or SAOUHSC_00266. Figure 31A discloses SEQ ID NOS 3 and 9, Figure 31B discloses SEQ ID NOS 5 and 10, Figure 31C discloses SEQ ID NOS 7-9, Figure 31D discloses SEQ ID NOS 11-12 and Figure 31E discloses SEQ ID NOS 13-14, all respectively, in order of appearance. [0066] FIGS.32A-C show that inhibition of fatty acid synthesis potentiated the effect of gentamicin. (A) Schematic of the bacterial fatty acid synthesis (FAS) pathway borrowed from a review article Zhang et al., 2006) with the following modifications. Boxes indicate genes in which crRNAs were significantly depleted after exposure to gentamicin (1.0 µg/mL) for 4.5 hours. TOFA (5-tetradecyloxy)-2-furoic acid) and cerulenin inhibits the acetyl-CoA carboxylase complex AccABCD and FabB/F, respectively. (B) Growth of S. aureus RN4220 cells in media containing various concentrations of gentamicin and TOFA. Units are in µg/mL. See also Figure 14I. (C) ^{Cerulenin potentiates the effect of gentamicin as shown by bacterial growth (OD} ₆₀₀ ^{measured at 18} h). [gentamicin]: 0, 0.125, 0.25 and 0.5 µg/mL; [cerulenin]: 0, 3.125, 6.25 and 12.5 µg/mL. [0067] FIGS.33A-D show growth curves and MICs. (A) Growth of S. aureus RN4220 cells harboring selected top enriched spacers shown in Figure 14B in plain medium or gentamicin (1 µg/mL). (B) Growth of S. aureus RN4220 cells harboring spacers that target the 16S rRNA (R0001- R0005) and two structural components of the ribosome, rplB and rplC, in plain or gentamicin (1.0 µg/mL). All five 16S rRNAs (R0001-R0005) share extensive sequence homology and therefore the spacers designed match all of them. The MICs (triplicates) are also shown. (C) Growth of S. aureus RN4220 cells harboring spacers targeting an intergenic region (genomic locations shown in parenthesis) in plain medium or gentamicin (1.0 µg/mL). The MICs (triplicates) are also shown. (D) Growth of S. aureus RN4220 cells harboring selected top depleted spacers shown in Figure 14F in plain medium or gentamicin (0.5 µg/mL). For all panels, representative growth curves of at least three independent assays are shown. [0068] FIGS.34A-B show inhibition of mevalonate pathway antagonized gentamicin. (A) Schematic of the mevalonate pathway. Underlined genes had crRNAs that were significantly enriched after exposure to low (1.0 µg/mL) and high (4.0 µg/mL) concentrations of gentamicin. Boxed genes had crRNAs that were significantly enriched only after exposure to high concentration of gentamicin when a second gene was co-repressed (i.e., identified in 2S or individual “one-vs-all” libraries). Vanadyl sulfate (VS) inhibits the mevalonate kinase (Gharehbeglou et al., 2015). (B) Growth of S. aureus RN4220 cells in media containing gentamicin (1.0 µg/mL) and various concentrations of VS. Units are in µg/mL. See also Figure 14H. [0069] FIGS.35A-E show that Log₂FCs of all individual crRNAs targeting representative essential genes enriched in “qoxA-vs-all” libraries in gentamicin (4.0 µg/mL) after 18-hour treatment. (A) Log₂FCs of all individual crRNAs targeting the mva operon (mvaK1DK2) after “qoxA-vs-all” libraries (triplicates) were subjected to gentamicin (4.0 µg/mL) for 18-hour. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Horizontal purple and orange dotted lines indicate the mean log₂FCs of all antisense and sense crRNAs, respectively. (B-E) Same as (A) except the genes are a part of the operon encoding ribosomal genes (rplP – rplC), atoB, mvaS and SAOUHSC_00760, respectively. [0070] FIG.36 shows MICs of gentamicin for S. aureus RN4220 cells harboring single- and dual-spacers. The MICs (triplicates) of gentamicin for S. aureus RN4220 cells harboring single- and dual-spacers are shown. [0071] FIGS.37A-G show fitness measured by pairwise competition. (A-E) Fitness of S. aureus RN4220 cells harboring single- and dual-spacers in gentamicin (4.0 µg/mL) measured by pairwise competition. (F) Correlation between the log₂FC of crRNAs (either alone or with qoxA) and their fitness measured by pairwise competition in (A-E). (G) Same as (E) except fitness was measured in gentamicin (1.0 µg/mL). [0072] FIG.38 shows abundance of crRNAs (gene level) in 2S library after selection in high [gent]. Pie chart showing the average abundance of spacer (gene level) of all pairs among 2S library after election in high concentration of gentamicin (4.0 µg/mL) calculated from triplicates. Stringent filter was applied such that only genes that were targeted by an average of 3.5 unique pacers across triplicates were considered, in order to avoid spacers with off-target potential. A total of 183 genes remained after this filter. [0073] FIGS.39A-C show that top enriched spacer pairs containing qoxA in 2S library were corroborated by qoxA-vs-all” library. (A) The top 26 most enriched spacer pairs containing qoxA in 2S library from Figure 16D and their log₂FC values. The log₂FCs measured from the “qoxA-vs-all” library are also shown. The 23 underlined hits were enriched in both 2S and “qoxA-vs-all” libraries. (B) Distribution of the number of unique crRNAs targeting the 23 underlined hits in (A) in their respective pre-selected libraries. (C) Distribution of the number of unique crRNAs targeting the 23 underlined hits in (A) in their respective post-selected libraries. [0074] FIGS.40A-B show that top enriched spacer pairs containing ndh in 2S library were corroborated by ndh-vs-all” library. (A) Comparison of the top enriched hits (i.e., spacer pairs) containing ndh in 2S library and “ndh-vs-all” library. Histograms showing the distributions of the log₂FC values (P₉₅-log₂FC) of genes after gentamicin (4 µg/mL) treatment for both libraries. Venn diagram showing the top enriched hits identified from both libraries. (B) Distribution of the number of unique crRNAs targeting genes in the pre-selected 2S library. These genes were either identified as top hits in both libraries (red), or in “ndh-vs-all” library alone (purple). Color codes match the Venn diagram in (A). [0075] FIGS.41A-B show top enriched spacer pairs containing qoxB in 2S library were corroborated by qoxA-vs-all” library. (A) Comparison of the top enriched hits (i.e., spacer pairs) containing qoxB in 2S library and “qoxA- s-all” library. Histograms showing the distributions of the log2FC values (P95-log2FC) of genes after gentamicin (4 µg/mL) treatment for both libraries. Venn diagram showing the top enriched hits identified from both libraries. (B) Distribution of the number of unique crRNAs targeting genes in the pre-selected 2S library. These genes were either identified as top hits in both libraries (red), or in “qoxA-vs-all” library alone (purple). Color codes match the venn diagram in (A). A Mann-Whitney U test was performed between the two groups and the P- value is shown. [0076] FIG.42 show that Spc2/Spc1 ratios of spacer pairs containing qoxB inversely correlated with Z- scores. Spc2/Spc1 ratios of spacer pairs containing qoxB in 2S library inversely correlated with Z-scores of the non-qoxB gene measured in 1 µg/mL gentamicin at 9 h (P₉₅-Z, gent 1, 9 h), the point at which the second spacer was adapted. Distribution of P₉₅-Zs is also shown. [0077] FIGS.43A-C show sequencing coverage of spacer 1 and spacer 2 of top hits in post- selected 2S libraries containing qoxABC. (A) Distribution of the number of reads corresponding to the non-qoxA spacer at the Spacer 1 position yellow) and Spacer 2 position (purple) for top hits in post-selected 2S libraries containing qoxA. The median number of reads for Spacer 1 and Spacer 2 are also shown. (B, C) Same as (A) except the spacer pairs contained qoxB or qoxC. [0078] FIG.44 shows oligonucleotides used in the subject matter described herein. Figure discloses SEQ ID NOS 15-44, 44-45, 45-46, 46, 47-64, 64, and 65-95, respectively, in order of appearance. [0079] FIG.45 shows spacer sequences used in the subject matter described herein. Figure discloses SEQ ID NOS 96, 96-100, 99, and 101-119, respectively, in order of appearance. [0080] FIG.46 shows plasmids used in the subject matter described herein. DETAILED DESCRIPTION OF THE INVENTION [0081] Genome-wide CRISPR screens enable systematic interrogation of gene function. However, the requisite guide RNA (gRNA) or CRISPR RNAs (crRNAs) libraries are costly to synthesize and their limited diversity substantially reduces coverage and sensitivity of CRISPR screens. [0082] CRISPR-cas9 systems revolutionized genome editing by targeting specific DNA sequences (Sharma S, Petsalaki E. Application of CRISPR-Cas9 Based Genome-Wide Screening Approaches to Study Cellular Signaling Mechanisms. Int J Mol Sci.2018 Apr; 19(4): E933). The complementary guide RNA (gRNA) is crucial in achieving on-target efficiency and minimizing off- target effects (Sharma et al.). Genome-wide CRISPR screens enable systematic interrogation of gene function (1). There is a need for custom gRNA or crRNA libraries in order to systematically screen across the genome for individual gene function (Sharma et al). Generating custom gRNA or crRNA library pools is expensive, time-consuming, and labor intensive and their limited diversity substantially reduces coverage and sensitivity of CRISPR screens (Kweon J, Kim DE, Jang AH, Kim Y. CRISPR/Cas-based customization of pooled CRISPR libraries. PLoS One. Jun 2018; 13(6): e0199473). Pooled oligonucleotide synthesis is almost exclusively used for the generation of the gRNA or crRNA sequences of the libraries (Koferle A, Stricker SH. A Universal Protocol for Large- Scale gRNA Library Production from any DNA source. J Vis Exp. Dec 2017; 130: 56264.). Pooled oligonucleotide synthesis, the gold standard method for generating gRNA or crRNA, takes 1-2 weeks to perform (Koferle et al). [0083] As described herein, the Streptococcus pyogenes CRISPR-Cas adaptation machinery can be reprogrammed to turn bacterial cells into “factories” for generating hundreds of thousands of CRISPR RNAs (crRNAs) covering up to 93% of microbial genomes (Figure 1). Sheared genomic DNA was introduced by electroporation to bacterial cells harboring an engineered, hyperactive CRISPR-Cas adaptation machinery, generating a diverse crRNA library targeting the genome of interest. While this approach is 10 to 100 times cheaper than pooled oligonucleotide synthesis (i.e., the existing technology) and has much faster turnaround time (2 days vs.1-2 weeks), the crRNA library that is generated achieved higher diversity. A hyperactive CRISPR-Cas adaptation machinery consists of the 89-nt tracrRNA, hdCas9, Cas1, Cas2 and Csn2. New spacers are integrated into the empty CRISPR array denoted as “R” (Figure 1A). A diverse gRNA library is generated via hyperactive CRISPR-Cas adaptation (Figure 1B). Number of reads and location of all 129,856 spacers matching the S. aureus RN4220 genome (Figure 1C) are obtained from deep sequencing of the gRNA library generated in (B). The three gap regions correspond to prophages present in the NCBI reference genome but missing in RN4220. Number of spacers mapped to each of all 2,666 annotated genes in RN4220 are shown versus gene length. The percentage of genes covered by one or more (1+) and three or more (3+) spacers are also shown (Figure 1D). Number of reads and location of all 462,382 spacers matching the E. coli MG1655 genome are obtained from deep sequencing. lacI is abnormally enriched due to an additional presence of the gene in helper plasmid, pCCC (Figure 1E). Figure 1F is the same as Figure 1D except the genome is MG1655. “gRNA” in the figure refers to crRNA. [0084] With an average gene targeted by more than 100 distinct crRNAs, it is demonstrated herein that the comprehensive library produced varying degrees of transcriptional repression, which was critical for uncovering novel genes and pathways contributing to phenotypes such as antibiotic sensitivity. These results would most likely be missed by conventional design-based, low-diversity gRNA libraries generated by pooled oligonucleotide synthesis. [0085] By iterating CRISPR adaptation, a diverse dual-crRNA library representing more than 100,000 dual-gene perturbations (Figure 2) was rapidly constructed. Importantly, polarized spacer adaptation, a feature absent in conventional dual‐crRNA libraries generated by synthesis, had the capacity of unraveling historical contingency of genetic interactions and recording evolutionary trajectory (Figure 2). A polarized CRISPR adaptation reveals historical contingency in the acquisition of increasing antibiotic resistance (Figure 2). A schematic shows the sequential construction of a dual-gRNA (2G) library to study the historical contingency of genetic interactions leading to increased antibiotic resistance in bacteria (Figure 2A). Scatter plot shows all gRNA pairs detected in the 2G library (Figure 2B). For each gene, the fraction of gRNAs targeting it among the single-gRNA library and number of genes paired with it in the 2G library are shown. A schematic shows polarized CRISPR adaptation (Figure 2C). Spacers are numbered to indicate the order of adaptation. The ratio of the second spacer and the first spacer (Spc2/Spc1) can be used as an indicator for historical contingency between genes along an evolutionary trajectory. Fitness of genes at 9 hours post-selection in gentamicin (1 ug/mL) was measured using Z-score. Using gRNA pairs containing qoxA as an example, Spc2/Spc1 negatively correlated with Z-scores, indicating that historical contingency can be revealed by the ratio between the two spacers (Figure 2D). “gRNA” in the figure refers to crRNA. “2G” in the figure refers to dual-spacer libraries also referred to as “2S” libraries. [0086] While this technique was performed in S. aureus, as shown, it can be readily adapted to other bacterial systems such as E. coli and wildtype bacteria (such as bacterial pathogens) that have been refractory to genetic manipulation. [0087] This technology is an efficient, cost-effective method for generating diverse crRNA libraries by reprograming the Streptococcus pyogenes CRISPR-Cas adaptation machinery to generate a large quantity of crRNAs. This method generates more than 100 distinct crRNAs per gene target, successfully creating a diverse and comprehensive crRNA library. It is 10-100 times cheaper and significantly faster than pooled oligonucleotide synthesis for crRNA library generation. Furthermore, the technique is adaptable to S. aureus, E. coli, and other bacterial systems that were previously refractory to genomic editing. [0088] The potential applications of this technology include but are not limited to cost- and time- efficient generation of diverse single and dual crRNA libraries for use in CRISPR screens, crRNA library generation in S. aureus, E. coli, and wild-type bacterial systems that were previously refractory to genetic manipulation, generation of crRNA libraries producing transcriptional repression for uncovering genes/pathways that contribute to phenotypes such as antibiotic resistance, generation of crRNA libraries that can record evolutionary trajectory and reveal historical contingency among genetic interactions, dual-spacer and single-spacer library generation. In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); a trans-activating crRNA sequence (tracrRNA); CRISPR-associated endonuclease Cas1 (Cas1); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR-associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to GTTTTAGAGCTATGCTGTTTTGAATGGTCCCAAAAC (SEQ ID NO: 1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; and c) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library. [0089] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: a hyperactive variant of CRISPR-associated endonuclease Cas9 (hCas9); a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); or a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); tracrRNA; Cas1; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising nucleotide sequence at least 80% identical to GTTTTAGAGCTATGCTGTTTTGAATGGTCCCAAAAC (SEQ ID NO: 1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding the CRISPR-associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; c) introducing genomic DNA into the bacterial cells; and d) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library. [0090] In some embodiments, the bacterial cells constitutively express dCas9, tracrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the bacterial cells constitutively express the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding dCas9, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid. [0091] In some embodiments, the method further comprises producing competent bacterial cells before introducing the genomic DNA. In some embodiments, the genomic DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is sheared before introducing the genomic DNA. In some embodiments, the genomic DNA is introduced into the bacterial cells using electroporation. [0092] In some embodiments, in step c) a nucleic acid sequence from the bacterial cell is integrated into the CRISPR array. In some embodiments, a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array. In some embodiments, transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed. In some embodiments, the CRISPR-associated endonuclease is hCas9. In some embodiments, the CRISPR- associated endonuclease is dCas9. In some embodiments, the CRISPR-associated endonuclease is hdCas9. In some embodiments, the CRISPR-associated endonuclease is a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9). In some embodiments, the method further comprises isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. [0093] In some embodiments, the method further comprises generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic. [0094] In some embodiments, the method further comprises isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array. [0095] In some embodiments, the method further comprises subcloning the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence into a plasmid. [0096] In some embodiments, the method further comprises providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a Cas9 lacking endonuclease activity (dCas9) and a /racrRNA, wherein the one or more nucleic acid sequences encoding dCas9 and /racrRNA is under the control of one or more inducible promoters; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array; inducing the expression of the said at least one nucleic acid sequence encoding dCas9 and /racrRNA that is under the control of the one or more inducible promoters; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a crRNA library. In some embodiments, the bacterial cells of the population of bacterial cells and the bacterial cells of the second population of bacterial cells are different species.

[0097] In some embodiments, the method further comprises generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic. In some embodiments, the method further comprises isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

[0098] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: hdCas9; /racrRNA; Casl;

Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

[0099] In some embodiments, the bacterial cells constitutively express hdCas9, /racrRNA, and the CRISPR array and expression of the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding hdCas9, /racrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the method further comprises producing competent bacterial cells before introducing the genomic DNA. In some embodiments, the genomic DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the genomic DNA is sheared before introducing the genomic DNA. In some embodiments, the genomic DNA is introduced into the bacterial cells using electroporation. In some embodiments, a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array. In some embodiments, the method further comprises isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

[0100] In some embodiments, the method further comprising generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound. In some embodiments, the test compound is an antibiotic.

