1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Running title: Electron Cryo-Tomography of Procentriole
Electron Cryo-Tomography Provides Insight into Procentriole
Architecture and Assembly Mechanism
Sam Li1*, Jose-Jesus Fernandez2, Wallace F. Marshall1 and David A. Agard1,3*
1. Department of Biochemistry and Biophysics, University of California, San
Francisco, CA 94158, USA.
2. Centro Nacional de Biotecnologia – CSIC, Campus Universidad Autonoma,
Cantoblanco, 28049 Madrid, Spain
3. Howard Hughes Medical Institute, University of California, San Francisco, CA
94158, USA.
21
22
23
24
25
26
27
28
29
30
31
32
33
*Correspondence:
Sam Li
Phone: 1-415-502-2930
Fax: 1-415-476-1902
Email: samli@msg.ucsf.edu
David A. Agard
Phone: 1-415-476-2521
Fax: 1-415-476-1902
Email: agard@msg.ucsf.edu
34
35
36
37
38
39
1
40
41
42
43
Abstract
44
biogenesis and homeostasis is tightly regulated. Using electron cryo-tomography
45
(cryoET) we present the structure of procentrioles from Chlamydomonas reinhardtii. We
46
identified a set of non-tubulin components attached to the triplet microtubule (MT), many
47
are at the junctions of tubules likely to reinforce the triplet. We describe structure of the
48
A-C linker that bridges neighboring triplets. The structure infers that POC1 is likely an
49
integral component of A-C linker. Its conserved WD40 β-propeller domain provides
50
attachment sites for other A-C linker components. The twist of A-C linker results in an
51
iris diaphragm-like motion of the triplets in the longitudinal direction of procentriole.
52
Finally, we identified two assembly intermediates at the growing ends of procentriole
53
allowing us to propose a model for the procentriole assembly. Our results provide a
54
comprehensive structural framework for understanding the molecular mechanisms
55
underpinning procentriole biogenesis and assembly.
Centriole is an essential structure with multiple functions in cellular processes. Centriole
56
57
Keywords: centriole, procentriole, basal body, probasal body, microtubule triplet,
58
electron cryo-tomography, subtomogram averaging, structure heterogeneity
59
2
60
61
62
63
64
Introduction
65
characteristic nine-fold symmetry. As an evolutionarily conserved organelle, the
66
centriole, also known as the basal body, fulfills many cellular functions. In cycling cells,
67
a pair of centrioles recruits pericentriolar material (PCM). Together they form the
68
centrosome, the primary microtubule organizing center (MTOC) in animal cells. Related
69
to its function as an MTOC, the centrosome is essential for mitotic spindle formation,
70
spindle orientation and faithful mitotic chromosome segregation and for intracellular
71
transport of cargoes. In non-dividing cells, the centriole functions as basal body to
72
template cilium formation. Based on their structure and functions, the cilia can be further
73
classified into primary or motile cilia. The primary cilium functions as a cell “antenna”
74
that senses diverse signals on both sides of cell membrane. Motile cilia are responsible
75
for cellular swimming and fluid flow. Given the array of diverse functions carried out by
76
centrioles and basal bodies, it is not surprising that mutations affecting centrioles and
77
basal bodies cause diverse human diseases, ranging from tumors to different forms of
78
ciliopathies (Nigg and Raff, 2009; Reiter and Leroux, 2017)
The centriole is a barrel-shaped structure composed of a set of MT triplets arranged in a
79
80
Centriole biogenesis is tightly controlled. Recent studies have illuminated the
81
mechanisms that regulate centriole assembly. Genetics and cell biological studies in
82
various model organisms show that centriole biogenesis occurs by a cascade of molecular
83
events performed by evolutionarily conserved components, reviewed by (Banterle and
84
Gonczy, 2017). In vertebrate cells, the sequential recruitment of centriole components
3
85
during assembly is accompanied by a series of morphological changes at the newly
86
emerging centrioles. First, the nine-fold symmetric structure called the cartwheel forms at
87
the proximal side of the mature centriole (mother centriole). This is followed by assembly
88
of MT triplets at the tip of the nine radial spokes in the cartwheel. Together, they form
89
the procentriole, a precursor of the daughter centriole. The procentriole continue to
90
elongate and develop as the new daughter centriole prior to mitosis. During mitosis, the
91
mother and daughter centrioles are “disengaged” but remain loosely linked. In the next
92
cell cycle, the daughter centriole will completely separate from the mother. Meanwhile, it
93
acquires the appendages and the PCM and becomes fully competent as an MTOC. These
94
mark the completion of the centriole duplication cycle (Kong et al., 2014). Similar to the
95
centriole duplication in metazoans such as in mammals, the centriole duplication process
96
in the unicellular organism Chlamydomonas reinhardtii is tightly controlled. The process
97
has been described in a series of seminal studies (Cavalier-Smith, 1974; Geimer and
98
Melkonian, 2004; O'Toole and Dutcher, 2014). Compared to the vertebrates, despite
99
many morphological and ultrastructural differences in the duplication steps, a number of
100
key components in the Chlamydomonas centriole assembly have been found conserved in
101
other organisms (Dutcher et al., 2002; Dutcher and Trabuco, 1998; Hiraki et al., 2007;
102
Keller et al., 2009; Matsuura et al., 2004; Nakazawa et al., 2007). In addition, proteomics
103
and bioinformatics studies in several model organisms have identified a list of major
104
structural components of centrioles (Andersen et al., 2003; Keller et al., 2005; Kilburn et
105
al., 2007; Li et al., 2004; Muller et al., 2010). Together, these studies concur that the
106
centriole is assembled by a series of evolutionarily conserved protein building blocks.
4
107
The process is tightly controlled spatially and temporally by a set of regulatory proteins
108
(Carvalho-Santos et al., 2010; Hodges et al., 2010).
109
110
Meanwhile, structural approaches, including super-resolution light microscopy, X-ray
111
crystallography and electron cryo-microscopy, have been applied to put the building
112
blocks into the context of the centriole’s 3D structure. Several crystal structures are now
113
available describing components of the centriole, including Plk4, Spd2, Sas6, Cep135,
114
STIL and CPAP. In addition, there have been cryoET studies on assembly of centriole in
115
several organisms (Greenan et al., 2018; Guichard et al., 2010; Guichard et al., 2012; Li
116
et al., 2012). In particular, the events of cartwheel assembly has been studied extensively
117
(Guichard et al., 2012; Guichard et al., 2017; Hilbert et al., 2016; Kitagawa et al., 2011;
118
van Breugel et al., 2011), leading to a molecular mechanism that at least in part
119
establishes the 9-fold symmetry, reviewed in (Guichard et al., 2018).
120
121
Despite the structural and functional study of many centriole components in the past
122
years, a complete picture of the centriole architecture and its assembly mechanism is
123
lacking. Using cryoET and subtomogram averaging, we describe the triplet structure of
124
the Chlamydomonas reinhardtii procentriole. We identify 11 non-tubulin components in
125
the structure that are associated to the triplet tubules in an asymmetric manner. We
126
further present the structure of the A-C linker that laterally bridges neighboring triplets.
127
Finally, using extensive classification and averaging in image processing, we identified
128
two partially assembled triplets at the growing ends to the procentrioles that shed light on
129
the mechanism of triplet and procentriole assembly. Overall, our work presented here
5
130
builds a framework for understanding the mechanism of centriole biogenesis in molecular
131
details.
6
132
Results
133
134
Overall Architecture of the Procentriole
135
To study the structure of both the centriole and procentriole, the nuclear-flagellar-
136
apparatus (NFAp) from the unicellular green algae Chlamydomonas reinhardtii were
137
isolated and visualized by cryoET. The collected tilt series and the reconstructed
138
tomograms show good conservation of the stereotypical structures of NFAp (Video 1).
139
These include the proximal and distal striated fibers that connect two mature centrioles, a
140
set of rootlet microtubules (rMT), in the distal end of centriole there is transition zone
141
where it connects to flagellum (Figure 1A). At the centriole proximal end, a cartwheel
142
structure, including the central hub and the radial spoke, are visible. In many tomograms,
143
two well-preserved procentrioles are attached to their respective mother centrioles via the
144
rMT (Figure 1B, Video 1). Here, we focused our study on both the proximal ~100 nm
145
region of mother centrioles and on the attached procentrioles. Using subtomogram
146
averaging, we obtained an averaged structure of MT triplet at 23.0 Å resolution (Figure
147
1C, Figure1-figure supplement 1). In this structure, all MT protofilaments (PF) can be
148
resolved, as well as the 4 nm repeat of tubulin along the PF. This confirms previous
149
observations from the central core region and from centrioles in other organisms that the
150
A-tubule is composed of 13 PFs as an elliptical ring deviated from the canonical MT
151
structure. Both B- and C-tubules are partial rings with 10 PFs. In additional, a number of
152
non-MT structures are readily visible. These include the pinhead, a structure that
153
connects the A-tubule to the cartwheel, and the A-C linker that bridges A-tubule to the C-
154
tubule of its neighboring triplet. To visualize the structure of the entire procentriole, we
7
155
took reverse steps of subtomogram averaging and put the averaged triplet into the context
156
of entire procentriole tomogram. The result is a model representing a cross section of the
157
procentriole (Figure 1D). In the model, without imposing symmetry, the pinhead is about
158
70 nm away from the central hub. The A-C linker is clearly visible and makes multiple
159
connections to both the neighboring A- and C-tubules, although with blurred details
160
compared to the MTs in the triplet, indicating structural heterogeneity and flexibility in
161
this region.
162
163
To assess whether there are structural difference between the proximal region of mother
164
centrioles (10854 subtomograms) and adjacent procentrioles (2083 subtomograms), we
165
averaged these two datasets independently. This results in two triplet averages at 23.0 Å
166
and 30.1 Å resolution respectively (Figure1-figure supplement 2). Even though the MT
167
backbones are nearly identical, there are notable differences between these two averages.
168
A number of microtubule inner proteins (MIPs) that are attached to the centriole B-
169
tubule, in particular MIPs present in the inner junction of A-B and B-C tubule, are absent
170
in the procentriole triplet. Conversely, some luminal densities in the procentriole are
171
missing in the centriole. These differences may reflect difference in kinetics or transient
172
association of proteins at various stages of assembly, or incompletion of procentriole
173
assembly (see below for further detailed analysis on the B- and C-tubules in the
174
procentrioles).
