Bl~hemica/Sysrernaticsand Ecok~, Vol. 12, No. 4, pp. 415-421,1984. Printed in Great Britain. 0305-1978/84$3.00+0.00 PergamonPressLtd. High Level of Gene Silencing in the Tetraploid Goldfish THADDEUS D. WOODS and DONALD G. BUTH Department of Biology, Universiw of California, Los Angeles, California 90024, U.S.A. Key Word I n d e x - Carassius auratus; Cyprinidae, Cypriniformes; isozymes, starch gel electrophoresis; polyploidy; gene expression; gene silencing. Abstract - The goldfish, Carassiusauratus, is a karyotypically tetraploid form that expresses only 19% of its enzyme-encoding loci in duplicate. This level of gene duplication is among the lowest reported among tetraploid cypriniform fishes and may be related to the intense selective and drift processes associated with its domestication. I~uction Polyploidization has played an important role in the evolution of many groups of fishes [1,2]. Among the Cypriniformes, a number of groups are believed to be of tetraploid origin including some cyprinid species, some cobitid species and genera, and the family Catostomidae [3-9]. To date, only eight species of cyprinid fishes have been identified as tetraploid forms [10]. Support for these tetraploid hypotheses have been primarily karyological; these cyprinids exhibit a chromosome complement of approximately 2n=100-104, a near-perfect doubling of the 2n = 48-52 of related diploid species [3, 4, 11-15]. Additional support for tetraploidy has come from electrophoretic studies of a few of these species [16-22] reporting the expression of duplicated enzyme loci (duplicated beyond the diploid condition) as might be expected in polyploid forms. However, not all genes duplicated via polyploidy are necessarily expressed. While some gene duplicates are free to acquire 'forbidden mutations' and perhaps develop new functions, others may be "silenced' returning the expression of this gene system to a diploid-like condition [1]. These nonfunctional genes, or pseudogenes, may be retained in the genome in silenced form [23]. The degree of functional diploidization of duplicated genes has been correlated with morphological advancement in cypriniform fishes [24] and this information can be systematically applied treating the gene silencing events as synapomorphies where appropriate [25, 26]. (Received for publication 24 November 1983) Among the tetraploid cyprinid fishes, only the carp, Cyprinus carpio, has had a comprehensive treatment of duplicate gene expression and gene silencing [20]; this species retains 52% of its duplicate enzyme loci. The goldfish, Carassius auratus, a cyprinid tetraploid related to C. carpio [11], has been the subject of duplicate gene expression studies of particular enzyme systems [e.g. 16-19]. However, a comprehensive gene expression data base comparable to that available for C. carpio [20] is lacking and is necessary for a comparative evaluation of tetraploidy in the Cyprinidae. In this study, gene expression data are developed for C. auratus and compared with C. carpio and diploid cyprinid controls. Results and Discussion Table 1 compares the number of genes expressed in the tetraploid Carassius auratus, another tetraploid cyprinid Cyprinus carpio [20], and diploid cyprinid controls in a number of selected enzyme systems. Only five out of twenty-six total loci, GpiA (Fig. 1), S-Icdh-A, S-Me-A (Fig. 1), Pgdh-A [16, 17], and S-Sod-A [22], express a duplicated condition in C. auratus. Five loci, M-Icdh-A, LdhC, Ldh-B (Fig. 2), M-Mdh-A, and S-Mdh-B that have been reported as perhaps expressing duplicated conditions in C. auratus, [17, 18, 22] were found to be expressed by single genes in this study. Thus, C. auratus retains only 19% of its loci (in our array) in the presumed original duplicate form. In regard to the number of genes controlling particular enzyme systems in C. auratus, our findings are consistent with previous reports of single locus expression of L-iditol (=sorbitol) 415 416 THADDEUS D. WOODS AND DONALD G. BUTH TABLE 1. NUMBER OF GENES EXPRESSED IN CARASS/US AURATUS COMPARED TO ANOTHER TETRAPLOID CYPRINID, CY;PRINUS CARPIO [20], AND A DIPLOID CYPRINID CONTROL Enzyme Enzyme commission number Aconitate hydratase (mitochondrial) Adenosine deaminase Adenylate kinase Alcohol dehydrogenase