Evidence for a monophyletic origin of triploid clones of the Amazon molly, Poecilia formosa.
ABSTRACT Asexual reproduction in vertebrates is rare and generally considered an evolutionary dead end. Asexuality is often associated with polyploidy, and several hypotheses have been put forward to explain this relationship. So far, it remains unclear whether polyploidization in asexual organisms is a frequent or a rare event. Here we present a field study on the gynogenetic Amazon molly, Poecilia formosa. We used multilocus fingerprints and microsatellites to investigate the genetic diversity in 339 diploid and 55 triploid individuals and in 25 P. mexicana, its sexual host. Although multilocus DNA fingerprints found high clonal diversity in triploids, microsatellites revealed only two very similar clones in the triploids. Phylogenetic analysis of microsatellite data provided evidence for a monophyletic origin of the triploid clones of P. formosa. In addition, shared alleles within the triploid clones between the triploid and diploid genotypes and between asexual and sexual lineages indicate a recent origin of triploid clones in Poecilia formosa.
- SourceAvailable from: Ingo Schlupp[Show abstract] [Hide abstract]
ABSTRACT: The evolution and maintenance of sexual reproduction is still one of the major unresolved problems in evolutionary biology. Sexual reproduction is fraught with a number of costs as compared to asexual reproduction. For example, sexuals have to produce males, which–given a 1:1 sex ratio—results in a two-fold advantage for asexuals that do not produce males. Consequently, asexuals will outperform and replace sexuals over time assuming everything else is equal. Nonetheless, a few cases of closely related asexuals and sexuals have been documented to coexist stably in natural systems. We investigated the presence of a two-fold cost in a unique system of three closely related fish species: the asexual Amazon Molly (Poecilia formosa), and two sexual species, Sailfin Molly (P. latipinna) and Atlantic Molly (P. mexicana). Amazon Molly reproduce gynogenetically (by sperm dependent parthenogenesis) and always coexist with one of the sexual species, which serves as sperm donor. In the laboratory, we compared reproductive output between P. formosa and P. mexicana as well as P. formosa and P. latipinna. We found no differences in the fecundity in either comparison of a sexual and the asexual species. Under the assumption of a 1:1 sex ratio, the asexual Amazon Molly should consequently have a full two-fold advantage and be able to outcompete sexuals over time. Hence, the coexistence of the species pairs in nature presents a paradox still to be solved. KeywordsPoeciliid-Asexual-Recombination-FecundityEnvironmental Biology of Fishes 06/2010; 88(2):201-206. · 1.31 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The maintenance of sex is still an evolutionary puzzle given its immediate costs. Stably coexisting complexes of asexually and sexually reproducing forms allow to study mechanisms that balance the costs and benefits of both asexual and sexual reproduction. Here, we tested whether coexisting asexual and sexual fish of the genus Poecilia differed in neonate mortality when exposed to environmental stress in the form of fluctuating temperatures and food deprivation. We find that asexual Amazon mollies, Poecilia formosa, are significantly more sensitive to food stress than their sexual relative Poecilia latipinna, but both are equally unaffected by variable temperatures. Differences in the susceptibility to environmental stress may contribute to diminishing the asexuals’ benefits of a higher intrinsic population growth rate and thus mediate stable coexistence of the two reproductive forms.Evolutionary Ecology 01/2010; 24(1):39-47. · 2.41 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The gynogenetic livebearing Amazon molly (Poecilia formosa) is a sexual parasite that exploits males of closely related species for sperm. This is needed as physiological stimulus for embryo development; however, none of the male’s genes are normally incorporated into the genome of the gynogenetic offspring. Mostly diploid individuals were reported from the natural habitats in North-Eastern Mexico and South-Eastern Texas but stable populations of triploids have been reported from the Río Soto la Marina drainage and in the Río Guayalejo in North-Eastern Mexico. Triploidy is the result of defects in the mechanisms that normally clear the host sperm from the ameiotic diploid egg. Triploids also reproduce gynogenetically and their frequencies fluctuate markedly between years, seasons, and localities. To understand the dynamics of this mating system, it is important to understand the relative reproductive success of triploids and diploids. We hypothesize that triploids should have a selective advantage over diploids due to heterosis and/or gene redundancy based on the additional genetic material from the sexual host. However, clonal competition experiments revealed a clear reproductive advantage of diploids competing with triploids. This result contradicts not only our hypothesis but also the stable co-existence of diploids and triploids in natural habitats. Frequency dependent selection, niche partitioning and environmental heterogeneity are discussed as possible explanations.Evolutionary Ecology 01/2009; 23(5):687-697. · 2.41 Impact Factor
? 2005 The Society for the Study of Evolution. All rights reserved.