175
176
Non-tubulin Components Associated with the Procentriole Triplet
8
177
Unlike many biological complexes for which the individual components might have rigid
178
and well-defined structures, large organelle-scale assemblages are often intrinsically
179
flexible and heterogeneous, creating challenges for their structure study. To improve
180
resolution, we applied a “focused refinement” strategy widely used in single particle
181
cryoEM (Scheres, 2016), to align the A-, B- and C-tubules of the triplet separately. These
182
resulted in three averaged tubule structures, with improved resolution at 21.4 Å, 22.3 Å
183
and 22.3 Å, respectively (Figures 2A-C, Figure1-figure supplement 1). Based on the
184
improved structures, we built a MT model for each tubule, allowing us to identify a total
185
of 11 MIPs associated to the MT wall (Figures 2A-C, Figure 2-figure supplements 1-11).
186
Among them, two components MIP1 and MIP2 share similar binding pattern as the MIPs
187
previously seen in the core region (Li et al., 2012). MIP1 is a cone-shaped structure
188
projecting from PF A5 into the lumen of the A-tubule and having an 8 nm longitudinal
189
periodicity. Each MIP1 has two legs that recognize the luminal side of α/β tubulin dimer
190
(arrowheads in Figure 2-figure supplement 1). By contrast, MIP2 is more filamentous,
191
forming a trellis-like meshwork, laterally spanning PFs A11 to A13. It follows the pitch
192
of the 3-start helical MT lattice running along the A-tubule luminal wall (Figure 2-figure
193
supplement 2), presumably to reinforce the “ribbon” structure that is shared by the A- and
194
B-tubules. Besides MIP1 and MIP2, the overall pattern of the MIPs bound within the
195
procentriole is markedly different from the core region. The procentriole MIP binding
196
patterns are shown in detail in Figure 2-figure supplements 1-11 and their characteristics
197
are summarized in Table 2. Of interest, 7 of the 11 MIPs (MIPs 3~6, MIPs 9~11) are
198
localized at or in the vicinity of the inner junctions of the A-B or B-C tubules. Among
199
them, the MIP9 at the inner B-C tubule junction, is of particular interest. Even though it
9
200
forms a filament exhibiting 4 nm periodicity, the structure deviates from a canonical MT
201
protofilament (Figure 2-figure supplement 12). Therefore, we assigned it as a non-tubulin
202
protein MIP9. Together, these inner junctions MIPs crosslink multiple PFs forming
203
intricate networks. They are likely important in strengthening the overall triplet structure.
204
205
The cartwheel is a unique procentriolar structure, emerging early in centriole assembly,
206
that is essential in establishing both the 9-fold symmetry and for scaffolding procentriole
207
construction. The overall structure of the cartwheel can be divided into 3 parts, the central
208
hub, the radial spoke and the pinhead (Figure 1B) (Hirono, 2014). The pinhead connects
209
the cartwheel to the MT triplets. Using local classification to focus on the pinhead, we
210
identified 4763 subtomograms having relatively intact pinhead structures, resulting in an
211
averaged pinhead structure at 23.1 Å resolution (Figures 2D, Figure 2-figure supplement
212
14). It is a hook-shaped structure anchored to PF A3 of the A-tubule. Consistent with a
213
previous study (Guichard et al., 2013), the pinhead can be partitioned into three parts,
214
namely pinbody (PinB), pinfoot 1 (PinF1) and pinfoot 2 (PinF2) where both pinfeet
215
attach to the MT wall. Each PinF1/PinF2 unit forms a ring structure bound to α/β tubulin
216
dimer on PF A3, therefore, the entire pinhead repeats every 8 nm. This is in agreement
217
with a recent triplet structure from the proximal region of mammalian centriole (Greenan
218
et al., 2018). The PinB turns 90 degree from the pinfeet, forming an inverted L-shaped
219
structure (Figure 2D). There is a longitudinal gap between neighboring PinB every 8 nm
220
without connection. The distal tip of PinB is flexible exhibiting less well-defined
221
structure in the average (arrowheads in Figure 2D and Figure 2-figure supplement 13).
222
Interestingly, among the 13 PFs that form the elliptical-shaped A-tubule, the PFs A2 and
10
223
A3 display the highest curvature of the MT wall with a large lateral gap between these
224
two PFs where the MIP4 laterally crosslinks. The high local curvature and the gap may
225
create a unique site for anchoring the pinhead to the A-tubule.
226
227
Structure of the A-C linker that Connects Neighboring Triplets
228
Procentriole is a cylindrical structure composed of 9 MT triplet blades. Each triplet is
229
laterally connected to its neighboring triplets by a structure called the A-C linker. In the
230
averaged triplet structure (Figure 1C), the A-C linker is weaker than the MT triplet,
231
indicating either compositional and/or conformational heterogeneity. By extensive
232
subtomogram classification and alignment using the A-tubule as a reference (Figure 3-
233
figure supplement 1 and detailed in Materials and Methods), we obtained 4 classes whose
234
averaged structures showing a more complete and detailed structure of the A-C linker
235
most closely associated with the A-tubule (Figure 3-figure supplements 2, 5). Similarly,
236
in a “reciprocal classification” by using the C-tubule as the reference point, we identified
237
6 classes of subtomograms whose averaged structure shows enhanced detail for the
238
portion of the A-C linker associated with the C-tubule (Figure 3-figure supplements 3-5).
239
The differences in the classes largely reflect the underlying flexibility of the A-C linker
240
about its midpoint. We combined the above two averages by docking them into a lower
241
resolution class-averages, resulting in a structure of the complete A-C linker for each
242
identified class (Figure 3-figure supplement 6).
243
244
Overall, the A-C linker forms a crisscross-shaped structure (Figure 3A, Video 2) that can
245
be divided into five parts: a central trunk region, from which two arms (Arm A, B) and
11
246
two legs (Leg A, B), extend out to the C- and A-tubules, respectively. Arm A forms a
247
longitudinal helical filament with 8 nm periodicity and binds to PF C8 of the C-tubule
248
(Figure 3B). The Arm B inserts into the trough between PFs C9 and C10 of the C-tubule,
249
likely binding to both PFs. Leg A and B fork out from the central trunk towards the A-
250
tubule. The Leg A has a thin rod shape that tilts ~30° towards the proximal end of
251
procentriole and extends about 10 nm to reach and contact PF A6 of A-tubule (Figure
252
3C). Leg B tilts ~30° towards the distal end of procentriole. It spans about 22 nm to
253
connect to both PFs A9 and A10 of the A-tubule. It continues to reach as far as to the
254
outer junction of A- and B-tubules (Figures 3A,B). Lastly, the central trunk can be further
255
divided into three longitudinal filaments bundled into a spiral: f1, f2, f3 (Figures 3A,C).
256
257
Strikingly there is a doughnut-shaped density with the Leg B. The averaged diameter of
258
the ring is ~4 nm and they stack longitudinally with 8 nm periodicity (Figures 3A,B).
259
Previous studies found the conserved centriole components, POC1, to localize to the
260
centriole proximal ends (Keller et al., 2009; Pearson et al., 2009b). Knockdown or
261
deletion of POC1 in several organisms resulted in defective centrioles or basal bodies
262
(Khire et al., 2016; Pearson et al., 2009a; Venoux et al., 2013). Human mutations of
263
either paralog, POC1A or POC1B, cause ciliopathy-like pathologies (Beck et al., 2014;
264
Roosing et al., 2014; Sarig et al., 2012; Shaheen et al., 2012). In Tetrahymena, POC1 is
265
important for maintaining A-C linker integrity. poc1 null mutants display basal body
266
defects ranging from missing or disintegrated triplets to disconnected neighboring triplets
267
and aberrant A-C linkers (Meehl et al., 2016; Pearson et al., 2009a). Interestingly, one of
268
predicted conserved signatures of POC1 is multiple WD40 repeats at its N-terminus
12
269
(Woodland and Fry, 2008), for example, in Chlamydomonas, POC1 is predicted to have 7
270
tandem WD40 repeats at its N-terminus (Keller et al., 2009). A crystal structure of a 7-
271
repeat WD40 β-propeller domain fits remarkably well into the ring density observed on
272
the Leg B (Figures 3D,E). Due to relatively modest resolution of our average map, the
273
precise orientation of the WD40 domain could not be defined. Nevertheless, in the
274
docked model, the WD40 β-propeller makes multiple contacts with the rest of Leg B,
275
consistent with the function of WD40 domain as one of the most abundant protein-
276
protein interaction domains that scaffolds multi-protein complexes (Stirnimann et al.,
277
2010). Notably, human ciliopathy mutations in POC1A or POC1B map to the WD40
278
repeats. Thus, combined with these data, our structure suggests that POC1 is likely an
279
integral component of the procentriole A-C linker that bridges neighboring triplets. Its N-
280
terminal WD40 β-propeller domain provides multiple sites for interacting with other A-C
281
linker components. However, in addition to procentriole, POC1 has also been found at
282
other locations in mother centriole and in flagellum (Keller et al., 2009; Pearson et al.,
283
2009b). A precise localization of POC1 gene product on the centriole has to wait for
284
future higher resolution structure by cryoET and other studies.
285
286
The Twist of A-C linker Results in Iris Diaphragm Motion of the Procentriole
287
As discussed above, analysis of the A-C linker heterogeneity led to the identification of 4
288
and 6 conformational classes in two classification schemes where either the A- or C-
289
tubule was used as a reference point (Figure 3-figure supplements 1,3). Overlaying
290
projections from these classes, in both schemes, reveals a large and continuous swinging
291
motion of the remaining portion of the linker relative to the reference point (Videos 3,4).
13
292
The swing angle θ (Figure 4A, Table 3) changes up to 10 or 16 degrees respectively in
293
two schemes. Since the proximal region of the mother centriole has a defined length,
294
calculating the weighted average of the swing angle at each longitudinal position along
295
the triplet (Figures 4B,C) shows that two neighboring triplets, by using the A-C linker as
296
a pivot point, twist progressively along the procentriole in a left-handed manner with the
297
thumb pointing towards the MT plus end, as illustrated in Figure 4A.
298
299
To further identify the origin of this twist motion, we overlaid the complete A-C linker
300
for each class obtained by following steps illustrated in Figure 3-figure supplement 6.