Aminopeptidase Aspartate aminotransferase (mitochondrial) Aspartate aminotransferase(supernatant) Calcium binding protein Calcium binding protein Crestine kinase Creatine kinase Fructose-bisphophatealdolase Glucosephosphate isornerase Glucosephosphate isomerase Glyceraldehyde-phosphatedehydrogenase Glycerol-3-phosphate dehydrogenase L-Iditol dehydrogenase 4.2.1.3 3.5.4.4 2.7.4.3 1.1.1.1 3.4.11.1 2.6.1.1 2.6.1.1 Locus Diploid control TetraDIoids C. carpio C. auratus Reference for C. auratus 1 1 Present study Present study 2.7.3.2 2.7.3.2 4.1.2.13 5.3.1.9 5.3.1.9 1.2.1.9 1.1.1,8 1.1,1.14 M-Acon-A Ada-A Ak-A Adh-A Ap-A M-Aat-A S-Aat-A Cbp-1 Cpb-2 Ck-A Ck-B Ald-C Gpi-A Gpi-B Gapdh-A G3pdh-A Iddh-A 1 1 2 2 1 1 2 1 1 1 1 1 1 1 1 2 1 1 1 Isocitrate dehydrogenase (mitochondrial) 1.1.1.42 M-Icdh-A 2 1, 2 Isocitrate dehydrogenase (supernatant) 1.1.1.42 S-Icdh-A 2 2 Lactate dehydrogenase 1.1.1.27 Ldh-A 1 1 Lactate dehydrogenase 1.1.1.27 Ldh-B 2 1, 2 Lactate dehydrogenase 1.1.1.27 Ldh-C 2 1, 2 Malate dehydrogenase (NAD; mitochondrial) 1.1.1.37 M-Mdh-A 2 1, 2 Malate dehydrogenase (NAD; supernatant) 1.1.1.37 S-Mdh-A 1 1 1 Present study Present study Presentstudy Presem study Present study Present study Present study Present study Present study Presentstudy Present study, [19, 28] Present study, [18} Present study, [18l Present study, [22, 271 Present study, [17, 22, 27] Present study, [17, 27] Present study, [22, 271 Present study, Malate dehydrogenase (NAD; supernatant) 1.1.1.37 S-Mdh-B 1 1 1,2 Present study, Malate dehydrogenase (NADP; supernatant) Phosphoglucomutase Phosphogluconate dehydrogenase 1.1.1.40 2.7.5.11 1.1.1.44 S-Me-A Pgm-A Pgdh-A 1 1 1 1 2 2 1 2 Superoxide dismutase (supernatant) 1.15.1.1 S-Sod-A 1 2 2 Present study Present study Present study, [16, 17} Present study, Xanthine dehydrogenase 1.2.1.37 Xdh-A 1 1 1 Present study 0% 54% 19-28% - 2 1 2 2 [22,27} 122,271 [22} Percentduplication dehydrogenase, Iddh-A, [19, 28], and duplicate gene expression of supernatant isocitrate dehydrogenase, S-Icdh-A [18, 19J, phosphogluconate dehydrogenase, Pghd-A [16, 17], and supemxide dismutase, S-Sod-A [22]. However, our gene expression data differ from that reported for the Ldh-B and Ldh-C loci [17], mitochondrial M-Mdh-A and one supernatant S-Mdh-B malate dehydrogenase [22], and mito- chondrial isocitrate dehydrogenase M-Icdh-A [18, 19]. Klose et al. [17] examined lactate dehydrogenase expression in selected diploid and tetraploid cyprinids including C. auratus. These investigators reported duplicated Ldh-B and Ldh-C ( = 'Ldh-L') genes in their tetraploid group (Barbus barbus, Cyprinus carpio, Carassius carassius and Carassius auratud. While these duplications are 417 ® Gpi-A Origin /~, 0125456 78910 Glucosephosphate isomerose ® S-Me-A M-Me-A Origin----~ QI B 25 4 5 6 78 910 Melic enzyme FIG. 1. SINGLE GENE EXPRESSION IN A DIPLOID CYPRINID CONTROL, RICHARDSONIUS BALTEATUS (SPECIMENS 1-5}, COMPARE[ TO DUPLICATE GENE EXPRESSION IN THE TETRAPLOID CARASSIUS AURATUS (SPECIMENS 6-10): A, Glucosephosphate isomeras, (Gpi-A from brain tissue extracts); B, NADP-dependent malate dehydrogenase ( = 'malic enzyme,' supernatant or soluble from S*Me-A fron muscle tissue extracts). 418 B B3 Bz B1 A]B AzB A3B A Origin (~) 1 2 54 5 6 78910 Lactate dehydrogenose FIG. 2. LACTATE DEHYDROGENASE EXPRESSION (BOTH Ldh-A AND Ldh-B GENE PRODUCTS FROM MUSCLE TISSUE EXTRACTS) IN A DIPLOID CYPRINID CONTROL, RICHARDSONIUS BALTEATUS (SPECIMENS 1-5), AND THE TETRAPLOID CARASSlUS AURATUS (SPECIMENS 6-10}. Both species express single Ldh-A and Ldh-B genes and all expected randomly-associated heterotetrameric combinations of isozymes are observed. The subunit combination of each electromorph in each species is indicated. GENE SILENCING IN CARASSIUS AURATUS cleady recognized in Cyprinus carpio, these are neither as clearly figured nor specifically discussed for Carassius auratus. Indeed, the LDH zymogram from C. auratus muscle extracts of Klosa etal. [17] shows the same five-banded pattern of single locus Ldh-A and Ldh-B expression as we have found from muscle (Fig. 2) and heart extracts. Thus, there may have been differential gene expression of Ldh-B among the tetraploid species studies by Klose et al. [17]; that is, C. auratus differs from the other three, including