Evolution, 59(4), 2005, pp. 881–889
EVIDENCE FOR A MONOPHYLETIC ORIGIN OF TRIPLOID CLONES OF THE AMAZON
MOLLY, POECILIA FORMOSA
KATHRIN P. LAMPERT,1,2,*DUNJA K. LAMATSCH,1,*JO¨RG T. EPPLEN,3AND MANFRED SCHARTL1
1Department of Physiological Chemistry I, University of Wuerzburg, Biozentrum, Am Hubland, 97074 Wuerzburg, Germany
3Department of Human Genetics, University of Bochum, 44801 Bochum, Germany
is often associated with polyploidy, and several hypotheses have been put forward to explain this relationship. So far,
it remains unclear whether polyploidization in asexual organisms is a frequent or a rare event. Here we present a field
study on the gynogenetic Amazon molly, Poecilia formosa. We used multilocus fingerprints and microsatellites to
investigate the genetic diversity in 339 diploid and 55 triploid individuals and in 25 P. mexicana, its sexual host.
Although multilocus DNA fingerprints found high clonal diversity in triploids, microsatellites revealed only two very
similar clones in the triploids. Phylogenetic analysis of microsatellite data provided evidence for a monophyletic origin
of the triploid clones of P. formosa. In addition, shared alleles within the triploid clones between the triploid and
diploid genotypes and between asexual and sexual lineages indicate a recent origin of triploid clones in Poecilia
Asexual reproduction in vertebrates is rare and generally considered an evolutionary dead end. Asexuality
ternal introgression, polyploidy.
Clonal diversity, gynogenesis, microsatellite marker, Muller’s ratchet, multilocus DNA fingerprint, pa-
Received July 21, 2004.Accepted January 5, 2005.
Strictly unisexual vertebrate species are rare and an ex-
ception to the dominant mode of sexual reproduction in an-
imals. The Amazon molly, Poecilia formosa, was the first
vertebrate discovered to reproduce clonally (Hubbs and
Hubbs 1932). This all-female species has become a model
system for studying the evolution of sex. Like many unisexual
organisms it originated from a hybridization event (Dawley
1989), most likely between P. mexicana limantouri and P.
latipinna (Avise et al. 1991; Schartl et al. 1995a). Its repro-
ductive mode is gynogenesis: sperm from a closely related
bisexual species is needed to stimulate the onset of embryo-
genesis of the unreduced diploid eggs (Monaco et al. 1984),
but usually makes no genetic contribution to the embryo
(Schartl et al. 1990). In very rare cases, however, the exclu-
sion mechanism fails and either small parts of male genetic
material remain inside the oocyte in the form of microchro-
mosomes (Schartl et al. 1995b), or the sperm nucleus fuses
with the oocyte nucleus leading to gynogenetic triploid in-
dividuals (Rasch and Balsano 1974; Turner et al. 1980a,b;
Lamatsch et al. 2000a).
Many asexually reproducing organisms are polyploid even
though their ancestral closely related sexual species are dip-
loid. Three main hypotheses have been suggested to explain
this: (1) polyploidization might interrupt meiosis, therefore
creating asexual lines. (2) There might not be a high selection
against polyploids in asexual lines. (3) It might be advan-
tageous to the asexuals to be polyploid (Mogie 1986). Be-
cause P. formosa is a hybrid species, and diploid genotypes
also reproduce asexually, it can be excluded for this species
that polyploidy interrupts meiosis and therefore is the basis
for asexuality. To distinguish between the other two hypoth-
eses we considered the genetic variability of triploid indi-
viduals compared to diploids in this species. An equal or
lower level of genetic diversity in triploids would suggest
that polyploidization might simply not be selected against.