301
The two sets of overlaid structures are colored as a “heat map” with increasing
302
temperature from the proximal to the distal position along the longitudinal axis of the
303
centriole. The overlays clearly show the central trunk region of the A-C linker is at the
304
pivot point of this twist motion (Figures 4D,E). Interestingly, the trunk is composed of a
305
bundle of 3 inter-connected filaments, f1, f2 and f3 (Figures 3A-C). Together they form a
306
spiral that likely drives, at least in part, the twist motion of neighboring triplets.
307
308
Finally, to directly visualize the twist of neighboring triplets in the context of the entire
309
procentriole, we placed the averaged triplet structure back into one of the procentriole
310
tomograms by reversing the steps of subtomogram averaging. The result is a longitudinal
311
segment of the procentriole. Proceeding from the proximal towards the distal end (Video
312
5), the 9 triplets rotate concomitantly while the A-C linkers are at the pivot points of the
313
twist. At the distal end, the triplets become more tangential to the circumference of
314
procentriole cylinder. As a consequence, the luminal diameter gradually increases as if
14
315
opening up an iris diaphragm. Similar diaphragm-like motion of the centriole has been
316
observed previously in several organisms, including Chlamydomonas, Tetrahymena and
317
mammalian cells (Anderson, 1972; Li et al., 2012; Meehl et al., 2016; Paintrand et al.,
318
1992). Interestingly, a Tetrahymena POC1 null mutation results in a reduced triplet twist
319
angle (Meehl et al., 2016), likely due to defective A-C linker structure. It is likely that
320
this longitudinal twist of triplets is a common structure feature of procentrioles and
321
centrioles. Taken together, our structural analysis of the A-C linker has revealed a
322
characteristic iris diaphragm motion of the procentriole that is accommodated by the twist
323
of a spiral in the central trunk of the A-C linker.
324
325
Intermediate State of the B-tubule during Triplet Assembly
326
Many of our tomograms contain procentrioles attached to mother centrioles that are in the
327
process of assembly (Figure 1B, Video 1). To identify any assembly intermediates and to
328
study the assembly process, we set out to analyze the triplets from these procentrioles.
329
Among 110 tomograms collected, the length of the procentriole triplet varied
330
substantially, with the A-tubule always being the longest tubule, followed by the B- then
331
the C-tubules. Most of the procentriole MTs are straight with a slightly flared end at the
332
MT tip (Figure 5A), consistent with the morphology of slow growing MT and elongating
333
triplets (Guichard et al., 2010; Hoog et al., 2011). The B- and C-tubules are attached to
334
the middle of A- or B-tubules at different heights, respectively, suggesting that the B- and
335
C-tubules initiate assembly by using the A- and B-tubule wall as a template rather than a
336
minus-end template such as the -tubulin ring complex. Once initiated, the B- and C-
337
tubules then extend longitudinally towards both the MT minus and the plus ends.
15
338
Histograms of the measured tubule length in 110 procentrioles show wide distributions
339
for all 3 tubules (Figure 5B), demonstrated by the large standard deviations. This length
340
variation also implies substantial structural heterogeneity in the procentriole triplets
341
where the B- and C-tubules are likely partially assembled. We set out to identify any
342
assembly intermediates using classification.
343
344
We first focused on classification of the B-tubules. Based on 2083 subtomogram volumes
345
from 110 procentrioles, we obtained 8 classes (Figure 5-figure supplement 1).
346
Surprisingly, one of the classes, composed of 157 subtomograms, shows partially
347
assembled B-tubules. The average of this class shows three PFs B1, B2 and B3 at outer
348
A-B junction having more prominent density than the other PFs in the B-tubule. We
349
mapped these 157 subtomograms to find their location in the procentriole. Based on their
350
longitudinal positions, we divided them into 5 classes. These are 1) at the proximal end of
351
the triplet (28.7%), 2) at the distal end (25.5%), 3) a nascent B-tubule where the entire B-
352
tubule is partially assembled (33.8%), 4) extremely short and noisy B-tubule less than 10
353
nm long (7.0%), 5) partial B-tubule structure found in the midst of complete tubule
354
(5.1%), likely these are defective structures or errors in classification. The first three
355
classes comprise 88% of the 157 subtomograms, and the averages are shown in Figure
356
5C. Interestingly, all three averages show an incomplete B-tubules with strong densities
357
for PFs B1~B3. As a check, we directly visualized the identified partial doublets in their
358
procentriole tomograms (Figure 5D), confirming that they are indeed partially assembled
359
doublets at the longitudinal extremity of the triplet. In summary, based on the
360
classification and the mapping analysis, we identified an intermediate in the B-tubule
16
361
assembly at both the polar ends of triplet. It shows the B-tubule initiates lateral expansion
362
from PF B1 at the outer A-B junction towards the luminal side.
363
364
Intermediate State of the C-tubule during Triplet Assembly
365
We applied a similar analysis to the C-tubule and after two rounds of classification, we
366
identified 208 subtomograms from 110 procentrioles that have partially complete C-
367
tubules (Figure 6-figure supplement 1). In contrast to the incomplete doublet, the average
368
of these 208 subtomograms shows PFs at both the inner and the outer B-C junctions with
369
a gap in between. We further mapped these incomplete triplets in their corresponding
370
tomograms. Similar to the B-tubule result, based on the location these triplets can be
371
divided into five classes. Three major classes, including the proximal end, the distal end
372
and the nascent C-tubule where the entire length of the triplet has partially assembled C-
373
tubule, comprise 83% of the 208 subtomograms. Averages from these three major classes
374
shows a consensus structure of incomplete triplet (Figure 6A), where the PFs C8~C10
375
along with the MIP9 at the inner B-C junction are clearly visible while the PFs C1~C3
376
are visible at the outer junction. Interestingly, in all three averages, even though the C-
377
tubule is incomplete, the A-C linker is visible albeit with weaker density likely due to its
378
flexible nature or incomplete assembly. This indicates that the linkage between
379
neighboring triplets has been established at this early stage of the triplet assembly. We
380
further confirmed these averages by directly visualizing corresponding structure in the
381
procentriole tomograms (Figure 6B). Taken together, our analysis has identified C-tubule
382
assembly intermediates at either end of the triplet longitudinal extremity, suggesting that
383
the C-tubule assembles laterally from both the inner and the outer junction in a bi-
17
384
directional fashion. This allows prompt establishing of the A-C linker during the triplet
385
assembly, reinforcing the procentriole as a barrel-shaped structure.
18
386
387
Discussions
388
389
Here, by using cryoET and image analysis, we present procentriole structures from
390
Chlamydomonas reinhardtii. These new structures and findings have implications
391
concerning the mechanism of centriole biogenesis and assembly.
392
393
Comparing the Chlamydomonas Procentriolar Triplet to Other Triplet Structures
394
As important cell organelles, there have been extensive electron microscopic studies on
395
both the procentriole and the centriole. Most were focused on the morphology of these
396
organelles in the context of their life cycles. In recent years, cryoET has been applied to
397
study molecular architecture of centriole and basal body in several organisms, including
398
Chlamydomonas, Drosophila, Trichonympha and mammalian cells (Greenan et al., 2018;
399
Guichard et al., 2013; Li et al., 2012). These have provided insights into mechanism of
400
triplet/doublet assembly and centriole biogenesis.
401
402
In Chlamydomonas, there are marked differences between the procentriole triplet
403
structure reported here and that in the core region (Li et al., 2012). First, a Y-shaped
404
structure observed in the luminal side of core region is absent in procentriole, instead, the
405
space is taken in part by the pinhead structure. Second, the A-C linker structures in these
406
two regions are different. Whereas the procentriole has a crisscross shaped A-C linker
407
structure that links PFs A6, A9/A10 and C8, C9/C10 from its neighboring triplets, in the
408
core region, the A-C linker bridges PF A6 to various positions on neighboring C-tubule
19
409
depending on the longitudinal position (Li et al., 2012). Third, the PF C1 in the
410
procentriole is composed of tubulin having a 4 nm periodicity, while the C1 in distal half
411
of the core region exhibits an 8 nm repeat, indicating a non-tubulin protein. Lastly, even
412
though a number of MIPs are shared in both regions, for example a cone-shaped structure
413
MIP1 at PF A5 and a number of MIPs spanning the luminal side of PFs A1-A3, the
414
overall patterns of MIPs decorated on the triplet are different in these two regions. This is
415
particularly noticeable in the inner junctions of A-B and B-C tubules. All these
416
differences suggest that two sets of proteins are enlisted sequentially during the centriole
417
assembly, controlled by as yet unknown mode of spatial and temporal regulation. For
418
example, both pinhead and the Y-shaped structure are near the inner junction of A-B
419
tubules. As part of the cartwheel structure, the pinhead function as a scaffold for initial
420
establishment of nine triplets in the procentriole. Once the procentriole is fully
421
assembled, the Y-shaped structure will follow to further strengthen the central core
422
region. Structural difference between these two regions has also been observed recently
423
in mammalian centriole (Greenan et al., 2018), but with details that are distinct from
424
those with Chlamydomonas. However, this does suggest that a multiple-step hierarchical
425
spatial and temporal control is likely a common theme of centriole assembly in many
426
eukaryotes.
427
428
The procentriole triplet structure presented in this work also exhibits several notable
429
differences from the triplets from other organisms. For example, in Trichonympha, the
430
pinhead exhibits 17 nm periodicity, while in Chlamydomonas, it is 8 nm. Interestingly, in
431
both organisms the A-C linker has 8.5 nm or 8 nm periodicity, respectively. This
20
432
structural difference might reflect the difference in mechanism of length control as the
433
procentriole in Trichonympha has an exceptionally long cartwheel.
434
435
The Roles of A-C Linker in the Procentriole Assembly
436
In addition to serving as a structural linkage between neighboring triplets, our structure
437
has revealed several important roles for the A-C linker during procentriole assembly.
438
In a Sas6 null mutation in Chlamydomonas, a circular centriole can still form without the
439
cartwheel, albeit at low efficiency (Nakazawa et al., 2007). Among these cartwheel-less
440
centrioles, the majority remains nine-fold symmetric, but other symmetries have been
441
observed as well, suggesting that, at least in some organism, the cartwheel does not fully
442
account for setting up the nine-fold symmetry in the centriole. Interestingly, in vitro
443
assembly of CrSas6 results in cartwheel-like structures with the majority having eight and
444
nine-fold symmetry (Guichard et al., 2017; Hilbert et al., 2016). Likewise, mammalian
445
cells can form centrioles de novo without the central hub formed by Sas6, although the
446
assembly is less efficient and more error-prone than the wild type (Wang et al., 2015).