the congeneric C. carassius, in expressing a single Ldh-B locus. The numerous LDH isozymes resolved from liver extracts in this study and probably in the zymogram of Kloee etal. [17] can be explained as all random combinations of Ldh-A, Ldh-B and Ldh-C subunits in this tissue. Again, Klose eta/. [17] clearly demonstrate duplication of the Ldh-C locus in Cyprinus carpio but do not designate multiple Ldh-C locus products for C. auratus nor specifically discuss the expression of Ldh-C in this species. Our interpretation of single locus control of Ldh-A, Ldh-B and Ldh-C in C. auratus is in agreement with that of Wilson et al. [27] who examined the expression of this enzyme system in C. auratus tissues while varying the acclimation temperatures of the specimens. Danzmann and Down [22] also reported single locus expression of Ldh-A and probable single locus expression of Ldh-B in C. auratus. Danzmann and Down [22] reported a five-locus control of the malate dehydrogenase system in C. auratus suggesting that a slowly-migrating set of isozymes (probably mitochondrial products) were controlled by two loci and that the musclepredominating supernatant products (S-Mdh-B) were also duplicated yielding two mitochondrial and three supernatant loci in this system. Their zymogram f'~ure clearly shows electromorphs consistent with this interpretation. However, Wilson et al. [27] provided zymograms that are consistent with the three-locus MDH system (one each M-Mdh-A, S-Mdh-A and S-Mdh-B) that we observe in C. auratus. Danzmann and Down [22] discuss this locus control discrepancy as follows: "Therefore it is possible that regulatory polymorphisms also exist in this species. It is equally possible, however, that such differences can be produced by the electrophoretic conditions used. Comparative studies of different populations using the same electrophoretic procedure are required to resolve this ambiguity." 419 Quiroz-Gutierrez and Ohno [18] and Engel etal. [19] examined gene expression in the multilocus isocitrate dehydrogenese enzyme system in selected diploid and tetraploid cyprinids including C. auratus. Engel et al. [19] reported duplicate expression of the mitochondrial form of this enzyme M-Icdh-A in the tetraploid assemblage, however it should be noted that Carassius auratus was omitted from the comparison of ICDH expression (Table 1 of [19]). Quiroz-Gutierrez and Ohno [18] reported that the "M-form IDH was seen as three closely spaced bands rather than as a single band" in C. auratus and concluded that this represented a gene duplication for this dimeric enzyme. Our study yielded a single band of M-Icdh-A activity in both C. auratus and the diploid control using a comparable phosphate-citrate buffer system. This difference in M-Icdh-A expression may be real; their stock(s) of C. auratus may have retained this locus in duplicate whereas it may have been silenced in our sample. However, this hypothesis is unlikely because gene duplication differences within species are extremely rare and their sample of C. auratus included 75 individuals "of various breeds.., purchased from local pat stores" presumably also from southern California. Thus, the problem of M-Icdh-A expression in C. auratus remains unresolved and deserves additional study. The high level of gene silencing in Carassius auratus yielding only 19% retention of duplicate genes is unusual for tetraploid cypriniform fishes. This low level is in marked contrast to the 52% retention of duplicated genes in another tetraploid cyprinid, Cypnnus carpio [20]. Despite the fact that C. carpio expresses more genes than C. auratus, the amount of cellular DNA of C. carpio is actually less than C. auratus, 1.7 vs 2.0 picograms haploid DNA content [29, 30]. Duplicate gene expression among the most advanced tetraploid catostomid fishes (range of 35-65%) [24, 26, 31] is comparable to the highest possible estimate (Table 1) for C. auratus. It is only among the cobitids (Ioaches) that comparably low levels of duplicate genes (15-30%) have been reported [15]. Thus, the high level of gene silencing (low level of retained duplicate genes) in C. auratus may be explained in two ways: (a) C. auratus is an 'old' tetraploid that has lost most of its duplicate genes, comparable to the