[0101] In some embodiments, the method further comprises isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array. [0102] In some embodiments, the method further comprises providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a tracrRNA; Cas1; Cas2; and Csn2; wherein at least one of the one or more nucleic acid sequences encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 that are not under the control of the one or more inducible promoters are constiutively expressed; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence; inducing the expression of the said at least one nucleic acid sequence encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 that is under the control of the one or more inducible promoters; and introducing genomic DNA into the bacterial cells of the second population of bacterial cells; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a dual crRNA library. [0103] In some embodiments, the method further comprises generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the dual crRNA library with a test compound. In some embodiments, test compound is an antibiotic. In some embodiments, the method further comprises isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises amplifying the one or more nucleic acid sequences encoding the CRISPR array. In some embodiments, the method further comprises detecting the amplified nucleic acid sequences encoding the CRISPR array. In some embodiments, the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing. In some embodiments, the method further comprises identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array. [0104] In certain aspects, the invention provides a method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing two populations of bacterial cells wherein the bacterial cells of each population comprise one or more nucleic acid sequences encoding: a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a trans-activating crRNA sequence (tracrRNA); CRISPR- associated endonuclease Cas1 (Cas1); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR- associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to GTTTTAGAGCTATGCTGTTTTGAATGGTCCCAAAAC (SEQ ID NO: 1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein the one or more nucleic acid sequences encoding hdCas9, tracrRNA, and the CRISPR array are constitutively expressed and are present on a first plasmid and wherein the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are under the control of one or more inducible promoters and are present on a second plasmid, wherein the second plasmid further comprises an origin of replication and a nucleic acid sequence encoding resistance to an antibiotic wherein the origin of replication and nucleic acid sequence encoding resistance to an antibiotic are different between the first and second populations of bacterial cells; b) inducing in both populations of bacterial cells the expression of the said at least one nucleic acid sequence encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of one or more inducible promoters; c) mixing the bacterial cells of both populations of bacterial cells to generate a third population of bacterial cells; d) introducing genomic DNA into the bacterial cells of the third population of bacterial cells; and e) incubating the bacterial cells of step d), thereby producing a population of bacterial cells comprising a crRNA library. [0105] In some embodiments, the bacterial cells of the population of bacterial cells are Gram- positive. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus. In some embodiments, the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA). In some embodiments, the bacterial cells of the population of bacterial cells are Escherichia coli. In some embodiments, the bacterial cells of the population of bacterial cells are pathogenic. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the genomic DNA is of Escherichia coli. In some embodiments, bacterial cells of the population of bacterial cells are Staphylococcus aureus and the bacterial cells of the second population of bacterial cells are Escherichia coli.

[0106] In some embodiments, the racrRN A is a minimal /racrRNA sequence required for targeting. In some embodiments, said minimal /racrRNA sequence comprises 89 nucleotides.

[0107] In certain aspects, the invention provides a population of bacterial cells capable of producing a CRISPR RNA (crRNA) library, said bacterial cells of the population of bacterial cells comprising one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; racrRNA; Casl; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to and (ii) a canonical

[0108] In some embodiments, the bacterial cells constitutively express the CRISPR-associated endonuclease, /racrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 is under the control of the one or more inducible promoters. In some embodiments, the one or more nucleic acid sequences encoding the CRISPR- associated endonuclease, /racrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Casl, Cas2, and Csn2 are present on a second plasmid. In some embodiments, the CRISPR array further comprises a nucleic acid sequence integrated from the bacterial cell.

[0109] In some embodiments, the CRISPR array further comprises a nucleic acid sequence integrated from an exogenous genomic DNA. In some embodiments, transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed. In some embodiments, the CRISPR-associated endonuclease is hCas9. In some embodiments, the CRISPR- associated endonuclease is dCas9. In some embodiments, the CRISPR-associated endonuclease is hdCas9.

[0110] In some embodiments, the exogenous genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells. In some embodiments, the bacterial cells of the population of bacterial cells are Gram-positive. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus. In some embodiments, the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA). In some embodiments, the bacterial cells of the population of bacterial cells are Escherichia coli. In some embodiments, the bacterial cells of the population of bacterial cells are pathogenic. In some embodiments, the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the exogenous genomic DNA is of Escherichia coli. [0111] In some embodiments, the tracrRNA is a minimal tracrRNA sequence required for targeting. In some embodiments, said minimal tracrRNA sequence comprises 89 nucleotides. [0112] In some embodiment, the crRNA library covers at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome. In some embodiments, the crRNA library covers at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome. In some embodiments, the method further comprises isolating said crRNA library. [0113] In certain aspects, the invention provides a first plasmid comprising a nucleic acid sequence encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; tracrRNA; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; operably linked to a constitutive promoter sequence; and a second plasmid comprising a nucleic acid sequence encoding: Cas1; Cas2; and Csn2; operably linked to an inducible promoter sequence. [0114] In some embodiments the repeat sequence comprises a nucleotide sequence at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identical to

ID NO: 1). In some embodiments the repeat sequence comprises (SEQ ID NO: 1). In some

embodiments the repeat sequence consists of In some

embodiments the repeat sequence comprises equivalent sequences to that can be duplicated

after each adaptation event. CRISPR ARRAY [0115] In some embodiments, the CRISPR array comprises (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event. In some embodiments the repeat sequence comprises a nucleotide sequence at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% identical to

ID NO: 1). In some embodiments the repeat sequence comprises

embodiments the repeat sequence consists of

after each adaptation event. In some embodiments, the canonical sequence comprises about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, or 200 nucleotides of the canonical sequence. In some embodiments, the canonical sequence comprises a promoter and a leader- anchoring sequence. In some embodiments, the acquired spacers are inserted into the leader- anchoring sequence, which is the site where newly acquired spacers are inserted. See McGinn, J and Marraffini, L.A., 2016, Mol Cell, 616-623, the contents of which are incorporated by reference. [0116] After one adaptation event, the repeat sequence becomes duplicated, and the newly acquired spacer is flanked by the two repeats. For example, the new CRISPR array becomes:

the stretch of Ns represents the spacer sequence. In some embodiments, the spacer sequence is about 30 or 31 nucleotides in length. EXAMPLES Example 1 – Genome-wide perturbations via CRISPR adaptation reveal complex genetics of antibiotic sensitivity Abstract [0117] Genome-wide CRISPR screens enable systematic interrogation of gene function. However, the requisite guide RNA (gRNA) libraries are costly to synthesize and their limited diversity substantially reduces coverage and sensitivity of CRISPR screens. The Streptococcus pyogenes CRISPR-Cas adaptation machinery was reprogrammed to turn bacterial cells into “factories” for generating hundreds of thousands of crRNAs s covering up to 93% of microbial genomes. With an average gene targeted by more than 100 distinct crRNAs, comprehensive CRISPRi library described herein produced varying degrees of transcriptional repression, which proved critical for uncovering novel pathways contributing to antibiotic sensitivity. Furthermore, a diverse dual-spacer library was rapidly constructed representing more than 100,000 dual-gene perturbations by iterating CRISPR adaptation. Importantly, polarized spacer adaptation, a feature absent in conventional dual-spacer libraries, revealed historical contingency of genetic interactions that led to increasing antibiotic resistance. The approach described herein circumvents the expense, labor, and time required for synthesis and cloning of gRNAs, thus allowing the generation of diverse CRISPRi libraries in wild-type bacteria that have been refractory to genetic manipulation. This approach can be readily adapted to other systems to unravel the genetic basis and architecture of diverse traits. Introduction [0118] Functional genetic screens help elucidate the genetic basis of cellular and organismal phenotypes. Recent advances in CRISPR-Cas technology have enabled a wealth of discoveries in diverse model systems by facilitating genome-wide mutation, transcriptional repression (CRISPRi) and activation (CRISPRa) (1-4). To date, the most widely used CRISPR-Cas technology is the S. pyogenes Cas9 system. By changing the sequence of a short guide RNA (gRNA) that associates with it, Cas9, the endonuclease, can be easily programmed to cleave any genetic locus of interest (5). [0119] Similarly, a catalytically inactive version of the endonuclease (dCas9) can sterically hinder transcription at any desired genetic locus that matches the gRNA (6, 7). Currently, genome- wide CRISPRi libraries are generated by rationally designing multiple gRNAs targeting each gene and synthesizing them in array-based oligonucleotide pools. However, these libraries are costly and contain many faulty guides as knowledge of molecular rules governing gRNA efficacy are incomplete. Consequently, most current libraries accommodate less than 10 gRNAs per gene, resulting in limited genome coverage that severely limits the sensitivity of CRISPR screens. Results [0120] Generation of genome-wide crRNA libraries by CRISPR-Cas adaptation as shown in Figures 3A-C. Generation of genome-wide crRNA libraries by CRISPR-Cas adaptation in S. aureus as shown in Figures 4A-F. Generation of genome-wide crRNA libraries by CRISPR-Cas adaptation in for other bacteria as shown in Figures 5A-F. Genome-wide crRNA libraries identify known and novel pathways of aminoglycoside sensitivity as shown in Figures 6A-E. Additional hits and validation of gentamicin sensitivity as shown in Figures 7A-D. “One-vs-all” libraries identify genetic interactions that strengthen antibiotic resistance as shown in Figures 8A-E. Construction of a dual- spacer (2S) library by iterating CRISPR adaptation as shown in Figures 9A-C. Polarized CRISPR adaptation reveals historical contingency in the acquisition of increasing antibiotic resistance as shown in Figures 10A-E. [0121] A schematic of the bacterial fatty acid synthesis is shown in Figure 11.5-(tetradecyloxy)-2- furoic acid (TOFA), a FAS inhibitor potentiated the effect of gentamicin as shown by bacterial growth (OD₆₀₀ measured at 18 h) and Cerulenin potentiates the effect of gentamicin as shown by ^{bacterial growth (OD} ₆₀₀ ^{measured at 18 h) (Figure 12).} Example 2 – Comprehensive genome-wide perturbations via CRISPR adaptation reveal complex genetics of antibiotic sensitivity Summary [0122] Genome-wide CRISPR screens enable systematic interrogation of gene function. However, guide-RNA libraries are costly to synthesize and their limited diversity compromises the sensitivity of CRISPR screens. The Streptococcus pyogenes CRISPR-Cas adaptation machinery was reprogrammed to turn bacterial cells into “factories” for generating hundreds of thousands of crRNAs covering up to 95% of all targetable genomic sites. With an average gene targeted by >100 distinct crRNAs, these CRISPRi libraries produced varying degrees of transcriptional repression which proved critical for uncovering novel antibiotic tolerance pathways. Furthermore, by iterating CRISPR adaptation, diverse dual- crRNA libraries representing >100,000 dual-gene perturbations were generated. The polarized nature of spacer adaptation revealed the historical contingency in the step-wise acquisition of genetic perturbations leading to increasing antibiotic resistance. This approach circumvents the expense, labor, and time required for synthesis and cloning of gRNAs, thus allowing generation of CRISPRi libraries in wild-type bacteria refractory to routine genetic manipulation. Introduction [0123] Functional genetic screens help elucidate the genetic basis of cellular and organismal phenotypes. Recent advances in CRISPR-Cas technology have enabled a wealth of discoveries in diverse model systems by facilitating genome-wide mutation, transcriptional repression (CRISPRi) and activation (CRISPRa) (Gilbert et al., 2014; Sanjana, 2017; Shalem et al., 2014; Wang et al., 2014). To date, the most widely used CRISPR-Cas technology is the S. pyogenes Cas9 system. By changing the sequence of a short guide RNA (gRNA) that associates with it, Cas9, the endonuclease, can be easily programmed to cleave any genetic locus of interest (Jinek et al., 2012). Similarly, a catalytically inactive version of the endonuclease (dCas9) can sterically hinder transcription at any desired genetic locus that matches the gRNA (Bikard et al., 2013; Qi et al., 2013).

[0124] Currently, genome-wide CRISPR libraries are generated by designing multiple gRNAs targeting each gene and synthesizing them in array-based oligonucleotide pools. However, these libraries are costly and contain many faulty guides as knowledge of molecular rules governing gRNA efficacy is incomplete. Consequently, most genome-wide libraries accommodate 10 or fewer gRNAs per gene (Figure 18 and references therein), resulting in limited genome coverage that severely compromises the sensitivity of CRISPR screens.

[0125] The S. pyogenes CRISPR-Cas adaptation machinery was re-purposed as a “factory” to turn externally supplied DNA into hundreds of thousands of unique crRNAs in bacteria (Figure 13 A). A hallmark of the CRISPR-Cas immune response is spacer adaptation (McGinn and Marraffmi, 2018), a process in which CRISPR machinery integrates foreign DNA such as fragmented phage DNA into the associated spacer-repeat array, the precursor to crRNAs. For this reason, “spacer” and “crRNA” are used interchangeably herein. Similar to conventional engineered gRNAs, crRNAs are composed of a 20-nt target-recognizing sequence and a short structural element. In addition, the canonical crRNA needs to base-pair with another small RNA called /racrRNA for proper processing and targeting (Deltcheva et al., 2011), whereas gRNAs are a single RNA species that combines the functional and structural features of both crRNA and /racrRNA