447
These observations have led to speculation that other factor(s) might be needed for
448
efficient and robust centriole assembly (Gönczy, 2012; Hirono, 2014). Our finding that
449
the A-C tubule linkage is established at an early stage of procentriole assembly while the
450
triplets are still elongating provides direct evidence supporting a model that the A-C
451
linker helps establish the 9-fold symmetry. In our model, A-C linker can accomplish this
452
simply by restricting the angle between neighboring MT triplets to be approximately 140
453
degree as needed for forming a nonagon. Thus in wild type cells, the combination of both
454
Sas6 oligomerization and a stabilizing linkage between neighboring triplets that has a
21
455
defined angle will provide a robust mechanism for establishing the nine-fold symmetry.
456
In addition to restricting the angle between neighboring triplets, the A-C linker also plays
457
a structural role that connects nascent incomplete triplets therefore bolstering the nascent
458
procentriole structure during its assembly.
459
460
Instead of being a static structure, the A-C linker displays substantial twist motion along
461
the triplet wall. This can accommodate an apparent left-handed iris diaphragm movement
462
that progresses along the longitudinal direction. The twist observed both in the
463
procentriole and in the core region results in change of angle for each triplet relative to
464
the cylindrical circumference, from the most acute angle at the very proximal side
465
towards more tangential in the distal region. As a result, the inner diameter of the
466
centriole gradually increases. This change may facilitate multiple processes: First, the
467
acute angle of the triplet at the proximal end places the A-tubule including the pinhead
468
close to the cartwheel, likely reflecting its attachment as the earliest stages of biogenesis.
469
Second, the gradual twisting will increase the exposed surface area, facilitating
470
recruitment of other centriole components, such as the subdistal and the distal
471
appendages. Third, the tangential angle of the triplet at the distal end of basal body must
472
match the angle of the axoneme doublets, where the basal body is ready to serve as a
473
template for the flagellum assembly. Lastly, this iris diaphragm motion of the MT triplet
474
might also have impact on controlling the procentriole length by twisting the pinheads
475
away from the hub while the luminal diameter becomes wider. This suggests that the
476
radial spokes will be under strain of pulling force that continually increases along the
477
procentriole. This pinhead movement can be directly visualized in Video 5. As part of the
22
478
radial spoke, the C-terminus of Sas6 and Cep135/Bld10 are predicted to have highly
479
coiled-coil structures with intrinsically elastic properties. It is appealing to envision that
480
this strain and elasticity will reach to a balance point once the MT triplets elongate to
481
certain length, suggesting an intrinsic length control mechanism due to the iris diaphragm
482
motion of the triplet.
483
484
A Model for the Procentriole Triplet Assembly
485
Previous EM studies on mammalian centriole and basal body showed the triplet assembly
486
was a non-uniform process where the A-tubule assembled first, followed by sequential B-
487
and C-tubule assembly. During the tubule elongation, their lengths varied substantially
488
(Anderson and Brenner, 1971; Guichard et al., 2010). Furthermore, by analyzing the
489
nascent procentrioles, Anderson and Brenner observed the B- and C-tubule first emerged
490
as sheets laterally attached to the A- and B-tubule respectively. Consistent with these
491
findings, by analyzing a population of procentrioles, we found the average length of three
492
tubules in the triplet varies substantially. In addition, we have identified two triplet
493
intermediates, showing that the B-tubule initiates assembly uni-directionally from the
494
outer A-B junction. By contrast, the C-tubule assembles laterally in a bi-directional
495
fashion, from both the inner and the outer B-C junctions. These lateral expansions are
496
concomitant with the longitudinal elongation of the triplets at their extremities. Based on
497
the analysis, we propose a possible model for the triplet assembly during the procentriole
498
biogenesis (Figure 7). In our model, the A-tubule first emerges and anchors to the
499
pinhead of cartwheel. It continues to grow. Then the B-tubule initially assembles at the
500
outer A-B junction by making an unusual tubulin-tubulin lateral interface and branching
23
501
out at the PF A10. This branch laterally expands from the outer junction to reach the
502
luminal inner junction where the PF B10 and other components including the MIPs 5~6
503
seal and complete the B-tubule. Unlike the B-tubule, the C-tubule initiates assembly on
504
both junctions. PFs C1 and C10 branch out from the B-tubule at the inner and the outer
505
junctions, respectively. These two branches expand laterally in a bi-directional fashion.
506
At this point, while the C-tubule is still incomplete, the newly assembled PFs C8~C10 are
507
readily available to provide a site for the A-C linker that connects the C-tubule to the A-
508
tubule of its (N-1) neighbor triplet. Equivalently, its A-tubule will connect to the C-tubule
509
of its (N+1) neighbor triplet. As a result, all nine triplets form an inter-connected barrel
510
even though some triplets are still incomplete. In the last step, the gap is filled by the PFs
511
in the C-tubule. This completes the assembly of all three tubules at this longitudinal
512
point, while the lateral expansion and the longitudinal elongation will continue in both
513
extremities of the triplet.
514
515
Unlike the rapid kinetics of the MT self-assembly, the dynamics of growing MT triplets
516
in both procentriole and centriole are slow but stable compared to the MT (Kochanski
517
and Borisy, 1990). Although we cannot exclude the possibility that there is a fraction of
518
partial B- or C-tubules that is due to their lability during the purification process, based
519
on the observed tubule length variation, the morphology of the MT ends, the stability of
520
centriole MT compared to the canonical MT and the consistent pattern of the partial B-
521
and C-tubule observed at the growing ends (Figures 5C-D, 6A-B), the identified
522
intermediate structures likely represent two snapshots of an underpinning mechanism for
523
the MT triplet assembly, which is tightly controlled by a set of factors. These include
24
524
structural components such as MIPs identified in this work or other microtubule
525
associated proteins (MAPs) (Sharma et al., 2016; Zheng et al., 2016). In addition, other
526
regulatory components such as, acetyltransferase or glutamylase will alter the
527
polymerization properties and stability of the MT triplet as well (Yu et al., 2015).
528
Interestingly, many MIPs identified in this work and in mammalian centrioles (Greenan
529
et al., 2018) are localized at the junction of A-B or B-C tubules in the triplet. For
530
example, MIPs 5~6 are found at the inner A-B junction, while MIPs 9~11 are at the inner
531
B-C tubules junction. This asymmetric localization of the MIPs will likely modulate the
532
assembly kinetics of the MT PFs. Furthermore, since the association of MIPs to triplet
533
might be concomitant with the MT elongation, the MIPs’ own assembly kinetics on MT
534
will in turn have directly impact on the assembly of the triplet. In the future, with
535
enhanced dataset size and improved algorithm, it is likely that more assembly
536
intermediates will be identified in Chlamydomonas and in other organisms. This will
537
provide insight into the centriole assembly mechanism with improved spatial and
538
temporal resolution. Our study presented here has paved the way in this direction.
539
540
Finally, like many large and functional biological assemblages, the procentriole structure
541
presented in this work is an example demonstrating the extensive diversity and
542
heterogeneity of a biological system in situ, including composition heterogeneity such as
543
the incomplete B- and C-tubules in triplet and conformation heterogeneity such as the
544
continuous twisting of the A-C linker. CryoET is well poised as a unique and
545
unparalleled tool to reveal this phenomenon and to provide high-resolution insights for
546
many biological mechanisms waiting to be explored.
547
25
548
549
550
Acknowledgement
551
collection, Matt Harrington for supporting the computational infrastructure, Hiroaki
552
Ishikawa for discussion on the biology of Chlamydomonas, centrioles and basal bodies,
553
Tom Goddart for help with UCSF Chimera program. We are grateful to many of our
554
colleagues for critical reading of the manuscript and for the encouragements. This work is
555
supported in part by NIH grants GM031627 (D.A.A.), GM118099 (D.A.A.), PO1
556
GM105537 (D.A.A.), GM113602 (W.F.M.) and by HHMI (D.A.A.) and by the Spanish
557
AEI/FEDER (SAF2017-84565-R) (J.-J.F.) and by Fundacion Ramon Areces (J.-J.F.).
We thank Michael Braunfeld and Shawn Zheng for advice on tomography data
558
559
The EM structures have been deposited in the EMDB with the following accession
560
numbers: EMD-9167, EMD-9168, EMD-9169, EMD-9170, EMD-9171, EMD-9172,
561
EMD-9173, EMD-9174.
562
26
563
564
Declaration of Interests
565
The authors declare no competing interests.
566
567
568
569
27
570
Figure Legends
571
572
Figure 1. Tomographic Reconstruction and Subtomogram Averaging of the MT
573
Triplet in Centriole and Procentriole
574
(A) A slice from the reconstructed tomogram showing a pair of mother centrioles, also
575
known as basal bodies. The central hub as part of the cartwheel is marked by a yellow
576
asterisk. The centriole is partitioned longitudinally into three regions. The procentriole
577
region spans 100 nm at the proximal end of the mother centriole. The central core region
578
spans 250 nm. The distal region spans 150 nm where the triplets become doublets before
579
reaching the transition zone (TZ). PSF: proximal striated fibers; DSF: distal striated
580
fibers; rMT: rootlet MT.
581
(B) A slice from the reconstructed tomogram of procentriole attached to the mother
582
centriole via the rootlet MT. The central hub, radial spokes and pinheads are clearly
583
visible, as well as protofilaments in some MT triplets. RS: radial spoke; PH: pinhead;
584
rMT: rootlet MT.
585
(C) Subtomogram average of the MT triplet from both the proximal region of mother
586
centrioles and the procentrioles.
587
(D) A model of procentriole viewed from its distal end. The model is generated by
588
docking the subtomograms average in (C) into the tomogram volume in (B). The A- B-
589
and C-tubules and the A-C linker are labeled. The central hub and the radial spokes are
590
depicted schematically. PH: pinhead.
591
592
Figure 2. Identifying Non-tubulin Components in Procentriole MT Triplet
593
(A) The A-tubule structure. A model of 13-pf MT (red) are docked into the subtomogram
594
average of the A-tubule. 4 MIPs identified on the A-tubule are highlighted in colors.
595
(B) The B-tubule structure. A model of 10-pf MT (blue) are docked into the
596
subtomogram average of the B-tubule. 4 MIPs identified on the B-tubule are highlighted
597
in colors.