condition exhibited by the most advanced catostomids (suckers) [24, 31]; or (b) C. 4~ auratus is a 'recent' (perhaps auto-) tetraploid that is retaining its duplicated genes but is expressing them, without allelic divergence, in a cryptic fashion that is not easily interpreted, comparable to the cobitid condition [15]. The first hypothesis may be favored. Ohno et al. [11] reported only bivalent chromosome pairing in C. auratus rather than quadrivalent formation that may be expected in recent autotetraploids. However, these authors point out that lack of quadrivalent formation does not eliminate the possibility of an autotetraploid origin: " . . . it is possible that an ancestral species to the goldfish arose as an autotetraploid yet four original homologues gradually diverged into two different pairs" [11]. Another line of evidence supporting actual dipIoidization rather than cryptic duplicate gene expression comes from the expression of intralocus variation within C. auratus. The ratio of staining intensities of the isozymes in the heterozygous condition (e.g. in ADA and PGM) ware consistent with two, or three in the case of Ada-A, alleles at a single locus instead of superimposed products of two loci. Therefore, although the allotetraploid vs. autotetraploid origin of C. auratus may still be debated, it appears that the vast majority of its duplicate genes have been silenced since that polyploidization event. If selection can play an important role in the silencing of duplicate genes, the condition revealed in C. auratus is easy to reconcile. This species has been bred for centuries for human delight in their specific shape, size and coloration patterns. Such artificial selection for certain variants may have selected inadvertantly for particular combinations of newly silenced genes. This process may have been accelerated by genetic drift via the 'genetic bottlenecks' that often accompany a selective breeding program. The widespread use of Carassius auratus as an experimental animal, especially in physiological and biochemical studies, may not be ideal especially if the experimenters wish to generalize regarding 'typical' expression in fishes or even vertebrates. As a tetraploid, the traits expressed by C. auratus may occur in ways quite different than those found in its more typical diploid relatives. However, as an advanced tetraploid having lost most of its duplicate genes, its resulting expression may be similar enough to the diploid condition to THADDEUS D. WOODS AND DONALD G. BUTH warrant its use as an experimental subject. Yet, reservations regarding the use of C. auratus as a typical fish may be appropriate because of its domestication and genetic specialization, not necessarily because it is a tetraploid form. tf C. auratus is indeed a domesticated descendant of Carassius carassius [32], it would be interesting to study feral populations of C. auratus: [e.g. 28] to discover if morphological reversion (color and shape) to a 'wild state' is accompanied by any genetic changes in terms of gene expression. These two species of Carassius already differ in gene expression of the Ldh-B locus [17] as previously discussed. A comprehensive examination of gene expression in C. carassius from native East Asian waters would be helpful in assessing the effects of domestication of gene expression in C. auratus. Experimental Specimens of Carassius auratus were obtained from a commercial distributor in Los Angeles, California. The diploid cyprinid control, Richardsonius balteatus, was obtained by seine in the Umpqua River system in Oregon. Information on diploid expression of certain enzyme systems was also drawn from that reported for Gila orcutti [31] and Notropis stramineus [20]. All specimens were frozen on dry ice upon purchase or capture and were stored at - 20 ° until examined. Skeletal muscle, liver, and brain tissues were dissected from each specimen. Extracts were prepared separately by mixing each tissue with an equal volume of 0.1 M Tris-HCI buffer at pH 7.0, mechanically homogenizing the sample, and centrifuging the homogenate for 15 min at 15 000 g at 4 °. The clear supernatant fractions were subjected to horizontal starch gel electrophoresis using 14% gels made with a combination of 80% electrostarch (lot no. 