[0126] By externally supplying genomic DNA of interest to Staphylococcus aureus cells harboring a hyperactive CRISPR-Cas adaptation machinery, near-saturating genome-wide crRNA libraries were generated in bacteria, with an average gene covered by up to 100 crRNAs. This library can be directly used in S. aureus , or sub-cloned into other organisms (Figure 13 A). Importantly, as this comprehensive pool of crRNAs produced varying degrees of transcriptional repression and significantly raised statistical power, novel pathways contributing to antibiotic sensitivity was discovered that would have been missed by conventional gRNA libraries with lower diversity. Furthermore, by iterating the CRISPR-Cas adaptation process, an economical dual-spacer library was rapidly constructed that represented more than 100,000 dual -gene perturbations, and identified pairwise perturbations that strengthened antibiotic resistance. Critically, polarized spacer adaptation, a feature absent in conventional dual-gRNA libraries, revealed how historical contingency among these perturbations could constrain adaptive evolution. [0127] This pipeline is highly efficient, inexpensive and portable. Genome-wide crRNA libraries via CRISPR adaptation can be made with standard lab equipment in as short as one day (Figure 19A). Circumventing tedious cloning and transformation steps, crRNA libraries were directly generate in wild-type bacterial strains such as methicillin-resistant Staphylococcus aureus (MRSA), which is not amenable to easy genetic manipulation, let alone CRISPR screens. It is demonstrated that the system can function as a robust and portable platform for generating genome-wide crRNA libraries for other microbes such as Escherichia coli. Results Rapid generation of genome-wide crRNA libraries using a hyperactive CRISPR-Cas adaptation machinery [0128] DNA breaks promote spacer adaptation in the type I E. coli CRISPR-Cas system (Levy et al., 2015). Without being bound by theory, it is thought that DNA may be in general a preferred substrate for the CRISPR adaptation machinery – a potential factory for the production of crRNA libraries. To test this, the type II S. pyogenes CRISPR-Cas adaptation machinery was employed (Heler et al., 2015), which includes a single CRISPR repeat, the minimal tracrRNA (89-nt) required for targeting (Deltcheva et al., 2011), and all four cas genes (Figure 13A) including hyper-dead-Cas9 (hdCas9), a nuclease-dead Cas9 variant that enables hyperactive spacer adaptation (Heler et al., 2017). These elements were over-expressed in S. aureus RN4220 (Nair et al., 2011), supplied cells with sheared genomic DNA by electroporation, and tested whether they could be “transformed” into functional spacers. Using enrichment primers developed previously (Heler et al., 2015), a PCR assay detected that the hyperactive adaption system allowed 0.1%-1% of the cell population to acquire a _{single spacer (Figure 19B). Given that bacterial cultures at exponential phase contain ~10} ⁸ _cells/mL, creating a diverse crRNA library covering the entire genome is thus, straightforward in principle. [0129] Deep sequencing of the adapted spacers confirmed the generation of a comprehensive genome-wide crRNA library. In the most optimized protocol, in which the CRISPR adaptation was included during re-growth (Figure 19A), 91% of all adapted spacers matched the host chromosome (Lib-2 in Figure 19C). The great majority of chromosomal spacers had the correct NGG PAM (Figure 19D) and the same length as canonical spacers (i.e., 30 or 31 nt, Figure 19E). An optimized procedure involving competent cell making and tightly controlled expression of CRISPR adaptation was critical for creating a diverse library (Figure 19A and Methods). For instance, though adaptation occurred in as many as 21% of the population using a constitutively expressed CRISPR machinery, the great majority of the spacers were derived from helper plasmids (Figure 20). S. aureus RN4220 cells can harbor a constitutively expressed hyperactive CRISPR adaptation machinery (Figure 20A). Plasmid pTHR (aka. pWJ402) contains tracr, hdcas9 (hyper dead Cas9, D10A, I473F, H840A), an empty CRISPR array (R) and a chloramphenicol resistance marker (cat). Plasmid pCCC aka. pAV112B contains cas1, cas2 and csn2 under a strong constitutive pTet promoter. pCCC also contains a tetracycline resistance marker (tet). Overnight cells harboring these two plasmids were re- grown, followed by electroporation of S. aureus genomic DNA (gDNA) (Figure 20B). Spacer adaptation was detected in both overnight (O/N) cells and ells electroporated with gDNA even using non-enrichment primers (W1307 and L401). The percentage of cells with an adapted spacer is also shown (Figure 20B). Deep sequencing revealed spacer origin of adapted overnight (O/N) cells and cells electroporated with gDNA (Figure 20C). [0130] As both spacer adaptation and CRISPR targeting rely on the presence of protospacer- adjacent-motifs (PAMs) (Mojica et al., 2009), the maximum number of targetable sites in a given genome is equal to the total number of PAMs within it. Given that there are 136,928 PAMs (i.e., NGG) in the S. aureus RN4220 genome, Lib-4, the library sequenced the deepest (~25 million reads), contained at least 129,856 unique chromosomal spacers, thus representing 95% of all targetable sites in the genome (Figure 13B). All 2,666 annotated genes were targeted by at least one spacer (Figure 13C). Enrichment PCR primers and downstream size exclusion steps ensured that the majority of sequenced reads contained one spacer (Figure 19A), such that 7-10 million sequencing reads provided sufficient coverage for libraries of this scale. [0131] The libraries generated via CRISPR adaptation had a bias toward the origin of replication (Ori) (Figure 13B), suggesting that replicating genomic DNA inside the cell may compete with externally supplied DNA for CRISPR adaptation. Indeed, upon electroporating E. coli genomic DNA, only 17% of the spacers were observed to be mapped to its genome (Figure 21A), while the great majority (78%) belonged to S. aureus with a similar bias toward Ori (Figure 21B). [0132] With the goal to turn S. aureus into a “factory” for generating comprehensive crRNA libraries for any organism of interest (Figure 13A), dCas9 was replaced with wild-type Cas9 (Figure 22A) in order to minimize adaptation of internal DNA. Sure enough, when electroporated with E. coli DNA, 90% of the spacers were mapped to its genome (Figure 22B) with no apparent bias toward Ori (Figure 13D). With 462,382 unique spacers (out of 542,073 targetable sites), the E. coli library was more uniform (CV_gene = 0.34) and diverse, covering all 4,498 genes with an average gene targeted by 103 spacers (Figures 13D and E). Finally, a functional genome-wide CRISPRi system was constructed by successfully sub-cloning 90% of the spacers into E. coli harboring an inducible dCas9 (Figure 22A). [0133] In addition to the laboratory strain, RN4220, strong adaptation events were also observed in wild-type, pathogenic S. aureus strains such as MRSA (Figure 19B). This is a critical advance as restriction-modification and low transformation efficiency had rendered them extremely difficult for genetic manipulation, let alone comprehensive libraries for genetic screens. Additionally, since the majority of adapted spacers were derived from internal genomic DNA (Figure 21A), the protocol can be further simplified by omitting electroporation of external DNA if necessary. Mild adaptation events were also detected when the CRISPR-Cas machinery was cloned into E. coli (Figure 23), suggesting general applicability of the system in other bacteria. As the CRISPR machinery is gram- positive in origin, codon optimization may be necessary for optimal activity. Comprehensive CRISPRi libraries identify known and novel pathways of aminoglycoside sensitivity [0134] To explore the utility of the comprehensive genome-wide CRISPRi library disclosed herein, S. aureus was treated with sub-lethal concentration of gentamicin in triplicates and uncovered known and novel pathways contributing to aminoglycoside sensitivity (Figure 14A). The fitness effect of each crRNA under antibiotic exposure was determined by its enrichment/depletion relative to the initial unselected library using the Z-score (Girgis et al., 2007), with Z-scores of all the individual crRNAs targeting a gene averaged into Mean-Z (Methods). As a result, genes involved in oxidative phosphorylation (qoxABCD) and heme biosynthesis (ctaB) had the highest Mean-Zs (Figure 24A), consistent with the established body of work that disruption of electron transport chain (ETC) changes membrane potential and reduces uptake of aminoglycosides (Girgis et al., 2009; Taber et al., 1987). [0135] However, even for genes with the highest Mean-Z, the average antisense crRNAs were substantially more enriched than sense crRNAs (Figures 14C, 25A and B), consistent with the known efficacy of antisense crRNAs in transcriptional repression (Bikard et al., 2013; Qi et al., 2013). This suggests that Mean-Z could severely underestimate fitness effects due to the conflation of ineffective sense crRNAs. To better quantify each gene’s contribution to drug resistance while avoiding crRNAs with potential off-target effect (see detailed discussion in Methods), the 95^th percentile (P₉₅-Z) of the Z-score of all crRNAs targeting each gene was calculated. This analysis led to the discovery of many genes involved in pathways that feed into the ETC (Figures 14B and 26A), which were reported in multiple different studies (Bayer et al., 2006; Kinkel et al., 2013; Rajagopal et al., 2016). These genes included ctaABM, ndh and mnhABCDEFG, and multiple antisense crRNAs targeting them were highly enriched (Figures 25B-D), providing strong evidence for their roles in aminoglycoside resistance. Notably, inactivation of genes in the hem and men operons were also seen in resistant clinical isolates (Kahl et al., 2016; Lannergard et al., 2008). [0136] crRNAs targeting essential genes (Chaudhuri et al., 2009) such as mvaD and mvaK2 (mevalonate pathway) and topA (topoisomerase I) were also enriched (Figure 14B), they are novel loci of aminoglycoside sensitivity. crRNAs targeting a few of the essential ribosomal genes were also significantly enriched, especially after a shorter period of gentamicin selection (Figures 14E and 27). The fact that transcriptional inhibition of a few ribosomal components improved gentamicin tolerance suggest that specific compositional or structural perturbations to the ribosome may abrogate its normal targeting by aminoglycosides (Taber et al., 1987). [0137] While many antisense crRNAs were almost always enriched for non-essential genes (Figure 25), a tendency was observed that for essential genes only a small portion of sense crRNAs targeting them were enriched, after both long and short periods of gentamicin selection (Figures 14D and 28). This is consistent with weak transcriptional inhibition by sense crRNAs (Bikard et al., 2013; Qi et al., 2013), as stronger effects on essential genes are expected to be lethal. Indeed, an attempt to clone three antisense crRNAs targeting the essential rplB gene failed (Figure 28C). These results strongly indicate that the comprehensive CRISPRi library as disclosed herein produced a broad range of scale in transcriptional repression, leading to the selection of mildly- suppressing sense crRNAs targeting regions of the essential genes not easily predictable. Z-scores of all individual crRNAs targeting the mva operon (mvaK1DK2) 18-hour and 4.5-hour post-selection in gentamicin (1.0 µg/mL) are shown in Figure 28A.18-hour experiments were done in triplicates (lib-1 through ib-3). 4.5-hour experiments were either treated with gentamicin or grown in plain media. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. For 18-hour experiments, horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs, respectively. For 4.5-hour experiments, horizontal purple and orange solid lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample treated with gentamicin, and horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense rRNAs from the sample grown in plain media. Genes can be a part of the operon encoding ribosomal genes (rplP – rplC) (Figure 28B). Z-scores of all individual crRNAs targeting rplB after 18-hour post-selection in gentamicin are shown in Figure 28C. [0138] The fact that only a handful of these weaker sense crRNAs were enriched highlights the challenge faced by conventional design-based, low-diversity libraries. To assess the sensitivity of these libraries, 100 computational simulations were performed. In each simulation, 10 crRNAs per gene were randomly selected from the comprehensive library and P₉₅-Z was calculated as before (Figure 29A).9 out of 30 top hits identified in the comprehensive library were no longer significantly enriched in more than 50% of the simulations. More severely, novel essential hits such as mvaD and mvaK2 were missed in more than 80% of the simulations. [0139] As long-term extreme selection of the CRISPRi library in antibiotics could obscure quantification of negative fitness effects, a shorter-period (4.5 h) selection was also performed (Figure 27A), which identified crRNAs that were both significantly enriched (Figures 14E and 27B) and depleted (Figure 14F). Particularly, crRNA perturbations that truly potentiated the effect of antibiotics were expected to be significantly depleted in gentamicin but not in plain medium (Figure 27A and Methods). These included a membrane protease (ftzH), the ATP synthase (atpABDFI) and a two-component regulatory system (graR), all known to sensitize cells to aminoglycosides when disrupted (Hinz et al., 2011; Vestergaard et al., 2017; Yang et al., 2012). New hits included accACD and fabZ, which belong to the essential fatty acid synthesis (FAS) pathway (Figures 14F, 30A-D and 32A). Importantly, the simulation showed that all these hits would be missed more than 50% of the time if using conventional low-diversity libraries (Figure 29B), highlighting the general challenge in capturing depletion events. As such, the comprehensiveness of the crRNA library significantly elevated the sensitivity and statistical power (Figure 29C) for the discovery of novel antibiotic- potentiating pathways in bacteria. [0140] Figures 30A-G show Z-scores of all individual crRNAs targeting representative genes depleted in gentamicin (1.0 µg/mL) after 4.5-hour treatment. Figure 30A shows Z-scores of all individual crRNAs targeting accD treated with gentamicin (1.0 µg/mL) or grown in plain media for 4.5 hours. Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Horizontal purple and orange solid lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample treated with gentamicin, and horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense crRNAs from the sample grown in plain media. The number of crRNAs targeting the gene (N) and the P value of the Mann-Whitney U test (P) are also shown. Figures 30B-G show Z-scores for the genes accA, accC, fabZ, atpB, atpF, and ftzH, respectively. [0141] Figures 32A-C show that inhibition of fatty acid synthesis potentiated the effect of gentamicin. Schematic of the bacterial fatty acid synthesis (FAS) pathway borrowed from a review article Zhang et al., 2006) with the following modifications. Boxes indicate genes in which crRNAs were significantly depleted after exposure to gentamicin (1.0 µg/mL) for 4.5 hours. TOFA (5- tetradecyloxy)-2-furoic acid) and cerulenin inhibits the acetyl-CoA carboxylase complex AccABCD and FabB/F, respectively. Growth of S. aureus RN4220 cells in media containing various concentrations of gentamicin and TOFA (Figure 32B). Cerulenin potentiates the effect of gentamicin ^{as shown by bacterial growth (OD} ₆₀₀ ^{measured at 18 h) (Figure 32C).} [0142] crRNAs targeting representative top enriched and depleted genes were validated for their effects on drug resistance in liquid culture (Figure 33). The great majority of them also showed expected changes in MIC measured on agar plates (Figure 14G), with the exception of some ribosomal genes (Figure 33B). Importantly, as the newly discovered pathways such as mevalonate and FAS are essential, gene knock-out experiments were not feasible. Instead, they were validated by chemical inhibition. Mevalonate and FAS inhibitors strongly antagonized (Figures 14H and 34) and synergized (Figures 14I and 32) with gentamicin, respectively, consistent with the enrichment and depletion of the respective crRNAs seen in the screen (Figures 14B and F). [0143] Figures 33A-D show growth curves and MICs. Growth of S. aureus RN4220 cells harboring selected top enriched spacers as shown in Figure 14B in plain medium or gentamicin (1 µg/mL) are shown in Figure 33A. Growth of S. aureus RN4220 cells harboring spacers that target the 16S rRNA (R0001-R0005) and two structural components of the ribosome, rplB and rplC, in plain or gentamicin (1.0 µg/mL) is shown in Figure 33B. All five 16S rRNAs (R0001-R0005) share extensive sequence homology and therefore the spacers designed match all of them. The MICs (triplicates) are also shown. Growth of S. aureus RN4220 cells harboring spacers targeting an intergenic region (genomic locations shown in parenthesis) in plain medium or gentamicin (1.0 µg/mL) is shown in Figure 33C. The MICs (triplicates) are also shown. Growth of S. aureus RN4220 cells harboring selected top depleted spacers as shown in Figure 14F in plain medium or gentamicin (0.5 µg/mL) is shown in Figure 33D. For all panels, representative growth curves of at least three independent assays are shown. [0144] Figures 34A-B show inhibition of mevalonate pathway antagonized gentamicin. Figure 34A shows schematic of the mevalonate pathway. Underlined genes had crRNAs that were significantly enriched after exposure to low (1.0 µg/mL) and high (4.0 µg/mL) concentrations of gentamicin. Boxed genes had crRNAs that were significantly enriched only after exposure to high concentration of gentamicin when a second gene was co-repressed (i.e., identified in 2S or individual “one-vs-all” libraries). Vanadyl sulfate (VS) inhibits the mevalonate kinase (Gharehbeglou et al., 2015). Figure 34B shows growth of S. aureus RN4220 cells in media containing gentamicin (1.0 µg/mL) and various concentrations of VS. Units are in µg/mL. See also Figure 14H. Dual-spacer perturbations reveal epistasis and historical contingency in the acquisition of antibiotic resistance [0145] Microbes can adapt to extreme environments through sequential accumulation of mutations. Disruption of ETC and related pathways (e.g., qoxA and ndh) allowed cells to grow better in sub-lethal dose of gentamicin (Figure 14B). Can repression of additional pathways further strengthen antibiotic resistance? To test this, a comprehensive “one-vs-all” library was created containing a universal qoxA-targeting spacer (“qoxA-vs-all” library, Figure 15A). After selection in high dose of gentamicin (4 µg/mL), many spacer pairs targeting ETC and pathways further upstream of it (Figure 26A), as well as novel operons such as 01269-01271 were significantly enriched (“gent4, log₂FC(qoxA+X)” column in Figures 15B). [0146] Again, sense rather than antisense crRNAs tended to be more enriched for essential genes (Figure 35). Figures 35A-E show Log₂FCs of all individual crRNAs targeting representative essential genes enriched in “qoxA-vs-all” libraries in gentamicin (4.0 µg/mL) after 18-hour treatment. Log₂FCs of all individual crRNAs targeting the mva operon (mvaK1DK2) after “qoxA-vs- all” libraries (triplicates) were subjected to gentamicin (4.0 µg/mL) for 18-hour (Figure 35A). Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Figures 35B- E show Log₂FCs the genes part of the operon encoding ribosomal genes (rplP – rplC), atoB, mvaS and SAOUHSC_00760, respectively. The fitness of some top hits was validated by MIC measurement (Figure 36) and pairwise competition (Figures 37A-E), which was more quantitative. Moreover, with the exception of mnh, most genes were not enriched without the qoxA-targeting spacer in the same condition (“gent4, log₂FC(X)” column in Figures 15B), suggesting positive epistatic interactions with qoxA. Epistasis (ɛ) was estimated using fitness effects measured by log₂FC (Figure 15B and Methods), as it was highly correlated with fitness as measured by pairwise competition (Figures 37A-F). Fitness was also measured in gentamicin (1.0 µg/mL) (Figure 37G). [0147] Beyond simplicity, another unique advantage of the CRISPR-Cas system is its multiplexity and polarity – the combination and order of spacers in the CRISPR array have the capacity to reveal the effects of multiple genetic perturbations and the historical contingency among them. To demonstrate this, a single-spacer (1S) library was treated with low concentration of gentamicin (low [gent]) for 9 hours, in order to enrich for spacers that confer mild fitness benefits. Next, the resulting spacer pool was sub-cloned, a dual-spacer (2S) library was generated by iterating CRISPR adaptation, and it was subjected to a higher concentration of gentamicin (high [gent]) (Figure 16A). [0148] This rapid and economical pipeline generated a 2S library with at least 237,650 unique spacer pairs, representing 105,030 non-redundant pairwise genetic perturbations (Figure 16B). This diversity is comparable to costly array-synthesized dual-gRNA libraries by recent studies, which included 2,628 and 23,652 pairwise genetic perturbations, respectively (Han et al., 2017; Shen et al., 2017). At the gene level, the greatest number of pairwise connections made in the 2S library skewed toward the more abundant genes in the preceding 1S library (Figures 16B and C), which was the result of selection in low [gent] (Figure 16A). The top 10 genes made an average of 1221 connections, and the top 500 genes made an average of 321 connections (Figure 16C). Genes such as qoxABC were the most abundant, making up ~20% of all crRNAs in the preceding 1S library. The abundance of qoxABC-targeting crRNAs further rose to 86% (Figure 38) after selection in high [gent], highlighting their importance in boosting drug resistance. As expected, crRNAs targeting other components of the ETC (Figure 26A) such as menA, mvaS and ndh were among the most abundant species. [0149] Importantly, a large proportion of the highly enriched pairwise perturbations after selection in high [gent] were corroborated by their respective “one-vs-all” libraries. For instance, this is the case for 88% (23/26) of enriched qoxA-containing (Figures 4D and 39A) and all ndh- containing pairs (Figure 40). Figures 40A-B show that top enriched spacer pairs containing ndh in 2S library were corroborated by ndh-vs-all” library. Comparison of the top enriched hits (i.e., spacer pairs) containing ndh in 2S library and “ndh-vs-all” library is shown in Figure 40A. Histograms showing the distributions of the log₂FC values (P₉₅-log₂FC) of genes after gentamicin (4 µg/mL) treatment for both libraries. Venn diagram showing the top enriched hits identified from both libraries. Distribution of the number of unique crRNAs targeting genes in the pre-selected 2S library is shown in Figure 40B. These genes were either identified as top hits in both libraries (red), or in “ndh-vs-all” library alone (purple). [0150] 91% (21/23) of enriched pairs containing qoxB were also corroborated using hits from “qoxA-vs-all” library as a proxy (Figure 41). Figures 41A-B show top enriched spacer pairs containing qoxB in 2S library were corroborated by qoxA-vs-all” library. Comparison of the top enriched hits (i.e., spacer pairs) containing qoxB in 2S library and “qoxA- s-all” library is shown in Figure 41A. Histograms showing the distributions of the log₂FC values (P₉₅-log₂FC) of genes after gentamicin (4 µg/mL) treatment for both libraries (Figure 41B). Venn diagram shows the top enriched hits identified from both libraries (Figure 41A). Distribution of the number of unique crRNAs targeting genes in the pre-selected 2S library is shown in Figure 41B. These genes were either identified as top hits in both libraries (red), or in “qoxA-vs-all” library alone (purple). Notably, genes that were hit only in “one-vs-all” libraries but missing in 2S library tended to have significantly fewer crRNAs targeting them in the pre-selected 2S library (Figures 16E, 40 and 41). This suggests that although the 2S library contained far more pairwise perturbations (~100,000 pairs) than “one-vs-all” libraries (2,666 pairs) at the gene level, the number of crRNAs per gene it contained were fewer than those of “one-vs-all” libraries (Figure 39B and C), limiting its sensitivity. Therefore, while the more diverse 2S libraries could rapidly capture major beneficial pairwise perturbations for given phenotypes, “one-vs-all” and potentially “several-vs-all” libraries were necessary to thoroughly sample and quantify the fitness effect of every gene in the genome. The top 26 most enriched spacer pairs containing qoxA in 2S library from Figure 16D and their log₂FC values are shown in Figure 39A. [0151] It is not surprising that operons and genes such as men, mva and ndh synergized with qoxA to confer resistance to high [gent], as they already boosted fitness in low [gent] (“gent1, log₂FC(X)” column in Figure 15B). By contrast, genes such as the cyd operon, which encodes a second terminal oxidase of the ETC in addition to qox (Figure 26A), were not enriched in either low or high [gent] alone (Figure 15B). Thus, enrichment of cyd in the presence of qoxA suggests historical contingency (Blount et al., 2008) – under the selection regime described herein, while repression of cyd was clearly a path toward higher resistance to gentamicin, it was not accessible unless repression of qoxA was reached first. Other genes and pathways in which crRNAs were highly enriched are shown in Figure 26B. [0152] Multiple sequencing rounds are often required to unravel the evolutionary trajectory toward given phenotypes (Good et al., 2017), which could inform strategies in controlling drug resistance. Since CRISPR adaptation is polarized (McGinn and Marraffini, 2018) – new spacers are always added at the leader-proximal end of the array (Figure 17A) – it has the capacity to record the temporal sequence of biological events (Shipman et al., 2016; Tang and Liu, 2018). Among spacer pairs containing both qoxA and cydA post-selection in high [gent], cydA was significantly more abundant as the second spacer (Spc2) than as the first spacer (Spc1), confirming its contingency on qoxA. This is further corroborated by the competition and MIC assays: whereas cydA-targeting cells showed equally poor fitness as WT cells in low [gent], dual-spacers targeting qoxA and cydA provided significantly higher fitness than qoxA alone in high [gent] (Figures 36, 37E and G). [0153] To investigate whether the order of CRISPR spacers is a good indicator for historical contingency, spacer pairs that were enriched in both “qoxA-vs-all” and 2S libraries were analyzed. Genes that are contingent on qoxA are expected to not confer high fitness by themselves in low [gent], and should therefore have low Z-scores and consequently, high Spc2/Spc1 ratios. Indeed, the Spc2/Spc1 ratios in the post-selected 2S libraries inversely correlated with the Z-scores in low [gent] at 9-hour (Figure 17B), which was the time the second spacer was adapted (Figure 16A). This is equally true for dual-spacers containing qoxB (Figure 42), the second most abundant species post- selection (Figure 38). Thus, by sequencing just a single terminal time point, the order of the two CRISPR spacers can reveal candidate loci with historical contingency in the acquisition of increasing antibiotic resistance. Such chronological information is not present in dual-gRNA libraries generated in the conventional way. Discussion [0154] Harnessing the natural capacity of a hyperactive CRISPR-Cas adaptation machinery, the current study establishes a rapid and economical pipeline that converts exogenous DNA into comprehensive genome-wide crRNA libraries targeting bacterial genomes of interest. With an average gene covered by up to 103 crRNAs, the libraries disclosed herein enabled a broad range of transcriptional perturbations. This comprehensiveness also significantly increased the statistical power and led to the discovery of novel and essential pathways for aminoglycoside sensitivity in ways that conventional low-diversity libraries miss. The pipeline can be readily adapted to other bacterial species, including wild-type, clinically relevant bacteria that have not been amenable to standard genetic manipulation, and potentially higher eukaryotes with further optimization. [0155] The crRNA libraries generated in situ (i.e., in S. aureus) could be directly used in both laboratory and wild-type strains without cloning, a significant improvement over current methodologies. However, as both the externally supplied and internally replicating DNA could compete for the access to the CRISPR adaptation machinery, these libraries had a bias toward Ori. Nevertheless, in making libraries of other species such as E. coli in S. aureus, dCas9 was replaced with Cas9, resulting in diverse libraries with much higher uniformity. As such, crRNA libraries targeting organisms of interest could be “manufactured” in S. aureus, followed by sub-cloning (Figure 22A). The genome-wide E. coli library consisted of at least 462,382 unique spacers, which would cost nearly $20,000 through conventional synthesis methods. The upper limit of library diversity has not been tested, however, without being bound by theory it is thought that the quality of crRNA libraries could be further improved in a few ways. Optimization by laboratory evolution of the adaptation machinery (Cas9, Cas1, Cas2 and Csn2) could enhance the frequency of spacer adaptation and thus the library diversity. Inducing CRISPR machinery at different growth phases where DNA replication at the Ori is minimal or actively inhibited may act to reduce biased spacer adaptation at this location when making in situ libraries. [0156] In one of the computational pipelines described herein, ~30% of spacers were filtered out with off-target potentials. While this filter effectively removed faulty hits, elimination of such a large number of spacers also compromised the sensitivity and statistical power of 1S and 2S libraries (See detailed discussion in Methods). However, because genes targeted by more than one spacer with true non-neutral off-target site were rare, it was shown that removal of crRNAs with the strongest and the weakest fitness effect for each gene alone effectively eliminated faulty hits. More importantly, it was found that the bona fide top hit genes validated by us and others were consistently targeted by multiple significantly enriched or depleted crRNAs across multiple replicates, whereas faulty hits contained only one such crRNA (Figure 31). Figures 31A-E show Z-scores of all individual crRNAs targeting representative genes containing non-neutral off-target sites. Figure 31A shows Z-scores of all individual crRNAs targeting SAOUHSC_0125718-hour post-selection in gentamicin (1.0 µg/mL) (triplicates). Vertical gray solid and dotted lines indicate the start and end of the genes, respectively. Horizontal purple and orange dotted lines indicate the mean Z-scores of all antisense and sense crRNAs, respectively. The crRNA with abnormally high Z-score is circled and the base-pairing between the spacer and the non-neutral off-target site is shown. A gene (e.g., qoxA, qoxC and mnhA) is defined as non-neutral if repression of it by CRISPRi increases cell’s fitness in gentamicin (Figure 14B). Figure 31 B-E show Z-scores for the genes SAOUHSC_00628, SAOUHSC_03037, SAOUHSC_03016, and SAOUHSC_00266. For these reasons, the off-target filter was not applied. [0157] By iterating CRISPR adaptation, a 2S library was created (Figure 16) whose diversity is comparable to costly array-synthesized gRNA libraries by recent studies (Han et al., 2017; Shen et al., 2017). The 2S library revealed many pairwise perturbations that boosted antibiotic resistance, with the great majority of which corroborated by the “one-vs-all” libraries and fitness measurement. One unique feature of the 2S library is that the order of the two acquired spacers revealed how historical contingency between genetic perturbations could constrain adaptive evolution of increasing antibiotic resistance. For instance, the results provided strong genetic evidence that inactivation of the qox operon unlocked multiple contingent adaptive paths toward acquiring higher gentamicin resistance, such as inactivation of cydABCD and atoB (Figure 17B). The ease with which this strategy can reveal epistasis and historical contingency should facilitate rapid determination of genetic architecture for diverse bacterial phenotypes. This foundation will be critical for understanding adaptive evolutionary trajectories, engineering of synthetic bacteria for industrial and therapeutic applications, and developing rational antimicrobial strategies that are refractory to resistance. Materials and Methods Microbial strains and culture conditions [0158] Cultivation of Staphylococcus aureus strains RN4220 (Nair et al., 2011), Newman (Bae et al 2006) TB4 (Bae et al 2006) and MW2 (Baba et al 2002) were carried out in tryptic soy broth (TSB) medium (BD) at 37 °C with shaking (rpm 220). Whenever necessary, tryptic soy agar (TSA) was supplemented with appropriate antibiotics to select for plasmid transformation. In these cases, the concentrations of antibiotics were as follows: chloramphenicol, 10 µg/mL; erythromycin, 10 µg/mL; tetracycline, 5 µg/mL. [0159] Cultivation of Escherichia coli MG1655 was carried out in Luria-Bertani (LB) medium (BD) at 37 °C with shaking (rpm 220). Whenever necessary, Luria-Bertani agar was supplemented with chloramphenicol (25 µg/mL) to select for plasmid transformation. Preparation of electrocompetent S. aureus cells [0160] Preparation of S. aureus competent cells and DNA transformation was performed as previously described (Goldberg et al., 2014). Measurement of MICs _{[0161] A single colony (10} ⁶ _-10 ⁷ _{CFUs) of S. aureus RN4220 cells harboring single- or dual-} spacers was suspended in 100 µL of TSB and 2.5 uL of the suspended cells were spotted on TSA containing a 2- old dilution series of gentamicin (0.1 – 6.4 µg/mL). Cells were allowed to grow for 16 h at 37 °C and he MIC was determined as the minimum antibiotic concentration where no bacterial growth was seen. All measurements were performed in triplicates. Measurement of bacterial growth [0162] Overnight cultures of S. aureus RN4220 cells harboring single- or dual-spacers were diluted 1:200 in 200 µL of fresh TSB medium supplemented with gentamicin (1 or 4 µg/mL) in 96- well flat bottom plates (Costar). Chloramphenicol (5 µg/mL) was always added to maintain plasmid pTHR (carrying racr, hdcas9 and the spacer). [0163] To check the effect of mevalonate and FAS inhibitors on gentamicin, overnight cultures of S. aureus RN4220 cells harboring pTHR (i.e., pWJ402) were diluted 1:200 in 200 µL of fresh TSB medium supplemented with various concentrations of gentamicin, mevalonate inhibitor (vanadyl sulfate) and FAS inhibitors (5-(tetradecyloxy)-2-furoic acid and cerulenin). [0164] In all cases, cells were grown in a Biotek Synergy MX plate reader shaking continuously for 18 h at 37°C. The absorbance at 600nm (OD₆₀₀) was measured every 10 minutes. Extraction and sonication of genomic DNA [0165] 10 – 30 mL of S. aureus or E. coli or culture grown to saturation were pelleted and washed with 1 volume of TE buffer (pH 8.0). Pellets were re-suspended in ~3 mL of ice-cold TE buffer (pH 8.0). For east cells, the buffer was supplemented with 0.5% Triton X-100. Every 500 µL of re-suspended cells was mixed with 500 µL of ice-cold Phenol/Chloroform/Isoamyl alcohol (25:24:1) (Fisher Scientific). [0166] For bacterial cells, the mixture was transferred into a 2 mL microtubes pre-filled with _{~0.25 cm} ³ _{of glass beads (0.1 mm) on ice. Cells were disrupted using FastPrep-245G™} Homogenizer (MPBio) at 4°C. The default S. aureus or E. coli setting was used. For yeast cells, the _{mixture was transferred into a 2 mL microtubes pre-filled with ~0.25 cm} ³ _{of glass beads (425 – 600} _{µm) on ice. Cells were disrupted using FastPrep-245G™ Homogenizer (MPBio) at 4°C. The} following customized setting was used. Speed: 6.0 m/sec; Adapter: QuickPrep; Time: 30 sec; Lysing Matrix: C; Quantity: 1 mg; Cycles: 9; Pause time: 240 sec. [0167] The homogenized mixture was centrifuged at 16,000 rcf for 10 min at room temperature. The aqueous phase was collected and mixed with 500 μL of chloroform and centrifuged as above. The aqueous phase was collected again, mixed with 1 mL of isopropanol, gently inverted several times, incubated for 10 min at room temperature and centrifuged. Precipitated genomic DNA was washed with 1 mL of 75% ethanol, air dried and dissolved in 50-300 μL of water. [0168] Genomic DNA was sonicated in 130 μL total volume in microTUBE AFA Fibre Pre-Slit Snap-Cap 6 × 16 mm tubes (Covaris) using the Covaris S220 Focused-ultrasonicator to a fragment size of 150 bp. The sonicated DNA was dialysed before electroporation. Cloning [0169] Plasmid pWJ402 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pRH154 and primers JW756 and W1301. Another PCR was performed using plasmid pGG32 and primers W1302 and JW755. [0170] Plasmid pWJ406 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pDB114 and primers W1309 and W1310. Another PCR was performed using plasmid pT181 and primers W1311 and W1312. [0171] Plasmid pWJ411 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pWJ402 and primers W1523 and W1778. Another PCR was performed using plasmid pJW105 and primers W1777 and W1524. [0172] Plasmid pWJ418 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pAV112B and primers W1339 and W1340. Another PCR was performed using plasmid pLM9 and primers W1341 and W1342. [0173] Plasmid pWJ420 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pE194 and primers W795 and W1011. Another PCR was performed using plasmid pWJ418 and primers W1363 and W1364. [0174] Plasmid pWJ424 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pWJ402 and primers W852 and W1374. Another PCR was performed using plasmid pDB182 and primers W1373 and W614. [0175] Plasmid pWJ444 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pWJ424 and primers W852 and W1540. Another PCR was performed using plasmid pWJ402 and primers W1541 and W614. [0176] Plasmid pWJ445 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pWJ40 (Goldberg et al., 2014) and primers W1581 and W1582. Another PCR was performed using plasmid pdcas9-bacteria (Qi et al., 2013) and primers W1583 and W1584. [0177] Plasmid pWJ450 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pJW104 and primers W1789 and W1792. Another PCR was performed using plasmid pWJ104 and primers W1790 and W1791. [0178] Plasmid pWJ571 was constructed by Gibson assembly of two PCR products. One PCR was performed using plasmid pZE21-MCS-1 (Lutz and Bujard, 1997) and primers W1578 and W1668. Another PCR was performed using plasmid pWJ402 and primers W1879 and W1880. [0179] Similar to a previous study (Bikard et al., 2013), spacer cloning was performed by ligation of annealed oligonucleotide pairs and BsaI-digested parent vector, pWJ444 or pWJ406. The spacer sequences are shown in Figure 45. _{[0180] All PCRs and Gibson assembly reactions were performed using Q5® High-Fidelity DNA} _{Polymerase and NEBuilder® HiFi DNA Assembly Master Mix supplied by NEB, respectively.} Cloning used S. aureus RN4220 or E. coli MG1655 electrocompetent cells. Name and content of all plasmids and shown in Figure 46. Name and sequence of all oligonucleotides are shown in Figure 44. Generation of single-spacer (1S) libraries by CRISPR-Cas adaptation in S. aureus with hdcas9 [0181] As explained in the main text, the terms “crRNA” and “spacer” are used interchangeably throughout. Generation of crRNA libraries by CRISPR-Cas adaptation [0182] The crRNA libraries generated in the following protocol are intended to be used directly in S. aureus. First, a single colony of S. aureus RN4220 or wild-type strains harboring the chloramphenicol-resistant pTHR (i.e., plasmid pWJ402, which carries tracr, hdcas9 and an empty CRISPR array) and he tetracycline-resistant pCCC (i.e., pWJ418, which carries cas1, cas2 and csn2 under an IPTG-inducible promoter, pSpac) was grown overnight in 4 mL of TSB with chloramphenicol (5 µg/mL) and tetracycline (2.5 µg/mL). Culture was diluted 1:200 in 15 mL of fresh TSB (no antibiotics) with 2 mM PTG to induce the expression of Cas1, Cas2 and Csn2 and ^{grown until OD} ₆₀₀ ^{reached 1.0 (typically 3 – 4 h). To make competent cells, cells were pelleted and} washed two times using one volume of sterile water at room temperature. (Cells prepared at 4 °C contained libraries with considerably more pacers matching the two helper plasmids.) Cells were _{ultimately re-suspended in 1/100} ^th _{volume of sterile water.} [0183] 50 µL of competent cells were mixed with 20 µg (usually in 10 µL) of sheared genomic DNA prepared from the same host and incubated 5 min at room temperature. Electroporation was performed using MicroPulser (Bio-Rad) with the default staph program (2 mm, 1.8 kV and 2.5 ms). After electroporation, cells were immediately re-suspended in 500 µL of TSB and recovered at 37 °C for 15 min with shaking. Next, 200 µL of recovered cells were transferred to 15 mL of pre-warmed TSB with chloramphenicol (5 µg/mL) and recovered for an additional 5 h at 37 °C with shaking. CRISPR adaptation happened during this recovery period. Recovery less than 5 h was not tested but may have still worked. [0184] Essentially, a crRNA library was made after the 5-hour CRISPR adaptation/recovery _{phase. At this point, this library culture was typically at OD600 = 3.0 or 10} ⁹ _{CFU/mL. As estimated} by enrichment PCR assays, 0.1 – 1% of cells had adapted one spacer. Typically, 15 mL of the library culture was pelleted, lysed, amplified by PCR and sent for Illumina sequencing while 6 mL of library culture was subjected to experimental procedures (e.g., gentamicin treatment). Exposure of single-spacer (1S) libraries to gentamicin [0185] S. aureus with a 1S library generated with a hyperactive CRISPR adaptation machinery carrying hdCas9 were subjected to treatment of low or high concentration of gentamicin (1.0 µg/mL or 4.0 µg/mL, respectively). [0186] For long-term exposure to the antibiotics, 6 mL of library (i.e., cells recovered after CRISPR adaptation) was transferred to 500 mL of TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) and appropriate concentration of gentamicin and grown for 18 h at 37 °C, 220 rpm. After selection, 12 mL of culture were pelleted, lysed, amplified by PCR and sent for Illumina sequencing. Selection in both low and high concentration of gentamicin was performed three times. [0187] For short-term exposure to the antibiotics, 20 mL of library was transferred to 500 mL of TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) and gentamicin (1 µg/mL) and grown for 4.5 h at 37 °C, 220 rpm. After selection, 400 mL of culture were pelleted, lysed, amplified by PCR and sent for Illumina sequencing. As a control, 20 mL of library was also transferred to 500 mL TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) but no gentamicin (plain control) and grown for 4.5 h at 37 °C, 220 rpm. After growth, 200 mL of culture were pelleted, lysed, amplified by PCR and sent for llumina sequencing. Illumina sequencing [0188] S. aureus crRNA libraries (15 mL) or cells after gentamicin treatment (12 mL) were pelleted and re-suspended in 5 mL of P1 buffer (Qiagen) supplemented with lysostaphin (1 µg/mL). Mixture was incubated for 1 h at 37 °C. Next, Miniprep was performed following the Qiagen protocol. To note, one column was used per sample and 70 – 120 µL of H₂O was used to elute plasmid DNA. The final concentration of DNA ranged from 100 to 400 ng/uL. [0189] As CRISPR adaptation is of low frequency, to detect adaptation, enrichment PCR was performed as described in Modell et al., 2017 with slight modifications. For each sample, a 60 µL reaction mix was prepared by adding 100 ng of plasmid DNA as template, 0.5 uM of forward primer W1201), 0.5 uM of reverse enrichment primers (equimolar mixture of W1202, W1203, W1204) and _{Q5® High-Fidelity DNA polymerase (NEB). PCR was performed using a C1000™ Thermal Cycler} Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 18-29 cycles (depending on ate of CRISPR adaptation): 98 °C for 10 s, 61 °C for 20 s and 72 °C for 30 s; final extension: 72 °C or 2 min. PCR products were either visualized on a 2% agarose gel or purified with AMPure XP beads (Beckman Coulter). The amount of the beads used was adjusted to maximize the removal of the smaller non-adapted amplicons and retention of the larger adapted ones. Beads were eluted in 1 volume of H₂O. [0190] To prepare samples for sequencing, a second PCR was performed to introduce the Illumina adapter sequences to the purified amplicons from the previous PCR. A 150 µL reaction mix was prepared by adding 3 µL of the purified amplicons as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB). One forward primer was chosen from the following: W1407, W1409, W1410, W1411, W1417, W1418 and W1419, all containing the Illumina universal adapter sequences and various customized internal barcodes. One reverse primer was chosen from the following: W1408 and W1426, both containing the Illumina adapter and index sequences. All these primers were PAGE purified (IDT). PCR was performed _{using a C1000™ Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C} for 30 s; 6 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 20 s; final extension: 72 °C for 2 min. PCR products were purified with AMPure XP beads (Beckman Coulter) two times. The amount of the beads used was adjusted to maximize the removal of non-adapted amplicons and retention of adapted ones. Purified amplicons were subjected to the Illumina NextSeq platform. Generation of single-spacer (1S) libraries by CRISPR-Cas adaptation in S. aureus with hCas9 Generation of crRNA libraries by CRISPR-Cas adaptation [0191] As hdCas9 was replaced by hCas9 to avoid CRISPR adaptation from self-DNA, the crRNA libraries generated in the following protocol are intended to be sub-cloned to other organisms (e.g., E. coli). First, a single colony of S. aureus RN4220 harboring the chloramphenicol-resistant pTHR (i.e., plasmid pWJ411, which carries tracr, hcas9 and an empty CRISPR array) and the tetracycline-or erythromycin-resistant pCCC (i.e., pWJ418 or pWJ420, respectively. Both plasmids carry cas1, cas2 and csn2 under an