598
(C) The C-tubule structure. A model of 10-pf MT (green) are docked into the
599
subtomogram average of the C-tubule. 3 MIPs found at the B-C inner junction are
600
highlighted in colors.
28
601
(D) Subtomogram averaging of the pinhead structure. The pinhead is highlighted in pink
602
color. Two dashed arrows indicate viewing directions for two images in the side view on
603
the right. The arrows indicate PinB, PinF1 and PinF2, respectively. The arrowheads mark
604
the tip of PinB.
605
606
Figure 3. The Structure of A-C Linker
607
(A) A top view of the A-C linker structure with its linked A- and C-tubules. The A-C
608
linker is highlighted in red. Three filamental structures f1, f2 and f3 making up the
609
central trunk running in longitudinal direction are marked with *.
610
(B) The A-C linker viewed from outside of procentriole. The dashed square highlights the
611
area that will be viewed in (D).
612
(C) The A-C linker viewed from the luminal side of procentriole. Three filaments f1, f2
613
and f3 that make up the central trunk are marked with vertical dashed lines. They are
614
interconnected and form a spiral.
615
(D and E) Close up views of the Leg B in the A-C linker show a possible location of
616
POC1. (D) is in side view from outside of the procentriole. (E) is in top view from the
617
distal end of the procentriole. An atomic model of WD40 β-propeller domain (PDB ID
618
1S4U) is fitted into the doughnut-shaped density in Leg B. The red * indicate connecting
619
points of this WD40 domain to the rest of Leg B structure and to the f3 in central trunk.
620
621
Figure 4. The Twist of A-C Linker Drives Procentriole Triplets in Iris Diaphragm
622
Motion
623
(A) Schematic diagram of two triplets connected by the A-C linker. θ is defined as the
624
angle between the Arm A and the Leg B. The results of mapping and angle measurement
625
as shown in (B-C, Table 3) are consistent with a twisting motion of two triplets as
626
indicated by two curved arrows when moving longitudinally from proximal to distal
627
direction. The pivot point of this twist is at the A-C linker as indicated by an arrow.
628
(B and C) Weighted averages of twist angle along the triplet wall in two classification
629
schemes shown in Figure 3-figure supplements 1,3. Six points along 90 nm longitudinal
630
length of centriole are sampled starting from the proximal end. The weighted average of
631
twist angle T at point i is defined as: Ti = ∑ θj*Nij / ∑ Nij
29
632
Nij: number of subtomogram belong to class j at point i.
633
θj: twist angle as defined in (A) for class j.
634
(D and E) Overlay of two sets A-C linker structure based on two classification schemes
635
(Figure 3-figure supplements 1,3). Based on its longitudinal position, each structure is
636
colored following a “heat map” scheme. Arrows point to the pivot point, the central
637
trunk.
638
639
Figure 5. Variation of the Tubule Length in Procentriole Triplets and Detecting the
640
B-tubule Intermediate.
641
(A) Two examples of procentriole in tomogram. In each, two triplets marked by yellow
642
rectangle box are selected to depict tubule length variation as shown on the right in the
643
longitudinal view. The distal end of each tubule is indicated by an arrowhead. Black
644
arrows indicate examples of the slightly flared morphology at the growing end of the
645
tubule. Scale bar: 100 nm.
646
(B) Histogram showing length distribution for A-, B- and C-tubules. For each tubule, the
647
average length, its standard deviation and the number of measurement are indicated.
648
(C) The averages of incomplete B-tubule from the proximal, distal and nascent doublet,
649
respectively. Each image is a z-projection of doublet of 9.6 nm long. The numbers of
650
subtomogram in each class and their percentages are indicated. The arrowhead indicates
651
PF B1.
652
(D) Examples of incomplete B-tubule in procentriole tomogram from three classes. The
653
incomplete B-tubules are marked by yellow arrowheads.
654
655
Figure 6. Detecting the C-tubule Intermediate
656
(A) The averages of subset of subtomogram with incomplete triplet at proximal, distal
657
end and from nascent triplet. The images are z-projection of the average where the triplet
658
length is 9.6 nm. The numbers of subtomogram in each class and their percentages are
659
indicated. The arrowhead points to PF C1. The arrow points to the A-C linker.
660
(B) Examples of incomplete C-tubule in procentriole tomogram in three classes. The
661
incomplete C-tubules are marked by yellow arrowheads. The A-C linkers are marked by
662
yellow arrows.
30
663
664
Figure 7. A Model for the MT Triplet Assembly in the Procentriole
665
In the model, the triplet assembly can be divided into 5 sequential steps: 1) A-tubule
666
emerges and anchors to the pinhead in the cartwheel. It continues to elongate
667
longitudinally. 2) B-tubule branches out at PF A10 in outer A-B junction. 3) B-tubule
668
expands laterally from outer A-B junction toward the luminal inner A-B junction where it
669
completes the B-tubule. 4) C-tubule branches out at both inner and outer B-C junctions.
670
Meanwhile, the A-C linker is established. 5) The bi-directional lateral expansion of C-
671
tubule completes the triplet assembly at this longitudinal position.
672
673
674
31
675
Supplemental Figure Legends
676
Figure 1-figure supplement 1 Assessing the Resolutions of Subtomogram Average by
677
Fourier Shell Correlation (FSC) Method.
678
FSC curves are for the MT triplet (black, 23.0 Å), the A-tubule (red, 21.4 Å), the B-tubule
679
(blue, 22.2 Å) and the C-tubule (green, 22.2 Å). All resolution assessments are carried out
680
in a “gold-standard” Scheme using FSC 0.143 criterion.
681
682
Figure 1-figure supplement 2 Comparing Subtomogram Averages from a Subset of
683
Data.
684
(A) The triplet in blue on left is an average based on 10854 subtomograms from the
685
proximal end of mother centrioles.
686
(B) The triplet in light green on right is an average based on 2083 subtomograms from
687
the 110 procentrioles that remain attached to the mother centrioles.
688
The differences between these two structures are indicated by red arrows and by red
689
asterisks, respectively.
690
691
Figure 2-figure supplements 1-11 The non-tubulin Procentriole Components Associated
692
to the MT Triplet.
693
The figures show the binding pattern of 11 MIPs and their periodicity along the MT wall.
694
In each figure, image on the left is a longitudinal cross section of the average of
695
corresponding tubule. The middle image is a surface-rendered map with the structure of
696
corresponding MIP highlighted in color. Image on the right is a cartoon model of the
697
tubule. The dashed line indicates the location where the cross section goes through. The
698
red arrow indicates viewing direction of the structure shown in the middle image. They
699
are also summarized in Table 2.
700
701
Figure 2-figure supplements 12. MIP9 forms a non-tubulin filament at the inner
702
junction of B- and C-tubule.
703
The MIP9 density is highlighted in blue. A model of MT protofilament in green is fit into
704
the MIP9 density. On the right panel, only the MIP9 filament is shown in two orthogonal
32
705
views, the rest triplet structure is removed for clarity. The MIP9 density differs
706
substantially from a canonical MY protofilament.
707
708
Figure 2-figure supplements 13. The anisotropic resolution in the pinhead structure.
709
The distal tip of PinB is flexible exhibiting the lowest resolution in the average
710
(arrowhead). The local resolution is estimated by blocres in Bsoft.
711
712
Figure 2-figure supplements 14. FSC curve for the averaged pinhead structure after
713
classification (resolution: 23.1 Å).
714
715
Figure 3-figure supplement 1. Schematic diagram shows the classification process for
716
identifying intact A-C linker associated to the A-tubule.
717
The number of subtomogram in each class and the 4 classes identified with intact A-C
718
linker are indicated. Total 3992 subtomograms from 4 classes are identified with relative
719
intact A-C linker. Among them, 2917 (73.1%) are from the proximal end of mother
720
centrioles. 1075 (26.9%) are from the procentrioles.
721
722
Figure 3-figure supplement 2. Subtomogram average of the A-tubule with associated A-
723
C linker.
724
This is based on 3992 subtomograms (Class 1-4) identified in classification scheme in
725
Figure 3-figure supplement 1.
726
727
Figure 3-figure supplement 3. Schematic diagram showing classification process for
728
identifying intact A-C linker associated to the C-tubule.
729
The number of subtomogram in each class and the 6 classes identified with relative intact
730
A-C linker are indicated. The Class 1 is not used for further refinement and average due
731
to its low quality. Total 3478 subtomograms from 6 classes are identified with relative
732
intact A-C linker. Among them, 2611 (75.1%) are from the proximal end of mother
733
centrioles. 867 (24.9%) are from the procentrioles.
734
33
735
Figure 3-figure supplement 4. Subtomogram average of the C-tubule with associated A-
736
C linker.
737
This is based on 3245 subtomograms (Class 2-6) identified in classification scheme in
738
Figure 3-figure supplement 3.
739
740
Figure 3-figure supplement 5. FSC curves for the averaged A-tubule and the C-tubule
741
with more intact A-C linker after classification.
742
The A-tubule with associated A-C linker (red, 23.1 Å), the C-tubule with associated A-C
743
linker (green, 23.1 Å).
744
745
Figure 3-figure supplement 6. The scheme for Reconstruction of Full A-C Linker
746
Structure.
747
By using this scheme, based on the 4 and 6 classes identified (illustrated in Figure 3-
748
figure supplements 1,3), total 10 structures with full A-C linker are reconstructed
749
showing different degrees of the twist motion. The Class 2 (568 subtomograms)
750
identified in the second classification scheme (Figure 3-figure supplement 3) is used as
751
an example shown in Figure 3.
752
753
Figure 5-figure supplement 1. Classification and Identification of Partially Assembled
754
B-tubule.
755
Total 157 subtomograms with incomplete doublet are identified as the result of
756
classification. Images displayed are the central z-sections from the average of each class.
757
The number of subtomogram in each class is indicated.
758
759
Figure 6-figure supplement 1. Classification and Identification of Partially Assembled
760
C-tubule.
761
Total 208 subtomograms with incomplete triplet are identified as the result of
762
classification. Images displayed are the central z-sections from the average of each class.
763
The number of subtomogram in each class is indicated.
