CYTOGENETIC STUDIES IN NORTH AMERICAN MINNOWS (CYPRINIDAE). IV. SOMATIC POLYPLOIDY IN GILA BICOLOR

The kidney tissue of a single individual of the California minnow Gila bicolor (Girard) contained polyploid cells in about 1.7% frequency. Chromosome spreads of triploid, tetraploid, hexaploid, octaploid, and dodecaploid cells were observed and may have arisen thmugh endoreduplication of ancestral diploid and triploid cells. The cytological mechanism producing the triploid cells is unknown. Diplochromosomes were not present. The distribution of pfoidy in cells of this individual is not random. In particular, cells having undergone one round of chmmosomal increase appear increasingly susceptible to additional rounds of chromosomal gain. On a tmuve dms le tissu du rein d'un seul specimen de vaimn de Califsrnie, Gila bicslor (Girard) des cellules polyploides i~ une fdquence d'envimn 1.7%. D%s pn5parations broyees de chmrmssomes ont rev616 des cellules triploides, tetraploides, hexaploides, octagloides et dod6caploides. Ces dernitres sont peut-kre le n5sultat d'une endomitose des cellules mcestrales diploides et triploides. Le mecanisme cytofogique qui provoque la formation des cellules triploides n'est pas connu. & n'a pas vu de diplochromosomes. Chez ce spkcimen la distribution des cellules gloiddes n9est pas aleamire. En particulier, les cellules qui ont &j8 subi un taux d9accroissemnt en m b r e de chromosomes semblent susceptibles h un autre accroissement .


Introduction
Scattered polyploid cells in predominantly diploid plant tissues were first reported early in the century (Stomps, 1918;Hustd, 1932;Larter, 1932;Lorz, 1937;Ervin, 1941). Somatic polygloidy is of wide-spread if not universal occurrence in plants, particularly Angiospemae (Lon, 1947;D' Amato, 1964). Many cases have also been reported in animals, including man (Holt, 1917;Schwarzacher and Schnedl, 1965;Kelly and Almy, 1969). Polyploid cells most often occur in strongly differentiated tissue. Typically, such cells contain (2n)2" chromosomes, where 2n is the diploid number andx is a small integer. This suggests that the extra chromosomes arise through duplications of entire chromosome sets without intervening mitotic divisions. The adaptive significance, if any, of somatic polyploidy is not known, although it has been suggested that the additional chromosome sets increase and economize the synthesizing capacity of a cell or tissue (Nagl, 1976a; see also Nagl, 1976b).
In this paper, we report the finding of a number of polyploid cells in the kidney of the California minnow Gila bicolor (Girard). In tissue from a single individual, we have observed diploid, triploid, tetraploid, hexaploid, octaploid, and dodecaploid cells. The distribution of ploidy within this individual, and among other minnows examined, does not appear random. Furthermore, the appearance of 3x, 6x, and l2x cells cannot be explained simply by the repated doubling of whole chromosome sets in a diploid ancestral cell.

Manuscript received May 9, 1977.
Can. J. Genet. Cytol. 19: 657462,1977. It is appropriate here to briefly review some of the terminology applied to somatic polyploid cells, since the terms are confusing and have not been used consistently. As originally used by Geitler ( 1939), "endsmitosis' referred to the appearance of additional chromosome-Bike structures during an abortive mitosis. With the recognition that chromosome replication and mitosis are separate events, "endsmitosis" has largely been discarded and 6endoredupBication" now refers to chromosomal replication in the absence of mitosis leading t s polyploidy (Mittwoch et a&. , 1965). "Endopolyploidy results from "'endsreeduplicatiofl as opposed to other possible mechanisms of generating polyploidy such as nuclear fusion. Since endoreduplication may not be the only process leading to polyploid Gila cells, we will use interchangeably the more general terns po1ysomaty9 ' or somatic polyploidy . ' '

Makerids and Methods
The specimen of Gila bicolor exhibiting polysornaty was found during a routine karyological study of Cyprinidae inhibiting California (Gold and Avise, 1977~1). A total of four Gila bicolor were collected by seine near Crowley Lake, Mono County, California, and returned live to the laboratory for karyotyping. Chromosome preparation and analysis was carried out using the technique of Gold (1974), employing kidney as the source tissue and colchicine as the mitotic inhibitor. Celchicine (0.5% in 0.85% sterile saline solution) was injected into the dorsal musculature three hours prior to dissection of the kidney.