IPTG-inducible promoter, pSpac.) was grown overnight in 4 mL of TSB with chloramphenicol (5 µg/mL) and tetracycline (2.5 µg/mL) or erythromycin (5 µg/mL). Culture was diluted 1:200 in 15 mL of fresh TSB (no antibiotics) with 2 mM IPTG to induce the expression of Cas1, Cas2 and Csn2 and grown until OD₆₀₀ reached 1.0 (typically 3 – 4 h). To make competent cells, cells were pelleted and washed two times using one volume of sterile water at room temperature. (Cells prepared at 4 °C contained libraries with considerably more spacers matching the two helper plasmids.) Cells were ultimately re-suspended in 1/100^th volume of sterile water. [0192] 50 µL of competent cells were mixed with 20 µg (usually in 10 µL) of sheared genomic DNA prepared rom organisms of interest (e.g., E. coli) and incubated 5 min at room temperature. Electroporation was performed using MicroPulser (Bio-Rad) with the default staph program (2 mm, 1.8 kV and 2.5 ms). After electroporation, cells were immediately re-suspended in 500 µL of TSB and recovered at 37 °C for 15 min with shaking. Next, 200 µL of recovered cells were transferred to 15 mL of pre- warmed TSB with chloramphenicol (5 µg/mL) and recovered for an additional 5 h at _{37 °C with shaking. At this point, the culture was typically at OD600 = 3.0 or 10} ⁹ _{CFU/mL. To further} eliminate undesired crRNAs targeting S. aureus or the helper plasmids, 7 mL of the culture was transferred to 500 mL TSB with chloramphenicol (5 µg/mL) and tetracycline (2.5 µg/mL) or erythromycin (5 µg/mL), and grown for 10 h. Typically, 15 mL of this library culture was pelleted, lysed and minipreped Qiagen), generating M2141 (if pCCC is pWJ418) and M2144 (if pCCC is pWJ420). These minipreped DNAs were later PCR amplified for sub-cloning (see below). [0193] Notice that two different pCCCs were constructed herein, pWJ418 and pWJ420, each with a different origin of replication and antibiotic resistant marker. crRNA libraries generated with pWJ418 had larger percentage of spacers matching the genome (pWJ418: 90% vs pWJ420: 83%, Figure 22B), while libraries generated with pWJ420 had greater spacer diversity pWJ418: 386,032 unique spacers vs pWJ420: 424,473 unique spacers). As the growth rate of cells harboring either plasmid was indistinguishable from each other, mixing these two cell types at a 1:1 ratio and preparing competent cells together is recommended in order to generate more diverse crRNA libraries. Sub-cloning [0194] To sub-clone this library to the organism of interest (e.g., E. coli), enrichment primers were necessary as only 0.1 – 1% of cells contained an adapted spacer after CRISPR adaptation. Sub- cloning was done by Gibson assembly of an insert PCR and a backbone PCR. [0195] For the insert PCR, every 50 µL reaction mix was prepared by adding 375 ng of equimolar mixture of M2141 and M2144 as template, 0.5 uM of forward enrichment primers (equimolar mixture of W1397, W1398 and W1399), 0.5 uM of reverse primer (W1887) and Q5® _{High-Fidelity DNA polymerase NEB). PCR was performed using a C1000™ Thermal Cycler (Bio-} Rad) with the following settings. initial denaturation: 98 °C for 30 s; 26 cycles: 98 °C for 10 s, 65 °C for 20 s and 72 °C for 30 s; final extension: 72 °C for 2 min. PCR products were subjected to a 2-step purification with AMPure XP in order to remove the large-sized plasmid template and the small- sized non-adapted amplicons. Briefly, PCR products were mixed with 0.5X volume of beads, let settled and supernatant was transferred to 0.3X volume of beads and proceeded with purification. Beads were eluted in H₂O, generating C2185. [0196] For the backbone PCR, every 50 µL reaction mix was prepared by adding 0.4 ng of plasmid pWJ571 as template, 0.2 uM of forward primer (W1889), 0.2 uM of reverse primers (W1891) and OneTaq® DNA polymerase (NEB). PCR was performed with the following settings. Initial denaturation: 94 °C for 30 s; 30 cycles: 94 °C for 20 s, 55 °C for 20 s and 68 °C for 2 min; final extension: 68 °C for 5 min. PCR products were purified with QIAquick PCR purification kit (Qiagen), generating C2184. [0197] 10 µL of Gibson reaction was performed by mixing 100 ng of insert PCR (in 2 µL), 800 ng of backbone PCR (in 3 µL) and NEBuilder HiFi DNA Assembly Master Mix (NEB) and incubating 30 min at 50 °C. The insert PCR was at ~1.7X excess. Ligated products were dialyzed, and 5 µL of it was transformed into electro-competent E. coli MG1655 cells harboring pWJ445 (plasmid with inducible dCas9). Electroporation was performed using MicroPulser (Bio-Rad) with the default E. coli program 11 mm, 1.8 kV and 6.1 ms) and cells were recovered in 500 µL of SOC medium for 1.5 h at 37 °C. Transformation efficiency was routinely at 5% ± 2% of the population. [0198] To select for transformants, recovered cells (500 µL) were transferred to 250 mL LB with _{kanamycin 50 µg/mL) and grown for 4.5 h at 37 °C (OD600 reached ~ 0.3, equivalent to ~10} ⁸ CFU/mL). This extent of selection was sufficient as ~100% of cells were kanamycin-resistant at this point. Illumina sequencing [0199] 100 mL of cells selected in liquid medium containing kanamycin were pelleted and minipreped Qiagen). Since sub-cloning was done and the majority of plasmids extracted from cells contained a pacer, enrichment PCR was not necessary. For each sample, a 60 µL reaction mix was prepared by adding 100 ng of plasmid DNA as template, 0.5 uM of forward primers (equimolar mixture of W1397, W1398, W1399 and W1400), 0.5 uM of reverse primer (W1699) and Q5® High- _{Fidelity DNA polymerase (NEB). PCR was performed using a C1000™ Thermal Cycler (Bio-Rad)} with the following settings. Initial denaturation: 98 °C for 30 s; 12 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 30; final extension: 72 °C for 2 min. PCR products were either visualized on a 2% agarose gel or purified with AMPure XP beads (Beckman Coulter). The amount of the beads used was adjusted to maximize the removal of the smaller non-adapted amplicons and retention of the larger adapted ones. Beads were eluted in 1 volume of H₂O. [0200] To prepare samples for sequencing, a second PCR was performed to introduce the Illumina adapter sequences to the purified amplicons from the previous PCR. A 150 µL reaction mix was prepared by adding 3 µL of the purified amplicons as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB). The forward primer was either W1434 or W1435, both containing the Illumina universal adapter sequences and customized internal barcodes. The reverse primer was W1427, which contained the Illumina adapter and index _{sequences. All these primers were PAGE purified (IDT). PCR was performed using a C1000™} Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 6 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 30 s; final extension: 72 °C for 2 min. PCR products were purified with AMPure XP beads (Beckman Coulter) two times. The amount of the beads used was adjusted to maximize the removal of the smaller non-adapted amplicons and retention of the larger adapted ones. Purified amplicons were subjected to the Illumina NextSeq platform. Generation of single-spacer (1S) libraries by CRISPR-Cas adaptation in E. coli with dCas9 [0201] A single colony of E. coli MG1655 harboring plasmid pWJ450 (which carries tracr, dcas9, cas1, cas2, sn2 and an empty CRISPR array) was grown overnight in 4 mL of LB with chloramphenicol (25 µg/mL). Culture was diluted 1:200 in 15 mL of fresh LB (no antibiotics) and ^{grown until OD} ₆₀₀ ^{reached 0.5 – 0.6. Cells were pelleted and washed two times using one volume} of sterile water at room temperature. Cells were ultimately re-suspended in 1/200^th volume of sterile water. [0202] 50 µL of cells were mixed with 20 µg (usually in 10 µL) of sheared genomic DNA prepared from MG1655 Electroporation was performed using MicroPulser (Bio-Rad) with the default E. coli program 1 (1 mm, 1.8 kV and 6.1 ms). After electroporation, cells were immediately re-suspended in 500 µL of LB and recovered at 37 °C for 15 min with shaking. Next, 200 µL of recovered cells were transferred o 15 mL of pre-warmed LB with chloramphenicol (12.5 µg/mL) and recovered for an additional 9 h at 37 °C with shaking. Generation of “one-vs-all” libraries by CRISPR-Cas adaptation in S. aureus with hdCas9 Generation of “one-vs-all” by CRISPR-Cas adaptation [0203] Two spacers targeting qoxA and ndh were cloned into pTHR, generating plasmids pWJ451 and pWJ455, respectively. Both plasmids carried tracr, hdcas9 and the respective spacer targeting the gene of interest. [0204] A single colony of S. aureus RN4220 harboring the chloramphenicol-resistant pTHR (e.g., pWJ451, targeting qoxA) and the tetracycline-resistant pCCC (i.e., pWJ418, which carries cas1, cas2 and csn2 under an IPTG-inducible promoter, pSpac) was grown overnight in 4 mL of TSB with chloramphenicol 5 µg/mL) and tetracycline (2.5 µg/mL). Culture was diluted 1:200 in 15 mL of fresh TSB supplemented with gentamicin (0.5 µg/mL) and 2 mM IPTG to induce the ^{expression of Cas1, Cas2 and Csn2 and grown until OD} ₆₀₀ ^{reached 1. Cells were pelleted and} washed two times using one volume of sterile water at room temperature. Cells were ultimately re- suspended in 1/100^th volume of sterile water. [0205] 50 µL of competent cells were mixed with 20 µg (usually in 10 µL) of sheared genomic DNA prepared rom RN4220 and incubated 5 min at room temperature. Electroporation was performed using MicroPulser (Bio-Rad) with the default staph program (2 mm, 1.8 kV and 2.5 ms). After electroporation, cells were immediately re-suspended in 500 µL of TSB and recovered at 37 °C for 15 min with shaking. Next, 200 µL of recovered cells were transferred to 15 mL of pre-warmed TSB with chloramphenicol (5 µg/mL) and gentamicin (1 µg/mL) and recovered for an additional 5-7 h at 37 °C with shaking. Exposure of “one-vs-all” libraries to high doses of gentamicin _{[0206] After recovery, the “one-vs-all” library culture was typically at OD600 = 3.0 or 10} ⁹ CFU/mL.15 mL of the library culture was pelleted, lysed, amplified by PCR and sent for Illumina sequencing.6 mL of library culture was transferred to 500 mL of TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) and high dose of gentamicin (4 µg/mL) and grown for 18 h at 37 °C, 220 rpm. After selection, 12 mL of culture were pelleted, lysed, amplified by PCR and sent for Illumina sequencing. Illumina sequencing [0207] As CRISPR adaptation is of low frequency, to detect adaptation, enrichment PCR was performed as described on Modell et al., 2017 with slight modifications. For each sample, a 60 µL reaction mix was prepared by adding 100 ng of plasmid DNA as template, 0.5 uM of forward primer, 0.5 uM of reverse primers and Q5® High-Fidelity DNA polymerase (NEB). The forward enrichment primers were an equimolar mixture of W1397, W1398 and W1399. The reverse primer was W1542 _{for “qoxA-vs-all”) and W1688 (for “ndh-vs-all”), respectively. PCR was performed using a C1000™} Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 18-29 cycles depending on rate of CRISPR adaptation): 98 °C for 10 s, 61 °C for 20 s and 72 °C for 30 s; final extension: 72 °C for 2 min. PCR products were either visualized on a 2% agarose gel or purified with AMPure XP beads (Beckman Coulter). The amount of the beads used was adjusted to maximize the removal of non-adapted amplicons and retention of adapted ones. Beads were eluted in 1 volume of H₂O. [0208] To prepare samples for sequencing, a second PCR was performed to introduce the Illumina adapter sequences to the purified amplicons from the previous PCR. A 150 µL reaction mix was prepared by adding 3 µL of the purified amplicons as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB). One forward primer was chosen from the following: W1412, W1420, W1421, W1422, W1423 and W1424, all containing the Illumina universal adapter sequences and various customized internal barcodes. One reverse primer was chosen from the following: W1425, W1427 and W1428, all containing the Illumina adapter and _{index sequences. All these primers were PAGE purified (IDT). PCR was performed using a C1000™} Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 6 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 20 s; final extension: 72 °C for 2 min. PCR products were purified with AMPure XP beads (Beckman Coulter) two times. The amount of the beads used was adjusted to maximize the removal of non-adapted amplicons and retention of adapted ones. Purified amplicons were subjected to the Illumina NextSeq platform. Generation of dual-spacer (2S) libraries in S. aureus with hdCas9 Generation of a 1S library [0209] Generation of the dual-spacer (2S) library is schematized in Figure 4A. First, a single- spacer (1S) library was generated by CRISPR-Cas adaptation in S. aureus RN4220 with hdCas9 as described above. After CRISPR adaptation and recovery, 20 mL of this library culture was transferred to 500 mL of TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) and gentamicin (1 µg/mL) and grown or 9 h at 37 °C, 220 rpm. After this mild selection, plasmid DNA was prepared from 30 mL of culture generating M1906 which represented the 1S library. Sub-cloning [0210] After CRISPR adaptation, since only 0.1 – 1% of cells contained an adapted spacer, sub- cloning of the 1S library using enrichment primers was necessary in order to substantially increase DNA with adapted spacers in the library. Sub-cloning was done by Gibson assembly of an insert PCR and a backbone PCR. [0211] For the insert PCR, a 140 µL reaction mix was prepared by adding 800 ng of M1906 as template, 0.5 uM of forward primer (W1521), 0.5 uM of reverse enrichment primers (equimolar mixture of W1202, W1203, W1204) and Q5® High-Fidelity DNA polymerase (NEB). PCR was _{performed using a C1000™ Thermal Cycler (Bio-Rad) with the following settings. Initial} denaturation: 98 °C for 30 s; 40 cycles: 98 °C for 10 s, 65 °C for 20 s and 72 °C for 30 s; final extension: 72 °C for 2 min. PCR products were subjected to a 2-step purification with AMPure XP in order to remove the large-sized plasmid template and the small-sized non-adapted amplicons. Briefly, PCR products were mixed with 0.8X volume of beads, let settled and supernatant was transferred to 0.3X volume of beads and proceeded with purification. Beads were eluted in H₂O, generating C1976. [0212] For the backbone PCR, a 50 µL reaction mix was prepared by adding 10 ng of plasmid pWJ402 as template, 0.5 uM of forward primer (W1522), 0.5 uM of reverse primer (W1525) and _{Q5® High-Fidelity DNA polymerase (NEB). PCR was performed using a C1000™ Thermal Cycler} (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 30 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 7 min; final extension: 72 °C for 15 min. PCR products were purified with QIAquick PCR purification kit (Qiagen), generating C1542. [0213] 10 µL of Gibson reaction was performed by mixing 70 ng of insert PCR, ~1 µg of backbone PCR and NEBuilder HiFi DNA Assembly Master Mix (NEB) and incubating 30 min at 50 °C. The insert PCR was at ~2.5X excess. Ligated products were transformed into electro-competent S. aureus RN4220 cells. [0214] Transformants were selected on TSA with chloramphenicol. After 16 h incubation at 37 °C, a total of 20,000 CFU (in two plates) were obtained (5,000 – 10,000 CFU for every 10 µL Gibson reaction). These colonies were scraped using an L-shaped spreader and suspended in 1X PBS, lysed with lysostaphin and minipreped, generating the sub-cloned 1S library (sub1-1S). As estimated by PCR, ~70% of the sub1-1G library contained an adapted spacer while the rest had an empty array. [0215] The sub1-1S library was concentrated by evaporation and transformed into electro- competent S. aureus RN4220 cells harboring pCCC (i.e., pWJ418). Transformed cells were selected on TSA with chloramphenicol and tetracycline. After 16 h incubation at 37 °C, a total of 63,000 CFU (in three plates) were obtained. These colonies were scraped using an L-shaped spreader and suspended in 45 mL of 1X PBS.10 mL of the suspended cells were lysed with lysostaphin and minipreped, generating M1983, which represented the sub-cloned 1S library immediately before the second round of CRISPR adaptation (sub2-1S). Generation of a 2S library _{[0216] For the second round of CRISPR adaptation, 450 µL of suspended cells (1.8 x 10} ⁹ _CFU) were transferred to 90 mL of TSB with 2 mM IPTG and gentamicin (0.5 µg/mL) and grown until OD₆₀₀ reached 1 (~3.5 h). Cells were washed and electroporated with genomic DNA prepared from RN4220 as described above.200 µL of electroporated cells were re-suspended in 2 mL of TSB (scaled-up our times) and recovered at 37 °C for 15 min with shaking. Next, 800 µL of recovered cells were transferred to 60 mL of pre-warmed TSB with chloramphenicol (5 µg/mL) and gentamicin (0.5 µg/mL) and recovered for an additional 6.5 h at 37 °C with shaking. _{[0217] After recovery, this library culture was at 2 x 10} ⁹ _{CFU/mL. 30 mL of the culture was} pelleted, lysed and minipreped, generating M1984 which represented the 2S library. On the other hand, 10 mL x 3 of the 2S library culture were transferred to three different Erlenmeyer flasks each containing 500 mL of TSB (pre-warmed at 37 °C) with chloramphenicol (5 µg/mL) and gentamicin (4 µg/mL), respectively. After 18 h growth at 37 °C, 220 rpm, 25 mL of cells from each flask were pelleted, lysed and minipreped, generating M1985, M1986 and M1987, each representing the 2S library post-selection in high concentration of gentamicin (4 µg/mL). Illumina sequencing [0218] Three types of libraries were being prepared for sequencing: sub2-1S (i.e., M1983), pre- selected 2S i.e., M1984) and post-selected 2S libraries (i.e., M1985, M1986 and M1987). These library cultures were pelleted and re-suspended in P1 buffer (Qiagen) supplemented with lysostaphin (1 µg/mL). Mixture was incubated for 1 h at 37 °C. Next, Miniprep was performed following the Qiagen protocol. To note, one column was used per sample and 70 – 120 µL of H₂O was used to elute plasmid DNA. The final concentration of DNA ranged from 500 to 900 ng/uL. Sub2-1S library [0219] To prepare the sub2-1S library (i.e., M1983) for sequencing, a 50 µL reaction mix was prepared by adding 50 ng of plasmid DNA as template, 0.5 uM of forward primer (W1782), 0.5 uM of reverse primer (M1783) and Q5® High-Fidelity DNA polymerase (NEB). PCR was performed using a C1000™ Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 22 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 20 s; final extension: 72 °C for 2 min. PCR products were purified with QIAquick PCR purification kit (Qiagen) and eluted in 50 µL of H₂O. [0220] To prepare samples for sequencing, a second PCR was performed to introduce the full Illumina adapter sequences to the purified amplicons from the previous PCR. A 250 µL reaction mix was prepared by adding 1 µL of the purified amplicons as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB). The forward primer was P067, which contained a portion of the Illumina universal adapter sequence. The reverse primer was P072, which contained the Illumina adapter and index sequences. PCR was performed using a C1000™ Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 5 cycles: 98 °C for 10, 55 °C for 20 s and 72 °C for 20 s; final extension: 72 °C for 2 min. PCR products were purified with AMPure XP beads (Beckman Coulter). The amount of the beads used was adjusted to maximize the removal of non-adapted amplicons and retention of adapted ones. Purified amplicons were subjected to the Illumina NextSeq platform (150 cycles). Pre- and post-selected 2S libraries [0221] To prepare the 2S libraries (i.e., M1984, M1985, M1986 and M1987) for sequencing, a 50 µL reaction mix was prepared by adding 500 ng of plasmid DNA as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB) for each of these libraries. One forward primer was chosen from the following: W1779, W1780, W1781 and W1782, all containing a portion of llumina universal adapter sequences and various customized internal barcodes. The reverse primer was W1783, which contained a portion of an Illumina adapter sequence. PCR was performed using a C1000™ Thermal Cycler (Bio-Rad) with the following settings. Initial denaturation: 98 °C for 30 s; 25 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 35 s; final extension: 72 °C for 2 min. PCR products were subjected to electrophoresis on a 2% agarose gel. Bands corresponding to amplicons with two spacers were excised and purified using an Ultrafree-DA Centrifugal Filter Unit (Millipore) and concentrated in 10 µL of H₂O using DNA Clean & Concentrator (DCC-5, Zymo Research). [0222] To prepare samples for sequencing, a second PCR was performed to introduce the full Illumina adapter sequences to the purified amplicons from the previous PCR. A 250 µL reaction mix was prepared by adding 1.5 µL of the purified amplicons as template, 0.5 uM of forward primer, 0.5 uM of reverse primer and Q5® High-Fidelity DNA polymerase (NEB). The forward primer was P067, which contained a portion of the Illumina universal adapter sequence. The reverse primer was either P071 or P072, both containing the Illumina adapter and index sequences. All these primers _{were PAGE purified (IDT). PCR was performed using a C1000™ Thermal Cycler (Bio-Rad) with} the following settings. Initial denaturation: 98 °C for 30 s; 5 cycles: 98 °C for 10 s, 55 °C for 20 s and 72 °C for 40 s; final extension: 72 °C for 2 min. PCR products were purified with AMPure XP beads (Beckman Coulter) two times. The amount of the beads used was adjusted to maximize the removal of non- adapted amplicons (i.e., amplicons with one or no spacer) and retention of adapted ones (i.e., amplicons with two spacers). Purified amplicons were subjected to the Illumina MiSeq platform (300 cycles). Data analysis of single-spacer (1S) and “one-vs-all” libraries Spacer identification and sequence alignment [0223] First, for any genome, a list of all functional spacers (i.e., spacers with an “NGG” PAM) was created. Each spacer was uniquely identified by its strandedness and location in the genome. For instance, here are 136,928 unique functional spacers (i.e., occurrence of “NGG” on both strands of the DNA) n the S. aureus RN4220 genome. [0224] From the FASTQ file generated by Illumina sequencing (NextSeq, 75 cycles), spacer sequences were identified by locating the two direct repeats flanking them. Spacers were aligned to chromosomal and plasmid genomes using “aln” and “samse” functions of the Burrows-Wheeler Alignment tool (Li and Durbin, 2009). The NCBI Staphylococcus aureus subsp. aureus NCTC8325 chromosome (NC_007795.1) and Escherichia coli str. K-12 substr. MG1655 (NC_000913.3) were used as reference genomes. [0225] From the output SAM file, only functional spacers (i.e., spacers with an “NGG” PAM) were considered or downstream analysis. The number of times each unique functional spacer appeared was recorded. The frequency of each spacer was calculated by dividing its number by the total number of chromosomal spacers. Genome-wide spacer distribution at a gene level [0226] To assess the genome-wide spacer distribution at a gene level, the number of reads of all spacers matching each gene was summed and normalized to the gene length, getting R_{gene 1}, R_{gene 2} … R_{gene N}, where N is the number of genes in the genome. Then the coefficient of variation was calculated (CV_gene): [0227] where σ_gene and μ_gene ar