764
765
766
34
767
768
769
770
Tables
Table 1. Summary of Statistics of Subtomogram Averages
Structure
Number of
subtomogram
Resolution
(Å)
Description
EMDB
ID #
1
12937
23.0
Triplet structure
9167
2
12517
21.4
A-tubule
9168
3
12179
22.3
B-tubule
9169
4
12075
22.3
C-tubule
9170
5
4763
23.1
Pinhead structure with its
associated partial A-tubule
9171
6
3992
23.1
A-tubule with more complete A-C
linker after classification
9172
7
3245
23.1
C-tubule with more complete A-C
linker after classification
9173
Full A-C linker structure with Aand C-tubule it links to
9174
8*
771
772
773
774
775
776
Composite map derived from
EMD-9172 and EMD-9173
All resolutions are assessed by using the “gold standard” scheme with FSC 0.143
criterion. * This is a composite map by merging EMD-9172 and EMD-9173 onto the map
of one of classes, class#2 (See Figure 3-figure supplement 6)
35
777
778
Table 2. Estimated Molecular Weight of MIPs associated to Triple
Name
779
780
781
782
783
784
Estimated MW Periodicity
(KD)
Location
MIP1
45
8 nm
cone-like structure associated to the lumen side
of A5
MIP2
n/a
4 nm
trellis-like structure spanning laterally from A11
to A13
MIP3
24
8 nm
laterally link A13, A1 and A2 in the lumen of Atubule
MIP4
28
8 nm
laterally link A2 and A3 in the lumen of A-tubule
MIP5
†
74
8 nm
inner junction of A, B-tubule, laterally link A1, A2
and B10
MIP6
27
8 nm
inner junction of A, B-tubule, laterally link A1,
A13, B9, B10
MIP7
n/a
8 nm
fin-like filamental structures running
longitudinally along B5, B4 B3 in the lumen of Btubule
MIP8
15
8 nm
outside of B-tubule but in the lumen of C-tubule,
laterally link B6 and B7
MIP9
92
8 nm
inner junction of B, C-tubule, link B8, B9 and
C10
MIP10
19
8 nm
inner junction of B, C-tubule, link B7 and C10
MIP11
6
8 nm
laterally crosslink C9 and C10 in the lumen of Ctubule
Protein density = 0.849 Dalton/Å3
Orange: MIPs associated to the A-tubule of triplet
Blue: MIPs associated to the B-tubule of triplet
Green: MIPs associated to the C-tubule of triplet
36
785
786
787
Table 3.
A. Measurement of Twist Angle at A-C Linker
Class θ Angle (degree)
788
789
790
791
792
793
794
795
I
93
II
90
III
87
IV
83
The assignment for each class is following convention defined in Figures 4D and it is
based on the resulting 4 classes in the first classification scheme, illustrated in Figure 3figure supplement 1. In the scheme, the A-tubule is used as a reference point.
The angle θ is defined as the angle between the Arm A and the Leg B as illustrated in
Figure 4A.
B. Measurement of Twist Angle at A-C Linker
Class θ Angle (degree)
796
797
798
799
800
801
802
803
I
96
II
93
III
90
IV
89
V
84
VI
80
The assignment for each class is following convention defined in Figures 4E and it is
based on the resulting 6 classes in the second classification scheme, illustrated in Figure 3figure supplement 3. In the scheme, the C-tubule is used as a reference point.
The angle θ is defined as the angle between the Arm A and the Leg B as illustrated in
Figure 4A.
37
804
Video 1 An Aligned Tomography Tilt Series of NFAp Showing a Pair of Mother
805
Centriole and the Attached Procentrioles.
806
807
Video 2 Surface Rendered Structure of the A-C Linker.
808
809
Video 3 Swing Motion of the A-C Linker and its Linked C-tubule. The A-tubule on the
810
left is static and is used as a reference point to show the motion.
811
812
Video 4 Swing Motion of the A-C Linker and its Linked A-tubule. The C-tubule on the
813
right is static and is used as a reference point to show the motion.
814
815
Video 5 An Iris-Diaphragm Motion of the Procentriole Triplets.. A longitudinal segment
816
of the procentriole is generated by putting the averaged triplet structure back into one of
817
the procentriole tomograms. The movie is displayed as moving through the longitudinal
818
direction from the proximal to the distal end then backwards. The triplets twist
819
progressively along the procentriole in a left-handed chirality with the thumb pointing
820
towards the distal end. Two yellow arrows mark the two A-C linkers. They appear at the
821
distal end of the longitudinal segment.
822
823
38
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
References:
Andersen, J.S., C.J. Wilkinson, T. Mayor, P. Mortensen, E.A. Nigg, and M. Mann. 2003.
Proteomic characterization of the human centrosome by protein correlation
profiling. Nature. 426:570-574.
Anderson, R.G. 1972. The three-dimensional structure of the basal body from the rhesus
monkey oviduct. J. Cell Biol. 54:246-265.
Anderson, R.G., and R.M. Brenner. 1971. The formation of basal bodies (centrioles) in
the Rhesus monkey oviduct. J. Cell Biol. 50:10-34.
Banterle, N., and P. Gonczy. 2017. Centriole Biogenesis: From Identifying the Characters
to Understanding the Plot. Annu. Rev. Cell. Dev. Biol. 33:23-49.
Beck, B.B., J.B. Phillips, M.P. Bartram, J. Wegner, M. Thoenes, A. Pannes, J. Sampson,
R. Heller, H. Gobel, F. Koerber, A. Neugebauer, A. Hedergott, G. Nurnberg, P.
Nurnberg, H. Thiele, J. Altmuller, M.R. Toliat, S. Staubach, K.M. Boycott, E.M.
Valente, A.R. Janecke, T. Eisenberger, C. Bergmann, L. Tebbe, Y. Wang, Y. Wu,
A.M. Fry, M. Westerfield, U. Wolfrum, and H.J. Bolz. 2014. Mutation of POC1B
in a severe syndromic retinal ciliopathy. Hum. Mutat. 35:1153-1162.
Carvalho-Santos, Z., P. Machado, P. Branco, F. Tavares-Cadete, A. Rodrigues-Martins,
J.B. Pereira-Leal, and M. Bettencourt-Dias. 2010. Stepwise evolution of the
centriole-assembly pathway. J. Cell Sci. 123:1414-1426.
Cavalier-Smith, T. 1974. Basal body and flagellar development during the vegetative cell
cycle and the sexual cycle of Chlamydomonas reinhardii. J. Cell Sci. 16:529-556.
Dutcher, S.K., N.S. Morrissette, A.M. Preble, C. Rackley, and J. Stanga. 2002. Epsilontubulin is an essential component of the centriole. Mol. Biol. Cell. 13:3859-3869.
Dutcher, S.K., and E.C. Trabuco. 1998. The UNI3 gene is required for assembly of basal
bodies of Chlamydomonas and encodes delta-tubulin, a new member of the
tubulin superfamily. Mol. Biol. Cell. 9:1293-1308.
Geimer, S., and M. Melkonian. 2004. The ultrastructure of the Chlamydomonas
reinhardtii basal apparatus: identification of an early marker of radial asymmetry
inherent in the basal body. J. Cell Sci. 117:2663-2674.
Gönczy, P. 2012. Towards a molecular architecture of centriole assembly. Nature reviews
Molecular cell biology.
Greenan, G.A., B. Keszthelyi, R.D. Vale, and D.A. Agard. 2018. Insights into centriole
geometry revealed by cryoTomography of doublet and triplet centrioles. Elife. 7.
Guichard, P., D. Chrétien, S. Marco, and A.-M. Tassin. 2010. Procentriole assembly
revealed by cryo-electron tomography. The EMBO journal. 29:1565-1572.
Guichard, P., A. Desfosses, A. Maheshwari, V. Hachet, C. Dietrich, A. Brune, T.
Ishikawa, C. Sachse, and P. Gonczy. 2012. Cartwheel architecture of
Trichonympha basal body. Science. 337:553.
Guichard, P., V. Hachet, N. Majubu, A. Neves, D. Demurtas, N. Olieric, I. Fluckiger, A.
Yamada, K. Kihara, Y. Nishida, S. Moriya, M.O. Steinmetz, Y. Hongoh, and P.
Gonczy. 2013. Native architecture of the centriole proximal region reveals
features underlying its 9-fold radial symmetry. Curr. Biol. 23:1620-1628.
Guichard, P., V. Hamel, and P. Gonczy. 2018. The Rise of the Cartwheel: Seeding the
Centriole Organelle. Bioessays. 40:e1700241.
39
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
Guichard, P., V. Hamel, M. Le Guennec, N. Banterle, I. Iacovache, V. Nemcikova, I.
Fluckiger, K.N. Goldie, H. Stahlberg, D. Levy, B. Zuber, and P. Gonczy. 2017.
Cell-free reconstitution reveals centriole cartwheel assembly mechanisms. Nat
Commun. 8:14813.
Hilbert, M., A. Noga, D. Frey, V. Hamel, P. Guichard, S.H. Kraatz, M. Pfreundschuh, S.
Hosner, I. Fluckiger, R. Jaussi, M.M. Wieser, K.M. Thieltges, X. Deupi, D.J.
Muller, R.A. Kammerer, P. Gonczy, M. Hirono, and M.O. Steinmetz. 2016. SAS6 engineering reveals interdependence between cartwheel and microtubules in
determining centriole architecture. Nat. Cell Biol. 18:393-403.
Hiraki, M., Y. Nakazawa, R. Kamiya, and M. Hirono. 2007. Bld10p constitutes the
cartwheel-spoke tip and stabilizes the 9-fold symmetry of the centriole. Current
biology : CB. 17:1778-1783.
Hirono, M. 2014. Cartwheel assembly. Philos Trans R Soc Lond B Biol Sci. 369.
Hodges, M.E., N. Scheumann, B. Wickstead, J.A. Langdale, and K. Gull. 2010.
Reconstructing the evolutionary history of the centriole from protein components.
J. Cell Sci. 123:1407-1413.
Hoog, J.L., S.M. Huisman, Z. Sebo-Lemke, L. Sandblad, J.R. McIntosh, C. Antony, and
D. Brunner. 2011. Electron tomography reveals a flared morphology on growing
microtubule ends. J. Cell Sci. 124:693-698.
Keller, L.C., S. Geimer, E. Romijn, J. Yates, I. Zamora, and W.F. Marshall. 2009.
Molecular architecture of the centriole proteome: the conserved WD40 domain
protein POC1 is required for centriole duplication and length control. Mol. Biol.
Cell. 20:1150-1166.
Keller, L.C., E.P. Romijn, I. Zamora, J.R. Yates, and W.F. Marshall. 2005. Proteomic
analysis of isolated chlamydomonas centrioles reveals orthologs of ciliary-disease
genes. Current biology : CB. 15:1090-1098.