Higher genetic variability as a precondition to adapt to newly
emerging selective pressure would provide evidence that trip-
loids enjoy the advantage of additional genetic material.
Various studies have tried to determine the amount of ge-
netic variability in diploid P. formosa populations. Immu-
nological techniques (tissue transplantations) were used first
(Kallman 1962), followed by allozyme studies (Turner et al.
1980a,b). DNA fingerprinting revealed very high levels of
genetic variation in natural populations (Turner et al. 1990).
Triploid clonal variability has only been estimated using al-
lozymes and was proposed to be rather high (Turner et al.
Since multilocus DNA fingerprints revealed the highest
number of clones for diploid unisexual fish, we used this
method for a study of genetic diversity in diploid and triploid
P. formosa. In addition, we used three microsatellite loci
because they allow work on a larger number of individuals.
We also wanted to validate this method for population genetic
analysis of clonally reproducing vertebrates. To investigate
whether polyploidy in clonal fish could bear the advantage
of higher genetic variability, we investigated the following
questions: Are triploids in natural populations genetically
more variable than diploid individuals? How frequently do
successful paternal introgressions lead to triploid individu-
als? Are these incorporations an old or a recent event in P.
Here we report a high degree of genetic variability in dip-
loid and less variation in triploid P. formosa indicating a
recent monophyletic origin of triploid clones in this species.
Microsatellites proved to be useful tools for the investigation
of clonal diversity in asexual organisms.
MATERIAL AND METHODS
Multilocus DNA fingerprints
and 39 diploid fish from two different years from the Rio
For multilocus DNA fingerprints, 10 triploid
KATHRIN P. LAMPERT ET AL.
The known occurrence of triploid P. formosa clones is marked with a dotted line and magnified. The magnified map inset shows the
Soto la Marina river system with the study sites (I–V). Arrows depict sampling sites; capital letters, documented catches of triploid P.
formosa (map modified from Balsano et al. 1989).
Distribution of Poecilia formosa (framed in black) and its host species P. mexicana (dark gray) and P. latipinna (light gray).
Purification (Mexico) were used. We mainly sampled one
population with especially high P. formosa abundances (I).
At this site we sampled four different habitat types: Ia, fast
flowing main river; Ib, slow flowing side arm of the river;
Ic, large, deep, cool pond (only temporarily connected to the
main river after rainy season); Id, small pond (rarely con-
nected to the main river). In addition, we investigated fish
from four more localities (II–V) to represent the full known
range of P. formosa triploids (Fig. 1). Liver, brain, eyes, and
muscles of each individual were pooled and stored at ?80?C.
DNA was extracted according to Blin
and Stafford (1976). HinfI was used for restriction digestion,
and the fragments were separated on a 0.8% agarose gel at
1 V/cm. In-gel hybridization was done essentially as de-
scribed by Nanda et al. (1988). The32P-labeled oligonucle-
otides (GATA)4, (GGAT)4, (GA)8, and (CA)8were used as
tioned above were used. Fish were captured in 1996, 1998,
In total, 450 fishes from the eight sites men-
MONOPHYLY OF POECILIA FORMOSA TRIPLOIDS
and 2002 using seine nets. One (5 mm2) piece of the fast
regenerating dorsal fin was cut from each individual and
stored in 70% alcohol. Immediately after taking the biopsy,
animals were released otherwise unharmed. In addition, fins
of 25 P. mexicana (the sexual host of P. formosa in this
region) were used. Only the subspecies P. mexicana liman-
touri lives here (hereafter P. mexicana).