392; Electrostarch Co., Madison, Wisconsin 53701 USA) and 20% Connaught starch (obtained from the Fisher Scientific Company). The optimal electrophoretic conditions for cyprinids, including relevant combinations of tissues and buffers, have been previously described [10]. The histochemical staining procedures used were modified from several sources [e.g. 33, 34]. Enzyme and locus terminology follows that recommended by the International Union of Biochemistry [35] and previous studies of cypriniform fishes [2, 10], respectively. The sample size criterion of five or more specimens yielding (P <0.05) to ascertain the number of functional gene loci for the assignment of single or duplicate gene expression, as discussed by Ferris and Whitt [25], was employed for all diploid-tetraploid comparisons at all loci; sample sizes of twelve to twenty-four C. auratus were usually used. Acknowledgements-This study was supported by the UCLA Department of Biology Fisheries Program, the UCLA Biomedical Support Grant, the UCLA Committee on Research (U.R. 3674 to D. G. Buth), and, in part, by the GENESILENCINGIN CAP=ASS/USAURATUS UCLA Undergraduate Summer Research Program for Minority Students (fellowship to T. D. Woods). We would like to thank C. Ben Crabtree for his r=sistance in both the field and laboratory and Kevin J. Collins for his laboratory assistance. S. D. Ferris and W. J. Rainboth critically evaluated the manuscript. References 1. Ohno, S. (1970) Evolution by GeneDuplication. Springer, New York. 2. Buth, D. G. (1983) in Isozyrnes: Current Topics in Bio logical and Medical Research, Vol. 10 (Rattazzi, M. C., Scandalios,. J. G. and Whitt, G. S., eds). Alan R. Liss, New York. 3. Wolf, U., Ritter, H., Atkin, N. B. and Ohno, S. (1969). Humangenetik 7, 240. 4. Suzuki, A. and Taki, Y. (1981)Jap. J. Ichthyology28, 173. 5. Ueno, K. and Ojima, Y. (1978) Proc. Jap. Acad. 52, 446. 6. Ueno, K., Iwai, S. end Ojima, Y. (1980) Bull. Jap. Soc. Sci. Fish. 46, 9. 7. Ferris, S. D. and Whirr, G. S. (1977) Biochem. Genet. 15, 1097. 8. Uyeno, T. and Smith, G.R. (1972). Science 175, 644. 9. Schultz, R. J. (1980) in Polyploidy, Biological Relevance (Lewis, W. H., ed.). Plenum Press, New York. 10. Buth, D. G. (1984) in Evolutionary Genetics of Fishes (Tumer, B. J., ed.). Plenum Press, New York. 11. Ohno, S., Muramoto, J., Christian, L. and Atkin, N. B. (1967) Chromosoma23, 1. 12. Berberovic, L., Hadziselimovic, R., Pavlovic, B. and Sofradzija, A. (1973) Bull. Sci. Acad. RSF Yougoslavie 18, 10. 13. Sofradzija, A. and Berberovic, L. (1973) Bull. Sci. Acad. RSF Yougoslavie18, 77. 421 14. Park, E. H. (1974) College Rev., College of Liberal Arts and Sciences, Seoul Nat. Univ. 20, 346. 15. Khuda-Bukhsh, A. R. (1980) Experientia33, 173. 16. Bender, K. and Ohno, S. (1968) Biochem. Genet. 2, 101. 17. Klose,J., Wolf, U., Hitzeroth, H., Ritter, H. and Ohno, S. (1969) Humangenetik 7, 245. 18. Quiroz-Gutierrez, A. and Ohno, S. (1970) Biochem. GeneL 4, 93. 19. Engel, J., Faust, J. and Wolf, U. (1971) Anim. Blood Grps. Biochem. Genet. 2, 127. 20. Ferris, S. D. and Whirr, G. S. (1977) Experientia33, 1299. 21. Triantaphyllidis, C. D., Damianakis, H., Economidis, P. S. and Karakousis, J. (1981) Comp. Biochem. Physiol. 75B, 278. 22. Danzmann, R. G. and Down, N. E. (1982) Biochem. Genet. 20, 1. 23. Li, W-H. (1983) in Evolution of Genesand Proteins (Ne[, M. and Koehn, R. K., eds). Sinauer, Sunderland, Massachusetts. 24. Ferris, S. D, and Whitt, G. S. (1977) Nature 265, 258. 25. Ferris, S. D. and Whitt, G. S. (1978) Syst. Zoo/. 27, 189. 26. Buth, D. G. (1979) Comp. Biochem. Physiol. 63B, 7. 27. Wilson, F. R., Whitt, G. S. and Proseer, C. L. (1973) Comp. Biochem. Physiol. 46B, 105. 28. Lin, C-C., Schipmann, G., Kittrell, W. A. and Ohno, S. (1969) Biochem. Genet. 3, 603. 29. Hinegardner, R. (1968)Am. Nat. 102, 517. 30. Hinegardner, R. and Rosen, D. E. (1972) Am. Nat. 106, 621. 31. Crabtree, C. B. and Buth, D. G. (1981) Copeia 1981,705. 32. Matsui, Y. (1934) J. Imp. Fisheries Inst. (Japan) 30, 1 (cited by Linet al. [28]). 33. Brewer, G. J. (1970) An Introduction to Isozyrne Techniques. Academic Press, New York. 34, Shaw, C. R. and Prasad, R. (1970) Biochem. Genet. 4, 297. 35. International Union of Biochemistry (1979) Enzyme Nomenclature, 1978.Academic Press, New York.