Khire, A., K.H. Jo, D. Kong, T. Akhshi, S. Blachon, A.R. Cekic, S. Hynek, A. Ha, J.
Loncarek, V. Mennella, and T. Avidor-Reiss. 2016. Centriole Remodeling during
Spermiogenesis in Drosophila. Curr. Biol. 26:3183-3189.
Kilburn, C.L., C.G. Pearson, E.P. Romijn, J.B. Meehl, T.H. Giddings, B.P. Culver, J.R.
Yates, and M. Winey. 2007. New Tetrahymena basal body protein components
identify basal body domain structure. J. Cell Biol. 178:905-912.
Kitagawa, D., I. Vakonakis, N. Olieric, M. Hilbert, D. Keller, V. Olieric, M. Bortfeld,
M.C. Erat, I. Fluckiger, P. Gonczy, and M.O. Steinmetz. 2011. Structural basis of
the 9-fold symmetry of centrioles. Cell. 144:364-375.
Kochanski, R.S., and G.G. Borisy. 1990. Mode of centriole duplication and distribution.
J. Cell Biol. 110:1599-1605.
Kong, D., V. Farmer, A. Shukla, J. James, R. Gruskin, S. Kiriyama, and J. Loncarek.
2014. Centriole maturation requires regulated Plk1 activity during two
consecutive cell cycles. J. Cell Biol. 206:855-865.
Li, J.B., J.M. Gerdes, C.J. Haycraft, Y. Fan, T.M. Teslovich, H. May-Simera, H. Li, O.E.
Blacque, L. Li, C.C. Leitch, R.A. Lewis, J.S. Green, P.S. Parfrey, M.R. Leroux,
W.S. Davidson, P.L. Beales, L.M. Guay-Woodford, B.K. Yoder, G.D. Stormo, N.
Katsanis, and S.K. Dutcher. 2004. Comparative genomics identifies a flagellar
and basal body proteome that includes the BBS5 human disease gene. Cell.
117:541-552.
40
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
Li, S., J.J. Fernandez, W.F. Marshall, and D.A. Agard. 2012. Three-dimensional structure
of basal body triplet revealed by electron cryo-tomography. EMBO J. 31:552-562.
Matsuura, K., P.A. Lefebvre, R. Kamiya, and M. Hirono. 2004. Bld10p, a novel protein
essential for basal body assembly in Chlamydomonas: localization to the
cartwheel, the first ninefold symmetrical structure appearing during assembly.
165:663-671.
Meehl, J.B., B.A. Bayless, T.H. Giddings, Jr., C.G. Pearson, and M. Winey. 2016.
Tetrahymena Poc1 ensures proper intertriplet microtubule linkages to maintain
basal body integrity. Mol. Biol. Cell. 27:2394-2403.
Muller, H., D. Schmidt, S. Steinbrink, E. Mirgorodskaya, V. Lehmann, K. Habermann, F.
Dreher, N. Gustavsson, T. Kessler, H. Lehrach, R. Herwig, J. Gobom, A.
Ploubidou, M. Boutros, and B.M. Lange. 2010. Proteomic and functional analysis
of the mitotic Drosophila centrosome. EMBO J. 29:3344-3357.
Nakazawa, Y., M. Hiraki, R. Kamiya, and M. Hirono. 2007. SAS-6 is a Cartwheel
Protein that Establishes the 9-Fold Symmetry of the Centriole. Curr. Biol.
17:2169-2174.
Nigg, E.A., and J.W. Raff. 2009. Centrioles, centrosomes, and cilia in health and disease.
Cell. 139:663-678.
O'Toole, E.T., and S.K. Dutcher. 2014. Site-specific basal body duplication in
Chlamydomonas. Cytoskeleton (Hoboken). 71:108-118.
Paintrand, M., M. Moudjou, H. Delacroix, and M. Bornens. 1992. Centrosome
organization and centriole architecture: their sensitivity to divalent cations. J.
Struct. Biol. 108:107-128.
Pearson, C.G., T.H. Giddings, and M. Winey. 2009a. Basal body components exhibit
differential protein dynamics during nascent basal body assembly. Mol. Biol. Cell.
20:904-914.
Pearson, C.G., D.P.S. Osborn, T.H. Giddings, P.L. Beales, and M. Winey. 2009b. Basal
body stability and ciliogenesis requires the conserved component Poc1. J. Cell
Biol. 187:905-920.
Reiter, J.F., and M.R. Leroux. 2017. Genes and molecular pathways underpinning
ciliopathies. Nat. Rev. Mol. Cell Biol. 18:533-547.
Roosing, S., I.J. Lamers, E. de Vrieze, L.I. van den Born, S. Lambertus, H.H. Arts, P.B.S.
Group, T.A. Peters, C.B. Hoyng, H. Kremer, L. Hetterschijt, S.J. Letteboer, E.
van Wijk, R. Roepman, A.I. den Hollander, and F.P. Cremers. 2014. Disruption of
the basal body protein POC1B results in autosomal-recessive cone-rod dystrophy.
Am. J. Hum. Genet. 95:131-142.
Sarig, O., S. Nahum, D. Rapaport, A. Ishida-Yamamoto, D. Fuchs-Telem, L. Qiaoli, K.
Cohen-Katsenelson, R. Spiegel, J. Nousbeck, S. Israeli, Z.U. Borochowitz, G.
Padalon-Brauch, J. Uitto, M. Horowitz, S. Shalev, and E. Sprecher. 2012. Short
stature, onychodysplasia, facial dysmorphism, and hypotrichosis syndrome is
caused by a POC1A mutation. Am. J. Hum. Genet. 91:337-342.
Scheres, S.H. 2016. Processing of Structurally Heterogeneous Cryo-EM Data in
RELION. Methods Enzymol. 579:125-157.
Shaheen, R., E. Faqeih, H.E. Shamseldin, R.R. Noche, A. Sunker, M.J. Alshammari, T.
Al-Sheddi, N. Adly, M.S. Al-Dosari, S.G. Megason, M. Al-Husain, F. AlMohanna, and F.S. Alkuraya. 2012. POC1A truncation mutation causes a
41
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
ciliopathy in humans characterized by primordial dwarfism. Am. J. Hum. Genet.
91:330-336.
Sharma, A., A. Aher, N.J. Dynes, D. Frey, E.A. Katrukha, R. Jaussi, I. Grigoriev, M.
Croisier, R.A. Kammerer, A. Akhmanova, P. Gonczy, and M.O. Steinmetz. 2016.
Centriolar CPAP/SAS-4 Imparts Slow Processive Microtubule Growth. Dev. Cell.
37:362-376.
Stirnimann, C.U., E. Petsalaki, R.B. Russell, and C.W. Muller. 2010. WD40 proteins
propel cellular networks. Trends Biochem. Sci. 35:565-574.
van Breugel, M., M. Hirono, A. Andreeva, H.A. Yanagisawa, S. Yamaguchi, Y.
Nakazawa, N. Morgner, M. Petrovich, I.O. Ebong, C.V. Robinson, C.M. Johnson,
D. Veprintsev, and B. Zuber. 2011. Structures of SAS-6 suggest its organization
in centrioles. Science. 331:1196-1199.
Venoux, M., X. Tait, R.S. Hames, K.R. Straatman, H.R. Woodland, and A.M. Fry. 2013.
Poc1A and Poc1B act together in human cells to ensure centriole integrity. J. Cell
Sci. 126:163-175.
Wang, W.J., D. Acehan, C.H. Kao, W.N. Jane, K. Uryu, and M.F. Tsou. 2015. De novo
centriole formation in human cells is error-prone and does not require SAS-6 selfassembly. Elife. 4.
Woodland, H.R., and A.M. Fry. 2008. Pix proteins and the evolution of centrioles. PLoS
One. 3:e3778.
Yu, I., C.P. Garnham, and A. Roll-Mecak. 2015. Writing and Reading the Tubulin Code.
J. Biol. Chem. 290:17163-17172.
Zheng, X., A. Ramani, K. Soni, M. Gottardo, S. Zheng, L. Ming Gooi, W. Li, S. Feng, A.
Mariappan, A. Wason, P. Widlund, A. Pozniakovsky, I. Poser, H. Deng, G. Ou,
M. Riparbelli, C. Giuliano, A.A. Hyman, M. Sattler, J. Gopalakrishnan, and H. Li.
2016. Molecular basis for CPAP-tubulin interaction in controlling centriolar and
ciliary length. Nat Commun. 7:11874.
42
989
990
Materials and Methods
991
Sample Preparation
992
Nucleo-flagellar apparatus (NFAp) from Chlamydomonas reinhardtii was purified based
993
on previous published method (Wright et al., 1985). Briefly, cells of a cell-wall defect
994
strain (cc849, cw10 mt-, Chlamydomonas Resource Center) were grown in TAP medium.
995
Cultures were bubbled continuously with filtered air and illuminated with continuous
996
white light. After reaching log phase, the cells were harvested by centrifugation at 250g
997
for 5 minutes. The cells were resuspended in MT Buffer (30mM Tris-acetate, 5mM
998
MgSO4, 5mM EGTA, 25mM KCl, 1mM DTT, pH=7.3). The suspension was overlaid on
999
a 30% Percoll (GE Health) in MT Buffer. The sample was centrifuged for 5 minutes at
1000
300g. The supernatant was discarded. The pellet was resuspended in MT buffer, placed
1001
on ice, equal volume of Lysis Buffer (MT Buffer, 1% Nonidet P-40 (Roche), EDTA-free
1002
protease inhibitor cocktails (Roche) was added, mixed vigorously. The cells were lysed
1003
rapidly and completely in 10 minutes.
1004
The lysate was overlaid on top of discontinuous gradient of 50%, 40% and 30% Percoll
1005
in MT Buffer, spun at 10,000g. Fractions in 30% Percoll was recovered and diluted 10-
1006
fold with MT buffer. The sample was pelleted at 800g followed by resuspension in a
1007
small volume of MT Buffer.
1008
1009
Cryo-ET Data Collection and Reconstruction
1010
The purified NFAp was mixed with BSA-coated 10 nm colloid gold (Ted Pella). 4 µl
1011
sample was applied onto Quantifoil grid Rh/Cu 200 R2/2 (Quantifoil, Inc), blotted for 2
1012
second in 100% humidity and flash-frozen in liquid ethane using a Vitrobot (FEI, Inc).