e the standard deviation and mean of all R_genes, respectively. For E. coli, gene lacI was excluded due to its abnormal enrichment caused by the presence of a helper plasmid (Figure 1D). Estimation of fitness effects using Z-score [0228] The terms “crRNA” and “spacer” are used interchangeably throughout. The fitness effect of each pacer under antibiotic exposure was determined by its enrichment/depletion relative to the initial unselected library using the Z-score (Girgis et al., 2007). In the current study, at least six single- pacer (1S) libraries were made independently via a single round of CRISPR adaptation. The frequency of each spacer was log₁₀ transformed (F) and its corresponding Z-score was calculated by:

[0229] where F_post is the log-transformed frequency of the spacer in the post-selected sample, and _{Fμ, pre} and σ_pre are the mean and the standard deviation of the log-transformed frequency of the spacer in the six pre-selected 1S libraries. ^{[0230] Next, all Z} _spacer ^{s were grouped based on the genes they target (The region within 100 bp} upstream from the start codon is also considered as part of the gene). To calculate the mean Z-score for each gene (Mean-Z), Z-scores of all individual spacers matching the gene were averaged after the removal of the lowest and highest Z-scores:

[0231] where spacers are ranked with their Z-scores from the highest to the lowest and N is the total number of spacers matching the gene. Removal of the lowest and highest Z-scores effectively eliminated pacers with off-target potentials (see below). [0232] However, as Mean-Z severely underestimated fitness effects due to the conflation of ineffective pacers, two other metrics were used: High_2/3-Z by averaging the second and third most enriched pacers (similar to work by Gilbert and colleagues who used the three strongest gRNAs (Gilbert et al., 2014)), or the 95^th percentile of the Z-score of all spacers (except the lowest and the highest) targeting each gene:

[0233] The two metrics showed good correlation (R = 0.92, Figure 24B) and shared 96% of the top 25 most enriched genes. P_95-Z was chosen to represent the fitness effect of genes, as it was less prone to outliers than High_2/3-Z while still effectively eliminated spacers with off-target potentials (see below). [0234] 1S libraries that were treated for 18 h with low (1.0 µg/mL) and high (4.0 µg/mL) concentration of gentamicin had three experimental replicates. Thus, for each gene, the mean of the Mean Z High Z and P Z from the triplicates were calculated and reported. [0235] For 1S libraries that were treated for 4.5 h with no or low concentration of gentamicin (1.0 µg/mL), crRNA perturbations were expected that truly potentiate the effect of antibiotics to be significantly depleted n gentamicin but not in plain medium. Here Mean-Z was chosen to represent the fitness effects, as the goal was to quantify the negative fitness effects (i.e., depletion of crRNAs) and neither High_2/3-Z nor P_95-Z was relevant. To do so, a filter was first applied so that only genes whose Mean-Z was greater than -0.2 in plain medium and less than 0 in gentamicin (1 µg/mL) were considered. Next, for each gene, Z-score of all individual crRNAs from the two conditions were ^{subjected to a Mann-Whitney U test in order to determine whether a significant portion of Z} _spacer ^s were lower in gentamicin. Consideration of spacers with off-target potentials ^{[0236] In all the aforementioned metrics (Mean-Z, High} _2/3 ^{-Z and P} ₉₅ ^{-Z), the lowest and highest} Z-scores for each gene were eliminated. Doing so effectively reduced the probability of including faulty spacers with potential off-target effect when quantifying each gene’s fitness effect. Indeed, when ranking genes by the highest Z-score (High-Z) targeting them, many top hits were found to be faulty, as they were targeted by only one spacer with high Z-score. These genes included SAOUHSC_01257, SAOUHSC_00628, SAOUHSC_03037, SAOUHSC_03016 and SAOUHSC_00266 (see Figure 31). Further examination revealed that these spacers had extensive seed-sequence homology (10-nt or more) and correct PAMs to non-neutral off-target sites within genes such as qoxA, qoxC and mnhA. A gene is defined as non-neutral if repression of it by CRISPRi increases cell’s fitness in gentamicin. [0237] In order to exclude off-target spacers, a filter was applied in which spacers with 10-nt or more of their seed sequence matching another genomic site with a correct PAM were excluded from analysis. As a result, ~30% spacers from the library were excluded by this filter, which effectively reduced the number of genes that were targeted by only one spacer with high Z-score. However, when comparing he P₉₅-Zs obtained with and without this off-target filter, 29 out of 30 top hits were found to be the same, suggesting that the off-target filter may not be necessary. [0238] Taken together, the off-target filter was not applied because: 1. The similarity between the results from these two pipelines indicates elimination of the highest Z-score for each gene alone effectively excluded spacers with off-target potentials; 2. Genes containing more than one spacer with off-target potential on other non-neutral genes are extremely rare. Plus, top enriched hits (e.g., top 30) identified with either pipeline contained multiple spacers with high Z-score (column “ranked Zs”), suggesting they were bona fide hits.3. The off-target filter excluded roughly 30% of spacers (the vast majority of these spacers have neutral off-targets) compromising the coverage and sensitivity of the 1S and 2S CRISPRi screens. For instance, the pipeline with off-target filter failed to identify the essential gene, mvaK2 as a hit after gentamicin (1 µg/mL) treatment. Estimation of fitness effects using Log₂FC (for 1S and “one-vs-all libraries) [0239] The fold-change between post-selected and pre-selected libraries was also used as a measure of fitness effect. Log₂ of the fold-change (log₂FC) of each spacer was calculated by:

[0240] where f_post and f_pre are the frequency of the spacer in the post-selected and pre-selected libraries, respectively. P₉₅-log₂FC was calculated similar to the Z-score. For each gene, the mean of the P₉₅-log₂FCs from the triplicates were reported, and a one-sample t-test (one-tailed) between the triplicates of P₉₅-log₂FCs and 0 (i.e., fold-change of 1) was performed. In cases where the P₉₅-log₂FC value was not present in triplicates post-selection, a t-test could not be performed and P-value was assigned “N/A”. [0241] The fold-change and Z-score calculations showed good correlation for enriched genes (R = 0.96, Figure 24C) and shared 92% (23/25) of the top enriched genes. However, using P₉₅-Z to P₉₅-log₂FC was preferred whenever possible as there were six pre-selected 1S libraries, making Z- score analysis more statistically robust. Estimating epistasis [0242] In the absence of a genetic interaction, the fitness of a double-mutant is expected to be the product of the individual fitness of the two single mutants (Phillips, 2008; Segre et al., 2005). Epistasis (ɛ) measures the deviation from this expectation:

[0243] where W_{X, Y} is the fitness of the double genetic perturbation, and W_X and W_Y are the fitness of the ingle genetic perturbations in genes X and Y, respectively. [0244] 10 representative strains were selected harboring single- and dual-spacers and measured their fitness by pairwise competition assays in gentamicin (See “Pairwise competition assay” below). As the measured fitness and log₂FC of individual spacers from deep sequencing exhibited a high correlation (Figure 37F), log₂FC was used as a proxy to estimate epistasis between qoxA and gene X:

[0245] where log₂FC_{qoxA, X} is the log₂FC of gene X measured in the “qoxA-vs-all” library, representing the combined fitness effect of repressing qoxA and gene X, log₂FC_qoxA and log₂FC_X are the log₂FCs measured in the 1S library, representing the fitness effect of repressing either qoxA or gene X, respectively. All measurements were performed in gentamicin (4.0 µg/mL). Pairwise competition assay [0246] S. aureus RN4220 cells harboring no spacer (pWJ402, representing wild-type) and single spacers targeting qoxA, ndh, mvaD, atoB, cydA and SAOUHSC_01269 were constructed. All these spacers were cloned into the pWJ402 backbone with a chloramphenicol-resistant (cm^R) marker. Cells harboring these plasmids were also co-transformed with a tetracycline-resistant (tet^R) plasmid harboring an empty CRISPR array. [0247] S. aureus RN4220 cells harboring dual-spacers were generated by co-transforming the _cm ^R _{pWJ451 qoxA-targeting) and a tet} ^R _{plasmid harboring a spacer that targets one of the following} genes: ndh, mvaD, atoB, cydA and SAOUHSC_01269. [0248] All strains harboring single-spacers and dual-spacers were competed with an RN4220 strain harboring two plasmids, pWJ451 (qoxA-targeting, cm^R) and pE194, an empty erythromycin- resistant erm^R) plasmid. This strain rather than the wild-type was chosen as the common competitor because measurements are more precise when the two competitors have similar fitness than when one is substantially more fit than the other (Wiser and Lenski, 2015). [0249] Importantly, the presence of all resistance cassettes (cm^R, tet^R and erm^R) did not change _{the MIC of gentamicin (data not shown). Additionally, plasmids carrying the tet} ^R _{and erm} ^R _cassettes (i.e., pT181 and pE194) were stable without selection, as 100/100 colonies checked by replica plating maintained the plasmid after growing in plain medium for 11 h (the same amount of time used for pairwise competition). Detailed experimental procedures of the pairwise competition assay are as follows. _{[0250] A single colony of S. aureus RN4220 harboring a cm} ^R _{plasmid (e.g., pWJ451, qoxA-} _{targeting) and et} ^R _{plasmid (e.g., pWJ481, ndh-targeting) was grown overnight in 4 mL of TSB with} chloramphenicol 5 µg/mL) and tetracycline (2.5 µg/mL). A single colony of the common competitor, S. aureus RN4220 harboring pWJ451 (qoxA-targeting, cm^R) and pE194 (erm^R) was grown overnight in 4 mL of TSB with chloramphenicol (5 µg/mL) and erythromycin (5 µg/mL). The two overnight ^{cultures were washed in 1X PBS and equal CFU of the two cultures were mixed based on OD} ₆₀₀ ^. ^{Next, the OD} ₆₀₀ ^{of the mixed culture was measured and 375 units of cells (e.g., if OD} ₆₀₀ ^{= 3.75,} use 100 µL) were transferred to 7.5 mL of TSB with chloramphenicol (5 µg/mL). Cultures were grown for 11 h and aliquots of cells at T = 0 and T = 11 h were plated onto TSA with chloramphenicol (10 µg/mL) and tetracycline (5 µg/mL) or TSA with chloramphenicol (10 µg/mL) and erythromycin (10 µg/mL). Agar plates were incubated 16 – 24 h at 37 °C and colonies were enumerated. [0251] Fitness of strain X (W_X) was calculated as relative to that of the common competitor:

[0252] where N_X and N_C are the population sizes of strain X and the common competitor, and subscripts f and i indicate the final and initial time points. Data analysis of dual-spacer (2S) libraries Spacer identification and sequence alignment [0253] First, for any genome, a list of all functional spacers (i.e., spacers with an “NGG” PAM) was created. Each spacer was uniquely identified by its strandedness and location in the genome. For instance, here are 136,928 unique functional spacers (i.e., occurrence of “NGG” on both strands of the DNA) n the S. aureus RN4220 genome. [0254] From the FASTQ file generated by Illumina sequencing (Miseq, 300 cycles), the two spacer sequences were identified by locating the three direct repeats flanking them. Spacers were aligned to chromosomal and plasmid genomes using “aln” and “samse” functions of the Burrows- Wheeler Alignment tool (Li and Durbin, 2009). From the output SAM file, only functional spacers (i.e., spacers with an “NGG” PAM) were considered for downstream analysis. Quantification of spacer frequency at a gene level [0255] Only reads containing two spacers were considered for this analysis. After selection in high [gent], pacer frequencies at a gene level from triplicates were calculated. For this analysis, a more stringent filter was applied in which the only considered genes were those targeted by an average of 3.5 unique spacers across triplicates, in order to avoid spacers with off-target potential. A total of 183 genes remained after this filter (Figure 38). Estimation of fitness effects using Log₂FC [0256] In order to estimate the fitness effect of spacer pairs containing qoxA (or qoxB, ndh, etc) in high [gent], he pairs were first sorted out from the 2S library. Then, the log₂ of the fold-change (log₂FC) of each pacer pair was calculated: [0257] where f_post and f_pre are t

he frequency of the spacer pair in the post- and pre-selected 2S libraries, respectively. ^{[0258] Similar to the 1S libraries, all log} ₂ ^FC _{spacer pair} ^{s were grouped based on the non-qoxA} gene they targeted, ranked, and P₉₅-log₂FC for each gene was calculated. For each gene, the mean of the P₉₅-log₂FCs from the triplicates were reported, and a one-sample T-test (one-tailed) between the triplicates of P₉₅-log₂FCs and 0 (i.e., fold-change of 1) was performed. In cases where the P₉₅-log₂FC value was not present in all triplicates post-selection, a T-test could not be performed and the P-value was assigned “N/A”. Only genes that were present in at least two of the three replicates were considered. [0259] When analyzing spacer pairs that were significantly enriched, a filter was applied in which only genes that were targeted by more than 2 crRNAs post-selection were included. This was to avoid spacers with off-target potential. Spacer order [0260] For all post-selected gene pairs containing qoxA (i.e., qoxA::geneX) or qoxB, the number of times gene X appeared as the first and the second spacer (N1 and N2, respectively) were recorded. To be stringent, a filter was applied such that only pairs in which gene X was targeted by more than one unique spacer at the dominant spacer position from triplicates were analyzed. In other words, if gene X appeared more often as Spacer 2 than as Spacer 1, Spacer 2 is the dominant position and the filter demanded that there were more than one unique spacer targeting gene X at the Spacer 2 position. [0261] For each gene, the ratio of Spacer 2 to Spacer 1 (Spc2/Spc1) was calculated as N2/N1. N1 and N2 rom triplicates were subjected to a Mann-Whitney U test to determine whether one spacer was significantly more abundant than the other. [0262] Spacer order was analyzed for pairs containing qoxA and qoxB, which were the two most abundant rRNAs among the 2S library after selection in high [gent] (Figure 38). For the most enriched pairs containing qoxC, the third most abundant crRNA, sequencing coverage dropped to a level (e.g., median number of reads for Spacer 2 was 8.5) where quantification of spacer ratio was no longer accurate (Figure 43). Figures 43A-C show sequencing coverage of spacer 1 and spacer 2 of top hits in post-selected 2S libraries containing qoxABC. Distribution of the number of reads corresponding to the non-qoxA spacer at the Spacer 1 position (yellow) and Spacer 2 position (purple) for top hits in post-selected 2S libraries containing qoxA are shown in Figure 43A. The median number of reads for Spacer 1 and Spacer 2 are also shown. Figures 43B and C show sequencing coverage for spacer pairs qoxB and qoxC, respectively. References for Example 2 [0263] Bayer, A.S., McNamara, P., Yeaman, M.R., Lucindo, N., Jones, T., Cheung, A.L., Sahl, H.G., and Proctor, R.A. (2006). Transposon disruption of the complex I NADH oxidoreductase gene (snoD) in Staphylococcus aureus is associated with reduced susceptibility to the microbicidal activity of thrombin-induced platelet microbicidal protein 1. Journal of bacteriology 188, 211- 222. [0264] Bikard, D., Jiang, W., Samai, P., Hochschild, A., Zhang, F., and Marraffini, L.A. (2013). Programmable repression and activation of bacterial gene expression using an engineered CRISPR- Cas system. Nucleic acids research 41, 7429-7437. [0265] Blount, Z.D., Borland, C.Z., and Lenski, R.E. (2008). Historical contingency and the evolution of a key innovation in an experimental population of Escherichia coli. Proceedings of the National Academy of Sciences of the United States of America 105, 7899-7906. [0266] Chaudhuri, R.R., Allen, A.G., Owen, P.J., Shalom, G., Stone, K., Harrison, M., Burgis, T.A., Lockyer, M., Garcia-Lara, J., Foster, S.J., et al. (2009). Comprehensive identification of essential Staphylococcus aureus genes using Transposon-Mediated Differential Hybridisation (TMDH). BMC genomics 10, 291. [0267] Deltcheva, E., Chylinski, K., Sharma, C.M., Gonzales, K., Chao, Y., Pirzada, Z.A., Eckert, M.R., Vogel, J., and Charpentier, E. (2011). CRISPR RNA maturation by trans-encoded small RNA and host factor RNase III. Nature 471, 602-607. [0268] Gilbert, L.A., Horlbeck, M.A., Adamson, B., Villalta, J.E., Chen, Y., Whitehead, E.H., Guimaraes, C., Panning, B., Ploegh, H.L., Bassik, M.C., et al. (2014). Genome-Scale CRISPR- Mediated Control of Gene Repression and Activation. Cell 159, 647-661. [0269] Girgis, H.S., Hottes, A.K., and Tavazoie, S. (2009). Genetic architecture of intrinsic antibiotic susceptibility. PloS one 4, e5629. [0270] Girgis, H.S., Liu, Y., Ryu, W.S., and Tavazoie, S. (2007). A comprehensive genetic characterization of bacterial motility. PLoS genetics 3, 1644-1660. [0271] Good, B.H., McDonald, M.J., Barrick, J.E., Lenski, R.E., and Desai, M.M. (2017). The dynamics of molecular evolution over 60,000 generations. Nature 551, 45-50. [0272] Han, K., Jeng, E.E., Hess, G.T., Morgens, D.W., Li, A., and Bassik, M.C. (2017). Synergistic drug combinations for cancer identified in a CRISPR screen for pairwise genetic interactions. Nature biotechnology 35, 463-474. [0273] Heler, R., Samai, P., Modell, J.W., Weiner, C., Goldberg, G.W., Bikard, D., and Marraffini, L.A. (2015). Cas9 specifies functional viral targets during CRISPR-Cas adaptation. Nature 519, 199-202. [0274] Heler, R., Wright, A.V., Vucelja, M., Bikard, D., Doudna, J.A., and Marraffini, L.A. (2017). Mutations in Cas9 Enhance the Rate of Acquisition of Viral Spacer Sequences during the CRISPR-Cas Immune Response. Molecular cell 65, 168-175. [0275] Hinz, A., Lee, S., Jacoby, K., and Manoil, C. (2011). Membrane proteases and aminoglycoside antibiotic resistance. Journal of bacteriology 193, 4790-4797. [0276] Jinek, M., Chylinski, K., Fonfara, I., Hauer, M., Doudna, J.A., and Charpentier, E. (2012). A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816-821. [0277] Kahl, B.C., Becker, K., and Loffler, B. (2016). Clinical Significance and Pathogenesis of Staphylococcal Small Colony Variants in Persistent Infections. Clinical microbiology reviews 29, 401-427. [0278] Kinkel, T.L., Roux, C.M., Dunman, P.M., and Fang, F.C. (2013). The Staphylococcus aureus SrrAB two-component system promotes resistance to nitrosative stress and hypoxia. mBio 4, e00696-00613. [0279] Lannergard, J., von Eiff, C., Sander, G., Cordes, T., Seggewiss, J., Peters, G., Proctor, R.A., Becker, K., and Hughes, D. (2008). Identification of the genetic basis for clinical menadione- auxotrophic small-colony variant isolates of Staphylococcus aureus. Antimicrobial agents and chemotherapy 52, 4017-4022. [0280] Levy, A., Goren, M.G., Yosef, I., Auster, O., Manor, M., Amitai, G., Edgar, R., Qimron, U., and Sorek, R. (2015). CRISPR adaptation biases explain preference for acquisition of foreign DNA. Nature 520, 505-510. [0281] McGinn, J., and Marraffini, L.A. (2018). Molecular mechanisms of CRISPR-Cas spacer acquisition. Nature reviews Microbiology. [0282] Mojica, F.J., Diez-Villasenor, C., Garcia-Martinez, J., and Almendros, C. (2009). Short motif sequences determine the targets of the prokaryotic CRISPR defence system. Microbiology 155, 733-740. [0283] Nair, D., Memmi, G., Hernandez, D., Bard, J., Beaume, M., Gill, S., Francois, P., and Cheung, A.L. (2011). Whole-genome sequencing of Staphylococcus aureus strain RN4220, a key laboratory strain used in virulence research, identifies mutations that affect not only virulence factors but also the fitness of the strain. Journal of bacteriology 193, 2332-2335. [0284] Qi, L.S., Larson, M.H., Gilbert, L.A., Doudna, J.A., Weissman, J.S., Arkin, A.P., and Lim, W.A. (2013). Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173-1183. [0285] Rajagopal, M., Martin, M.J., Santiago, M., Lee, W., Kos, V.N., Meredith, T., Gilmore, M.S., and Walker, S. (2016). Multidrug Intrinsic Resistance Factors in Staphylococcus aureus Identified by Profiling Fitness within High-Diversity Transposon Libraries. mBio 7. [0286] Sanjana, N.E. (2017). Genome-scale CRISPR pooled screens. Analytical biochemistry 532, 95-99. [0287] Shalem, O., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelsen, T.S., Heckl, D., Ebert, B.L., Root, D.E., Doench, J.G., et al. (2014). Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84-87. [0288] Shen, J.P., Zhao, D., Sasik, R., Luebeck, J., Birmingham, A., Bojorquez-Gomez, A., Licon, K., Klepper, K., Pekin, D., Beckett, A.N., et al. (2017). Combinatorial CRISPR-Cas9 screens for de novo mapping of genetic interactions. Nature methods 14, 573-576. [0289] Shipman, S.L., Nivala, J., Macklis, J.D., and Church, G.M. (2016). Molecular recordings by directed CRISPR spacer acquisition. Science 353, aaf1175. [0290] Taber, H.W., Mueller, J.P., Miller, P.F., and Arrow, A.S. (1987). Bacterial uptake of aminoglycoside antibiotics. Microbiological reviews 51, 439-457. [0291] Tang, W., and Liu, D.R. (2018). Rewritable multi-event analog recording in bacterial and mammalian cells. Science 360. [0292] Vestergaard, M., Nohr-Meldgaard, K., Bojer, M.S., Krogsgard Nielsen, C., Meyer, R.L., Slavetinsky, C., Peschel, A., and Ingmer, H. (2017). Inhibition of the ATP Synthase Eliminates the Intrinsic Resistance of Staphylococcus aureus towards Polymyxins. mBio 8. [0293] Wang, T., Wei, J.J., Sabatini, D.M., and Lander, E.S. (2014). Genetic screens in human cells using the CRISPR-Cas9 system. Science 343, 80-84. [0294] Yang, S.J., Bayer, A.S., Mishra, N.N., Meehl, M., Ledala, N., Yeaman, M.R., Xiong, Y.Q., and Cheung, A.L. (2012). The Staphylococcus aureus two-component regulatory system, GraRS, senses and confers resistance to selected cationic antimicrobial peptides. Infection and immunity 80, 74-81. [0295] Baba, T., Takeuchi, F., Kuroda, M., Yuzawa, H., Aoki, K., Oguchi, A., Nagai, Y., Iwama, N., Asano, K., Naimi, T., et al. (2002). Genome and virulence determinants of high virulence community-acquired MRSA. Lancet 359, 1819-1827. [0296] Bae, T., Baba, T., Hiramatsu, K., and Schneewind, O. (2006). Prophages of Staphylococcus aureus Newman and their contribution to virulence. Molecular microbiology 62, 1035-1047. [0297] Bassett, A.R., Kong, L., and Liu, J.L. (2015). A genome-wide CRISPR library for high- throughput genetic screening in Drosophila cells. Journal of genetics and genomics = Yi chuan xue bao 42, 301- 309. [0298] Bikard, D., Jiang, W., Samai, P., Hochschild, A., Zhang, F., and Marraffini, L.A. (2013). Programmable repression and activation of bacterial gene expression using an engineered CRISPR- Cas system. Nucleic acids research 41, 7429-7437. [0299] Cui, L., Vigouroux, A., Rousset, F., Varet, H., Khanna, V., and Bikard, D. (2018). A CRISPRi screen n E. coli reveals sequence-specific toxicity of dCas9. Nature communications 9, 1912. [0300] Gharehbeglou, M., Arjmand, G., Haeri, M.R., and Khazeni, M. (2015). Nonselective mevalonate inase inhibitor as a novel class of antibacterial agents. Cholesterol 2015, 147601. [0301] Gilbert, L.A., Horlbeck, M.A., Adamson, B., Villalta, J.E., Chen, Y., Whitehead, E.H., Guimaraes, C., Panning, B., Ploegh, H.L., Bassik, M.C., et al. (2014). Genome-Scale CRISPR- Mediated Control of Gene Repression and Activation. Cell 159, 647-661. [0302] Girgis, H.S., Liu, Y., Ryu, W.S., and Tavazoie, S. (2007). A comprehensive genetic characterization of bacterial motility. PLoS genetics 3, 1644-1660. [0303] Goldberg, G.W., Jiang, W., Bikard, D., and Marraffini, L.A. (2014). Conditional tolerance of temperate phages via transcription-dependent CRISPR-Cas targeting. Nature 514, 633- 637. [0304] Lee, H.H., Ostrov, N., Wong, B.G., Gold, M.A., Khalil, A.S., and Church, G.M. (2019). Functional genomics of the rapidly replicating bacterium Vibrio natriegens by CRISPRi. Nature microbiology 4, 1105-1113. [0305] Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with Burrows- Wheeler transform. Bioinformatics 25, 1754-1760. [0306] Lutz, R., and Bujard, H. (1997). Independent and tight regulation of transcriptional units in Escherichia oli via the LacR/O, the TetR/O and AraC/I1-I2 regulatory elements. Nucleic acids research 25, 1203- 1210. [0307] Modell, J.W., Jiang, W., and Marraffini, L.A. (2017). CRISPR-Cas systems exploit viral DNA injection to establish and maintain adaptive immunity. Nature 544, 101-104. [0308] Nair, D., Memmi, G., Hernandez, D., Bard, J., Beaume, M., Gill, S., Francois, P., and Cheung, A.L.2011). Whole-genome sequencing of Staphylococcus aureus strain RN4220, a key laboratory strain used in virulence research, identifies mutations that affect not only virulence factors but also the witness of the strain. Journal of bacteriology 193, 2332-2335. [0309] Phillips, P.C. (2008). Epistasis--the essential role of gene interactions in the structure and evolution of genetic systems. Nature reviews Genetics 9, 855-867. [0310] Qi, L.S., Larson, M.H., Gilbert, L.A., Doudna, J.A., Weissman, J.S., Arkin, A.P., and Lim, W.A. (2013). Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173-1183. [0311] Sanson, K.R., Hanna, R.E., Hegde, M., Donovan, K.F., Strand, C., Sullender, M.E., Vaimberg, E.W., Goodale, A., Root, D.E., Piccioni, F., et al. (2018). Optimized libraries for CRISPR-Cas9 genetic creens with multiple modalities. Nature communications 9, 5416. [0312] Segre, D., Deluna, A., Church, G.M., and Kishony, R. (2005). Modular epistasis in yeast metabolism. Nature genetics 37, 77-83. [0313] Sidik, S.M., Huet, D., Ganesan, S.M., Huynh, M.H., Wang, T., Nasamu, A.S., Thiru, P., Saeij, J.P.J., Carruthers, V.B., Niles, J.C., et al. (2016). A Genome-wide CRISPR Screen in Toxoplasma Identifies Essential Apicomplexan Genes. Cell 166, 1423-1435 e1412. [0314] Wiser, M.J., and Lenski, R.E. (2015). A Comparison of Methods to Measure Fitness in Escherichia oli. PloS one 10, e0126210. [0315] Zhang, Y.M., White, S.W., and Rock, C.O. (2006). Inhibiting bacterial fatty acid synthesis. The Journal of biological chemistry 281, 17541-17544.