All fins were cut in half to use one
part to determine ploidy levels via flow cytometry. For this
fins were stained with DAPI, and DNA contents were deter-
mined using chicken blood as a reference, essentially as de-
scribed (Lamatsch et al. 2000b). DNA for polymerase chain
reaction (PCR) analysis was extracted according to Altsch-
mied et al. (1997). For isolation of primers that amplify in-
formative microsatellite loci in various Poecilia species, total
genomic DNA from P. latipinna (from Key Largo, FL) and
P. catemaconis (from Laguna Catemaco, Veracruz, Mexico)
was digested to completion with MboI and cloned into p-
BSKII?. A total of 3000 recombinant clones was individu-
ally spotted on replica plates using a robotic device. The
library was screened with oligonucleotide probes (GGAT)4,
(CA)8, (TAC)8and (TAA)8TA that were end-labeled with
?32P-ATP using terminal transferase. Positive clones were
isolated and sequenced. Primers were designed from the sin-
gle copy flanking sequences of simple repeat sequences that
had a repeat number greater than eight leading to primer-pair
Sat1 (forward primer: 5?ACAGCAGCTGTTCACGGC 3?,
backwards primer: 5?TCGGAGGAAAAACAGATGAC 3?
repeat motive: (CA)CGTA(CA)10allele size range: 106–
152bp). Additionally, cross-species reactions with primers
working in Xiphophorus hellerii (Seckinger et al. 2002) and
P. reticulata (Becher et al. 2002) were used to identify mi-
crosatellites in P. formosa. A total of 35 primer pairs was
tested, three of which revealed polymorphic loci in P. for-
mosa: KonD15 (Seckinger et al. 2002); PR39 (Becher et al.
2002), and Sat1. Polymerase chain reactions were done at:
90 sec of initial denaturation, 30 sec at 94?C, 30 sec at an-
nealing temperature (52?C for KonD15, 58?C for Sat1, 54?C
for PR39) and 30 sec at 72?C for 35 cycles followed by a
300 sec elongation. PCR products were analyzed on an ALF
Express sequencer (Amersham Biosciences, Freiburg, Ger-
viations from Hardy-Weinberg equilibrium for the sexual P.
mexicana were calculated using Arlequin (Schneider et al.
2000). Genetic diversity among triploid and diploid P. for-
mosa and P. mexicana was described following Menken et
al. (1995). We calculated the proportion of distinguishable
clones (PDC: number of clones divided by sample size; a
value of 1.00 means every individual has a different geno-
type), the effective number of genotypes (ENC: 1/?
frequency of the ith clone in a population), genotypic diver-
sity (CD: 1 ? ?
) and genotypic evenness (CE: (1/?
Nc, with Nc being the number of genotypes). Genotypic even-
ness (CE) gives a relative and therefore comparable measure
for the distribution of genotypes. This was important because
sample sizes varied extensively between ploidy levels. A CE
Population differentiation (FST) and de-
value of 1.00 means all clones are equally distributed. To
estimate the effect of additional microsatellites on the ability
of this method to detect genetic differentiation, we calculated
all measurements suggested by Menken et al. (1995) not only
for the final combination of three microsatellite loci, but also
for each locus and for each combination of two loci sepa-
We used several different approaches
to analyze the phylogenetic relationships between triploid
and diploid P. formosa clones and their sexual host, P. mex-
icana. The general genetic identity between diploid and trip-
loid P. formosa and P. mexicana was estimated using the
measurement proposed by Tomiuk and Loeschke (1991). This
measure was introduced to infer genetic identity between
asexual and sexual species and populations of differentploidy
levels. For a more detailed evaluation of relationships be-
tween individual genotypes we created neighbor-joining(Sai-
tou and Nei 1987) trees using an allele sharing index (DAS;
Chakraborty and Jin 1993) and Cavalli-Sforza and Edwards
chord distance (CSE; Cavalli-Sforza and Edwards 1967).
DAS was recommended for microsatellites by Goldstein et
al. (1995) as a useful tool for clustering related individuals,
and CSE by Takezaki and Nei (1996) as the most reliable
tool to generate tree topologies. Both methods have recently
been used to investigate phylogenetic relationships between
sexual and asexual populations of animals (Simon et al. 1999;
Delmotte et al. 2003). Calculations for the Tomiuk and Loes-
chke (1991) genetic identity matrix were done using the pro-
gram Popdist 1.1.1 (Guldbrandtsen et al. 2000). All other
distances and the neighbor-joining trees were calculated us-
ing the program Populations 1.2.28 (Langella 1999). Boot-
strapping values for the DAS and CSE trees were derived
from 1000 iterations. Because the reduction of discrete mor-
phologies to genetic distances might result in a loss of in-
formation, we additionally calculated a Wagner maximum
parsimony tree (Felsenstein 1989). This program uses binary
data; we therefore coded our microsatellite size data as n
discrete characters (n, number of alleles; Schalkwyk et al.