43
1013
1014
Tomography tilt series were collected on a field emission gun (FEG) microscope (Polara,
1015
FEI, Inc) operated at 300kV. The microscope was equipped with a post-column energy
1016
filter Bio-Quantum GIF (Gatan, Inc). The slit width was set at 25 eV. UCSF Tomography
1017
software (Zheng et al., 2007) was used for collecting single-axis tilt series. For screening
1018
condition purpose, initial tomographic tilt series were collected on a CCD camera
1019
(UltraCam, Gatan Inc). But all tilt series used for subtomogram averaging and for
1020
classification were collected on a K2 direct electron detector in the counted mode (Gatan,
1021
Inc), in a nominal magnification of 50,000, the effective pixel size is 4.82 Å. The
1022
specimen was tilted from -60° to +60° in 1° increment. The exposure time at each tilt was
1023
1.0 second with 0.2 second per frame. This resulted in 5 frames at each tilt. The dose rate
1024
was set at 8 electron/pixel/second. The total accumulative dose for each tilt series was
1025
about 60 electron/Å2. The defocus was set in the range of 3~6 μm.
1026
1027
For tomogram reconstruction, a movie of 5 frames at each tilt was corrected of motion
1028
using MotionCorr (Li et al., 2013). Tomographic tilt series were aligned in IMOD
1029
(Kremer et al., 1996) by using 10 nm colloid gold beads as fiducials. The contrast transfer
1030
function for each tilt series was determined and corrected by TOMOCTF (Fernández et
1031
al., 2006). 3D tomograms were calculated by TOMO3D (Agulleiro and Fernandez,
1032
2015).
1033
1034
Subtomogram Averaging and Classification
44
1035
From 193 reconstructed tomograms containing NFAps, 201 centrioles and 110
1036
procentrioles were identified and selected for further processing. They were first
1037
annotated to find the proximal end of MT triplet. Based on these, the longitudinal spans
1038
of the MT triplet were defined. The subtomograms that contained segment of MT triplet
1039
were boxed out. Initially the triplet segment length was set to 38 nm with pixel size in
1040
2xbinned format. This was followed by ab initio alignment without a reference. The
1041
alignment of subtomograms were carried out by the programs MLTOMO implemented in
1042
the software package Xmipp (Scheres et al., 2009). After generating the initial reference,
1043
the subtomograms were re-extracted from the tomogram volumes in unbinned format.
1044
The MT triplet length in each subtomogram was gradually decreased. The final triplet
1045
length in the averaged structures for triplet, A-, B- and C-tubules was 29 nm, containing
1046
more than 3 repeats as the largest periodicity detected in the procentriole triplet is 8 nm.
1047
Since MT triplet is a continuous filament, the neighboring subtomograms should have
1048
similar geometric parameters. We have developed a program called RANSAC to impose
1049
constraint on the neighboring subtomograms from the same triplet. It was used to detect
1050
any alignment outliers and to impose constraint to correct misaligned subtomograms by
1051
regression. The program is available upon request. The program MLTOMO was also
1052
used in the classification of various structures such as the pinhead, the A-C linker, the A-
1053
tubule, B-tubule and C-tubule to detect heterogeneity. The processes are illustrated
1054
schematically in the supplementary figures.
1055
1056
For each tubule structure from 110 procentrioles, we first carried out focused
1057
classification on 2083 subtomograms on the A-, B- and C-tubule. This allowed us to
45
1058
identify 420, 758 and 862 defective or incomplete structure in each tubule respectively.
1059
These subtomograms were excluded from further focused alignment. The results are
1060
more homogeneous datasets for each tubule that combine procentriole and the proximal
1061
region of the mother centriole. The improvements are evident in the increased resolutions
1062
shown in Figure 1-figure supplement 1 and in Table 1 for all 3 tubules. These averaged
1063
tubule structures were used for further segmentation and estimation of the molecular
1064
weight for each MIP associated to the tubule.
1065
1066
For all the structures reported, resolution assessments were carried out in a “gold-
1067
standard” scheme (Chen et al., 2013). A post-processing program in the software package
1068
Relion (Scheres, 2012) was used to calculate the FSC between two halves of the maps.
1069
The statistics of the structures are summarized in Table 1 and the FSC plots are shown in
1070
the supplementary figures. The local resolution of the pinhead structure, as shown in
1071
Figure 2-figure supplement 13, is estimated by blocres in Bsoft (Heymann and Belnap,
1072
2007).
1073
1074
UCSF Chimera was used for displaying the surface-rendered averaged structures, the
1075
model building and the volume segmentation (Goddard et al., 2005).
1076
1077
To segment out the non-tubulin components, the MIPS, bound to the triplet and to
1078
estimate their molecular weights, an atomic model of MT protofilament was docked into
1079
the subtomogram averages of the A-, B- and C-tubules, respectively. These resulted in
1080
the atomic models for all 3 tubules (Figure 2). The differences between the subtomogram
46
1081
average map and the model density were attributed to the MIPs. The MIP density was
1082
segmented in UCSF Chimera. Based on the volume of segmented MIPs and assuming the
1083
protein density is 0.849 Dalton/Å3 (Fischer et al., 2004), the approximate molecular
1084
weight of each MIP was estimated (Table 2).
1085
1086
To model Poc1 in the A-C linker structure, a crystal structure of WD40 β-propeller
1087
domain with 7 repeats (PDB ID 1S4U) was first manually docked into the doughnut-
1088
shaped density in Leg B of the A-C linker. The docking was further optimized by using
1089
the “Fit in Map” function in UCSF Chimera. A local cross correlation coefficient
1090
between the model and the density map is 0.80, indicating a good fit.
1091
1092
Subtomogram Classification and Alignment on the A-C Linker
1093
For classification on the A-C linker structure, based on the initial alignment result of the
1094
triplet and by using the A- or C-tubule as reference point, respectively, we searched for
1095
subtomograms that exhibit intact A-C linker. In two parallel schemes, we identified 3992
1096
subtomograms from 4 classes and 3245 subtomogram from 5 classes (Class I with 233
1097
subtomograms was excluded due to its small number and low SNR) that show relatively
1098
intact A-C linker attached to the A- or the C-tubule, respectively (Figure 3-figure
1099
supplements 1,3). These two datasets are subjected to further refinement. The results are
1100
two structures, both at 23.1 Å resolution, showing more complete and detailed structure
1101
of the A-C linker (Figure 3-figure supplement 2,4,5).
1102
47
1103
To generate a complete structure of the A-C linker, since we have identified multiple
1104
classes of subtomogram showing intact A-C linker (Figure 3-figure supplements 1,3),
1105
though each in relative low resolution, we docked the above two averages (Figure 3-
1106
figure supplements 2,4) into these low-resolution class-averages. The result is a structure
1107
having complete A-C linker for each identified class. An example process is illustrated in
1108
Figure 3-figure supplement 6.
1109
1110
Analysis on the Periodicity of the Pinhead Structure
1111
For the pinhead structure, we carried out extensive studies on its longitudinal periodicity.
1112
There are several factors that can potentially affect the periodicity in the average. These
1113
are, 1) the length of the triplet in subtomogram, 2) the percentage of overlap between
1114
neighboring triplet segments, 3) the structural component included in the soft-edge mask
1115
that is used in subtomogram alignment, 4) a reference is provided or not during the initial
1116
alignment, 5) the structure difference between the triplet in the procentriole and the triplet
1117
from the proximal region of the mother centriole. All these factors can potentially
1118
introduce errors in the subtomogram alignment. We took these factors into account and
1119
did extensive tests. All results consistently show 8 nm periodicity in the pinhead
1120
structure. In the final map of the pinhead, the longitudinal length is 19 nm, including
1121
more than two repeats of the pinhead.
1122
48
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
References for Materials and Methods
Agulleiro, J.I., and J.J. Fernandez. 2015. Tomo3D 2.0--exploitation of advanced vector
extensions (AVX) for 3D reconstruction. J. Struct. Biol. 189:147-152.
Chen, S., G. McMullan, A.R. Faruqi, G.N. Murshudov, J.M. Short, S.H. Scheres, and R.
Henderson. 2013. High-resolution noise substitution to measure overfitting and
validate resolution in 3D structure determination by single particle electron
cryomicroscopy. Ultramicroscopy. 135:24-35.
Fernández, J.J., S. Li, and R.A. Crowther. 2006. CTF determination and correction in
electron cryotomography. Ultramicroscopy. 106:587-596.
Fischer, H., I. Polikarpov, and A.F. Craievich. 2004. Average protein density is a
molecular-weight-dependent function. Protein science : a publication of the
Protein Society. 13:2825-2828.
Goddard, T.D., C.C. Huang, and T.E. Ferrin. 2005. Software extensions to UCSF
chimera for interactive visualization of large molecular assemblies. Structure
(London, England : 1993). 13:473-482.
Heymann, J.B., and D.M. Belnap. 2007. Bsoft: image processing and molecular modeling
for electron microscopy. J. Struct. Biol. 157:3-18.
Kremer, J.R., D.N. Mastronarde, and J.R. McIntosh. 1996. Computer visualization of
three-dimensional image data using IMOD. J. Struct. Biol. 116:71-76.
Li, X., P. Mooney, S. Zheng, C.R. Booth, M.B. Braunfeld, S. Gubbens, D.A. Agard, and
Y. Cheng. 2013. Electron counting and beam-induced motion correction enable
near-atomic-resolution single-particle cryo-EM. Nat. Methods.
Scheres, S.H. 2012. RELION: implementation of a Bayesian approach to cryo-EM
structure determination. J. Struct. Biol. 180:519-530.
Scheres, S.H.W., R. Melero, M. Valle, and J.-M. Carazo. 2009. Averaging of electron
subtomograms and random conical tilt reconstructions through likelihood
optimization. Structure (London, England : 1993). 17:1563-1572.
Wright, R.L., J. Salisbury, and J.W. Jarvik. 1985. A nucleus-basal body connector in
Chlamydomonas reinhardtii that may function in basal body localization or
segregation. J. Cell Biol. 101:1903-1912.
Zheng, S.Q., B. Keszthelyi, E. Branlund, J.M. Lyle, M.B. Braunfeld, J.W. Sedat, and
D.A. Agard. 2007. UCSF tomography: an integrated software suite for real-time
electron microscopic tomographic data collection, alignment, and reconstruction.
J. Struct. Biol. 157:138-147.
49