Results
The diploid kayotype of Gila bicolor consists of 44 chromsomes with median or submedian centromeres, and 6 chromosomes with subterminal or terminal centromeres (Fig. 1). This karyotype differs relatively little in gross configuration from those of eight other California minnow species, or from many of the nearly 250 members of North American Cyprinidae . The diploid number was determined from actual counts of 145 well-spread cells, and more cursory examination of an additional 477 cells, taken from four specimens of Gila bicokor ( Table I ) .
An examination of 43 cells in three Gila individuals revealed no evidence of polysomaty. In cells from a fourth individual, however, we observed apparently intact nuclei containing approxinlately integer multiples of the haploid chromosome number ( Table 1; Fig. 2). About 1.740 of the chromosome spreads of kidney cells were polyploid (one dodecaploid , one octaploid , one hexaploid , four tetrapioid , and three triploid cells were observed). The polyploid knryotypes did not exhibit exactly the expected numbers of chromssames. For example, we count about 291 chrsmosomes in the presumed dodecaploid cell; the additional chromosomes were no doubt obscured by overlap, or perhaps lost during cell preparation.
With our sample sizes, the frequency of occurrence of polyploid cells in this individual GGi is not demonstrably different from that of the other Gila examined. However, we have also examined a total of 696 cells in 33 diploid individuals belonging to nine cyprinid species, without evidence of any polysomaty . We may also ask whether the chromosome replicating episodes occur independently of one another in the kidney cells of the fish exhibiting polyssmaty. In order to do this we must make certain assumptions about the mode sf generation of new chromosome complements. We assume that many new chromossmai sets are not generated spontaneously, in saltatory fashion, but rather arise in a sequential manner.
We d w assume that at least one sf the ancestral cells sf a cell now pslyplsid was diploid. For example, the minimum number of chromosome increasing events needed to generate a hexaploid e l 1 is two (2n -+ 4x, + 2se -+ 6x; or 2a + 3x -+ 6x1; the minimum number for an octaploid cell is also two (2n 4 4-x -+ 8x1; and the minimum number for a dodecaploid cell is three (i.e., 2p1 -+ 3, x -+ 6x + I&). This is true whether new chromosome sets are generated through endopolyploidy of whole or partial chmrnosome comp8ements, or whether nuclear fusions are involved.
The observed numbers of kidney cells having undergone a given minimum number of chromosome increasing episodes are listed in Table 11, and compared to frequencies expected under the relevant Poisson distribution. The discrepancy is highly significant. The excess numbers of cells exhibiting higher ploidy levels (dodecaploid, octaploid, hexaploid) , and the deficiency of cells with low ploidy levels (tetraploids , triploids) , suggest that the occurrence of one chromosome increasing event enhances the probability of second and later such events in the same cell line.

Discussion
In the first mitosis of nuclei following two series of replications, chromosomes characteristically occur in closely associated pairs in many organisms (Berger, 1941;Kelly and Almy, 1969;Levan, 1939;Levan and Hauschka, 1953). These pairs are called diplochromosomes, and provided the earliest and perhaps strongest evidence that endoreduplication rather than nuclear fusion is responsible for polysomaty (De Litarditre, 1923;see Lorz, 1947). However, in subsequent mitoses the duplicated chromosomes become dispersed by movements of cell division, and diplochromosomes m y be rare in a polysomatic cell population, The ckbfomosomes in d l of the polyploid Gila cells appear unpaired, and hence we cannot convincingly argue from our data that endoreduplication was the cause of polysomaty. Endoreduplication of entire chromosome sets could readily account for the appearance of tetraploid and octaploid cells from a diploid ancestor, and hexaploid and dodecaploid cells from a triploid ancestor. In most organisms, endoreduplication involves all the chromosomes of a somatic cell. Hence the appearance of triploid cells and their multiples in Gila is most surprising. For the present, we can offer no empirical rationale for the origin of a triploid somatic cell.
The colchicine treatment probably did not cause the polysomaty in Gila. Using the same technique, we have now examined numerous karyotypes of more than 500 fish without observing m y other instances of higher ploidy levels. Colchicine is widely used as a polyploid inducing agent in plants; its success in inducing polyploidy in animals has been much more limited (Dermen, 1940). It seems unlikely that if colchicine were the causal agent in Gila its effect would be so unevenly distributed among cells. Furthermore, only three hours elapsed between injection and sacrifice, probably too little time for the appearance of colchicine-induced ploidy at the higher levels observed. Finally, plysomaty is probably not the result of artificial treatment of any kind in many other organisms (Berger, 1941).
Whatever the cause of polysomaty in Gila, the data suggest that cells which contain more than two genomes are increasingly susceptible to additional rounds of chromosome increase. This finding is potentially important, since any mutationally derived genetic differences between cells, or any environmental parameters to which cells are differentially exposed (e, g . , position effects or viruses), might produce this result. This observation in Gila needs to be verified by additional study as an important first step in determining the genetic and environmental factors responsible for polysomaty .
A few cases of polysomaty in fishes have previously been noted (Nygren el a&., 1968(Nygren el a&., , 1971(Nygren el a&., , 1975. One of us (J.R. 6.) has also observed tetraploid cells in about 18-12 percent of kidney cells in Salmo aguabonita (unpublished data). Lorz (1947) writes, "It is significant that in most cases polysomaty is not manifest in cells which undergo meiotic reduction into spores or gametes. Otherwise, naturally occurring polyploid forms would be much more frequent than they are." However, we have recently reported the discovery of a single triploid individual of Hesperoleucus symmetricus, a minnow not too distantly related to Gila (Gold and Avise, 1977b). A likely origin of the triploid Hesperoleucus was fusion of a haploid sperm with a diploid ovum. Thus in fishes, the occurrence of cells containing multiples of the normal haploid or diploid numbers may occasionally be of some significance during sexual reproduction as well.