Claims

What is claimed is: 1. A method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); a trans-activating crRNA sequence (tracrRNA); CRISPR-associated endonuclease Cas1 (Cas1); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR-associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to ; and

(ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding dCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; and c) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library.

2. A method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: a hyperactive variant of CRISPR associated endonuclease Cas9 (hCas9); a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); or a CRISPR-associated endonuclease Cas9 lacking endonuclease activity (dCas9); tracrRNA; Cas1; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising nucleotide sequence at least 80% identical to ; and

(ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding the CRISPR- associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) inducing the expression of the said at least one nucleic acid sequence encoding the CRISPR-associated endonuclease, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; c) introducing genomic DNA into the bacterial cells; and d) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library.

3. The method of claim 1, wherein the bacterial cells constitutively express dCas9, tracrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters.

4. The method of claim 2, wherein the bacterial cells constitutively express the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters.

5. The method of claim 1 or 3, wherein the one or more nucleic acid sequences encoding dCas9, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid.

6. The method of claim 2 or 4, wherein the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid.

7. The method of claim 2, 4, or 6, further comprising producing competent bacterial cells before introducing the genomic DNA.

8. The method of claim 2, 4, 6, or 7, wherein the genomic DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells.

9. The method of claim 2, 4, 6, or 7, wherein the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells.

10. The method of claim 2, 4, 6, 7, 8, or 9, wherein the genomic DNA is sheared before introducing the genomic DNA.

11. The method of claim 2, 4, 6, 7, 8, 9, or 10, wherein the genomic DNA is introduced into the bacterial cells using electroporation.

12. The method of claim 1, 3, or 5, wherein in step c) a nucleic acid sequence from the bacterial cell is integrated into the CRISPR array.

13. The method of claim 2, 4, 6, 7, 8, 9, or 10, wherein a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array.

14. The method of claims 12 or 13, wherein transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed.

15. The method of claim 2, 4, 6, 7, 8, 9, 10, 12, 13, or 14, wherein the CRISPR-associated endonuclease is hCas9.

16. The method of claim 2, 4, 6, 7, 8, 9, 10, 12, 13, or 14, wherein the CRISPR-associated endonuclease is dCas9.

17. The method of claim 2, 4, 6, 7, 8, 9, 10, 12, 13, or 14, wherein the CRISPR-associated endonuclease is hdCas9.

18. The method of claim 1, 3, 5, 12, or 14, wherein the CRISPR-associated endonuclease is a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9).

19. The method of any one of claims 1 to 18, further comprising isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array.

20. The method of claim 19, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

21. The method of claim 20, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

22. The method of claim 21, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

23. The method of any one of claims 1 to 19, further comprising generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound.

24. The method of claim 23, wherein the test compound is an antibiotic.

25. The method of claim 23 or 24, further comprising isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array.

26. The method of claim 25, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

27. The method of claim 26, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

28. The method of claim 27, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

29. The method of claim 28, further comprising identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

30. The method of claim 20, further comprising subcloning the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence into a plasmid.

31. The method of claim 30, further comprising: providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a Cas9 lacking endonuclease activity (dCas9) and a tracrRNA, wherein the one or more nucleic acid sequences encoding dCas9 and tracrRNA is under the control of one or more inducible promoters; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array; inducing the expression of the said at least one nucleic acid sequence encoding dCas9 and tracrRNA that is under the control of the one or more inducible promoters; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a crRNA library.

32. The method of claim 30 or 31, wherein the bacterial cells of the population of bacterial cells and the bacterial cells of the second population of bacterial cells are different species.

33. The method of claim 31 or 32, further comprising generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the crRNA library with a test compound.

34. The method of claim 33, wherein the test compound is an antibiotic.

35. The method of claim 33 or 34, further comprising isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array.

36. The method of claim 35, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

37. The method of claim 36, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

38. The method of claim 37, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

39. The method of claim 38, further comprising identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

40. A method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing a population of bacterial cells that comprises one or more nucleic acid sequences encoding: hdCas9; tracrRNA; Cas1; Cas2; Csn2; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to GTTTTAGAGCTATGCTGTTTTGAATGGTCCCAAAAC (SEQ ID NO: 1); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constiutively expressed; b) isolating a single colony from the population of bacterial cells; c) incubating the single colony of bacterial cells to produce a population of clonal bacterial cells; d) inducing the expression of the said at least one nucleic acid sequence encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of the one or more inducible promoters; e) introducing genomic DNA into the bacterial cells of the population of clonal bacterial cells; and f) incubating the bacterial cells, thereby producing a population of bacterial cells comprising a crRNA library.

41. The method of claim 40, wherein the bacterial cells constitutively express hdCas9, tracrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters.

42. The method of claim 40 or 41, wherein the one or more nucleic acid sequences encoding hdCas9, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid.

43. The method of claim 40, 41, or 42, further comprising producing competent bacterial cells before introducing the genomic DNA.

44. The method of claim 40, 41, 42, or 43, wherein the genomic DNA is of the same bacterial species as the bacterial cells of the population of bacterial cells.

45. The method of claim 40, 41, 42, or 43, wherein the genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells.

46. The method of claim 40, 41, 42, 43, 44, or 45 wherein the genomic DNA is sheared before introducing the genomic DNA.

47. The method of claim 40, 41, 42, 43, 44, 45, or 46 wherein the genomic DNA is introduced into the bacterial cells using electroporation.

48. The method of claim 40, 41, 42, 43, 44, 45, 46, or 47 wherein a nucleic acid sequence from the genomic DNA is integrated into the CRISPR array.

49. The method of any one of claims 40 to 48, further comprising isolating from the population of bacterial cells or from a portion of the population of bacterial cells comprising the crRNA library at least the one or more nucleic acid sequences encoding the CRISPR array.

50. The method of claim 49, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

51. The method of claim 50, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

52. The method of claim 51, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

53. The method of any one of claims 40 to 48, further comprising generating a selected population of bacterial cells by contacting the population of bacterial cells or a portion of the population of bacterial cells comprising the crRNA library with a test compound.

54. The method of claim 53, wherein the test compound is an antibiotic.

55. The method of claim 53 or 54, further comprising isolating from the selected population of bacterial cells or from a portion of the selected population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array.

56. The method of claim 55, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

57. The method of claim 56, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

58. The method of claim 57, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

59. The method of claim 58, further comprising identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

60. The method of claim 30, further comprising providing a second population of bacterial cells that comprises one or more nucleic acid sequences encoding a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a tracrRNA; Cas1; Cas2; and Csn2; wherein at least one of the one or more nucleic acid sequences encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 that are not under the control of the one or more inducible promoters are constiutively expressed; introducing into the second population of bacterial cells the subcloned plasmid comprising the one or more amplified nucleic acid sequences encoding the CRISPR array, wherein said CRISPR array comprises at least a partial repeat-spacer-repeat sequence; inducing the expression of the said at least one nucleic acid sequence encoding the hdCas9, tracrRNA, Cas1, Cas2, and Csn2 that is under the control of the one or more inducible promoters; and introducing genomic DNA into the bacterial cells of the second population of bacterial cells; and incubating the second population of bacterial cells, thereby producing a second population of bacterial cells comprising a dual crRNA library.

61. The method of claim 60, further comprising generating a selected population of the second population of bacterial cells by contacting the second population of bacterial cells or a portion of the second population of bacterial cells comprising the dual crRNA library with a test compound.

62. The method of claim 61, wherein the test compound is an antibiotic.

63. The method of claim 61 or 62, further comprising isolating from the selected second population of bacterial cells or from a portion of the selected second population of bacterial cells at least the one or more nucleic acid sequences encoding the CRISPR array.

64. The method of claim 63, further comprising amplifying the one or more nucleic acid sequences encoding the CRISPR array.

65. The method of claim 64, further comprising detecting the amplified nucleic acid sequences encoding the CRISPR array.

66. The method of claim 65, wherein the amplified nucleic acid sequences encoding the CRISPR array is detected by sequencing.

67. The method of claim 66, further comprising identifying one or more target genes that are targeted by crRNA molecules encoded by the CRISPR array.

68. A method of producing a population of bacterial cells comprising a CRISPR RNA (crRNA) library, the method comprising: a) providing two populations of bacterial cells wherein the bacterial cells of each population comprise one or more nucleic acid sequences encoding: a hyperactive variant of CRISPR-associated endonuclease Cas9 lacking endonuclease activity (hdCas9); a trans-activating crRNA sequence (tracrRNA); CRISPR-associated endonuclease Cas1 (Cas1); CRISPR-associated endoribonuclease Cas2 (Cas2); CRISPR-associated protein Csn2 (Csn2); and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

); and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein the one or more nucleic acid sequences encoding hdCas9, tracrRNA, and the CRISPR array are constitutively expressed and are present on a first plasmid and wherein the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are under the control of one or more inducible promoters and are present on a second plasmid, wherein the second plasmid further comprises an origin of replication and a nucleic acid sequence encoding resistance to an antibiotic wherein the origin of replication and nucleic acid sequence encoding resistance to an antibiotic are different between the first and second populations of bacterial cells; b) inducing in both populations of bacterial cells the expression of the said at least one nucleic acid sequence encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that is under the control of one or more inducible promoters; c) mixing the bacterial cells of both populations of bacterial cells to generate a third population of bacterial cells; d) introducing genomic DNA into the bacterial cells of the third population of bacterial cells; and e) incubating the bacterial cells of step d), thereby producing a population of bacterial cells comprising a crRNA library.

69. The method of any one of claims 1 to 68, wherein the bacterial cells of the population of bacterial cells are Gram-positive.

70. The method of any one of claims 1 to 68, wherein the bacterial cells of the population of bacterial cells are Staphylococcus aureus.

71. The method of any one of claims 1 to 68, wherein the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA).

72. The method of any one of claims 1 to 68, wherein the bacterial cells of the population of bacterial cells are Escherichia coli.

73. The method of any one of claims 1 to 68, wherein the bacterial cells of the population of bacterial cells are pathogenic.

74. The method of claim 10 or 45, wherein the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the genomic DNA is of Escherichia coli.

75. The method of claim 31, wherein bacterial cells of the population of bacterial cells are Staphylococcus aureus and the bacterial cells of the second population of bacterial cells are Escherichia coli.

76. The method of any one of claims 1 to 75, wherein the tracrRNA is a minimal tracrRNA sequence required for targeting.

77. The method of claim 76, wherein said minimal tracrRNA sequence comprises 89 nucleotides.

78. A population of bacterial cells capable of producing a CRISPR RNA (crRNA) library, said bacterial cells of the population of bacterial cells comprising one or more nucleic acid sequences encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; tracrRNA; Cas1; Cas2; Csn2; and CRISPR i i (i) a repeat sequence comprising a nucleotide sequence at least 80% identical to

and (ii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; wherein at least one of the one or more nucleic acid sequences encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array is under the control of one or more inducible promoters and wherein the one or more nucleic acid sequences encoding hdCas9, tracrRNA, Cas1, Cas2, Csn2, and the CRISPR array that are not under the control of the one or more inducible promoters are constitutively expressed.

79. The population of bacterial cells of claim 78, wherein the bacterial cells constitutively express the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array and wherein expression of the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 is under the control of the one or more inducible promoters.

80. The population of bacterial cells of claim 78 or 79, wherein the one or more nucleic acid sequences encoding the CRISPR-associated endonuclease, tracrRNA, and the CRISPR array are present on a first plasmid and the one or more nucleic acid sequences encoding Cas1, Cas2, and Csn2 are present on a second plasmid.

81. The population of bacterial cells of claim 78, 79, or 80, wherein the CRISPR array further comprises a nucleic acid sequence integrated from the bacterial cell.

82. The population of bacterial cells of claim 78, 79, or 80, wherein the CRISPR array further comprises a nucleic acid sequence integrated from an exogenous genomic DNA.

83. The population of bacterial cells of claim 81 or 82, wherein transcription of a gene corresponding to the nucleic acid sequence integrated into the CRISPR array is repressed.

84. The population of bacterial cells of any one of claims 78 to 83, wherein the CRISPR-associated endonuclease is hCas9.

85. The population of bacterial cells of any one of claims 78 to 83, wherein the CRISPR-associated endonuclease is dCas9.

86. The population of bacterial cells of any one of claims 78 to 83, wherein the CRISPR-associated endonuclease is hdCas9.

87. The population of bacterial cells of claim 82, wherein the exogenous genomic DNA is of a different bacterial species as the bacterial cells of the population of bacterial cells.

88. The population of bacterial cells of any one of claims 78 to 87, wherein the bacterial cells of the population of bacterial cells are Gram-positive.

89. The population of bacterial cells of any one of claims 78 to 87, wherein the bacterial cells of the population of bacterial cells are Staphylococcus aureus.

90. The population of bacterial cells of any one of claims 78 to 87, wherein the bacterial cells of the population of bacterial cells are Methicillin-resistant Staphylococcus aureus (MRSA).

91. The population of bacterial cells of any one of claims 78 to 87, wherein the bacterial cells of the population of bacterial cells are Escherichia coli.

92. The population of bacterial cells of any one of claims 78 to 87, wherein the bacterial cells of the population of bacterial cells are pathogenic.

93. The population of bacterial cells of claims 82 or 87, wherein the bacterial cells of the population of bacterial cells are Staphylococcus aureus and the exogenous genomic DNA is of Escherichia coli.

94. The population of bacterial cells of any one of claims 78 to 93, wherein the tracrRNA is a minimal tracrRNA sequence required for targeting.

95. The population of bacterial cells of claim 94, wherein said minimal tracrRNA sequence comprises 89 nucleotides.

96. The method of any one of claims 1 to 77, wherein the crRNA library covers at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome.

97. The population of bacterial cells of any one of claims 78 to 96, wherein the crRNA library covers at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99% or 100% of a target genome.

98. The method of claim 1, further comprising isolating said crRNA library.

99. A first plasmid comprising a nucleic acid sequence encoding: a CRISPR-associated endonuclease selected from: hCas9; hdCas9; or dCas9; tracrRNA; and a CRISPR array comprising: (i) a repeat sequence comprising a nucleotide sequence at least 80% identical and

Ĩii) a canonical sequence located upstream of the repeat, wherein said canonical sequence comprises a promoter and a leader-anchoring sequence and wherein said repeat sequence is duplicated after each adaptation event; operably linked to a constitutive promoter sequence; and a second plasmid comprising a nucleic acid sequence encoding: Cas1; Cas2; and Csn2; operably linked to an inducible promoter sequence.