1999). To circumvent the problem of diploid and triploid
genotypes in the same dataset, we first calculated a tree cod-
ing every genotype as triploid. We added the same allele to
all diploid P. formosa and P. mexicana. Because this approach
might overestimate the degree of relatedness within the dip-
loids, we performed another analysis in which all triploid
individuals were treated as diploid by randomly leaving out
one of their three alleles in each locus. Bootstrapping values
were obtained from 1000 (all-triploid approach) and 1350
(all-diploid approach—50 per allele combination) replicates.
We used the programs Seqboot to generate the jackknife rep-
licates, MIX to generate the Wagner parsimony trees, and
Consense to generate a consensus tree from all the replicate
trees (Felsenstein 1989).
Ploidy level and genetic variability between years and
For all individuals of P. formosa, ploidy was determined
by flow cytometry. Ploidy levels varied between years: one
triploid and 20 diploids in the samples from 1996, 11 triploids
KATHRIN P. LAMPERT ET AL.
and 15 diploids in 1998, and 43 triploid and 316 diploid P.
formosa in 2002. In two of the eight populations no triploids
were found. Therefore, these sampling sites were excluded
from further analysis. For the remaining six sites sample sizes
varied greatly (n: Ia ? 85, Ib ? 105, Ic ? 60; Id ? 21; II
? 3; III ? 29). Because we could not find a significant level
of genetic diversification between the sampling sites (FST:
0.00–0.07; P ? 0.05), all data were pooled.
Clonal variability detected by multilocus DNA fingerprints
Multilocus DNA fingerprints of 10 triploid and 39 diploid
P. formosa revealed five genotypes in the triploids and 12
different diploid genotypes (Table 1). Statistical analysis re-
vealed a higher proportion of distinguishable clones in trip-
loids. The effective number of clones was clearly higher in
diploids than in triploids, whereas clonal diversity was only
slightly higher in diploids. Clonal evenness was higher in
triploids (Table 2A).
Low clonal variability of triploids detected by microsatellite
We found three polymorphic microsatellite loci for P. for-
mosa: Sat1, KonD15, and PR39. Analysis of their alleles
(Tables 1, 2B) led to the definition of 10 genotypes for primer
Sat1, seven for KonD15, and five in PR39 in diploid fish
(Table 1A). In triploids only two very similar genotypes were
found for each microsatellite (Table 1B). All diploids were
heterozygous at all three loci. Triploids always showed three
alleles. Genotypes varied highly in abundance. Intriguingly,
we found the same two triploid clones in all years and at all
sites. Triploid and diploid individuals shared a high number
of alleles: 71% (Table 1). An overall higher clonal variety
was observed in diploids (26 clones in 339 individuals) than
in triploids (two clones in 55 individuals). Consequently, a
higher proportion of distinguishable clones in diploids than
in triploids was found, as well as a higher number of effective
clones in diploids but lower clonal evenness than in triploids
Twenty-three genotypes were found for the 25 P. mexicana
investigated (Table 1C). They had an effective number of
21.6 different genotypes and a proportion of distinguishable
genotypes of 0.93. Genotypic diversity was very high and
different genotypes almost equally distributed (Table 2B). A
slight heterozygote deficiency in the Sat1 and KonD15 loci
was found in P. mexicana, probably due to the low number
of individuals scored. Increasing the number of microsatellite
loci revealed more genotypes in sexual (P. mexicana) and
asexual (P. formosa) diploids. Consequently, the proportion
of distinguishable genotypes increased. However, triploids
did not show any difference in their genotypic pattern wheth-
er one, two, or three loci were included in the analysis.
Microsatellites as tools to detect genetic variation even in
Since microsatellites have so far not been used extensively
to estimate genetic differentiation in clonal vertebrates, we
compared the microsatellite data with the results of the mul-
tilocus DNA fingerprints. A direct comparison of methods
was possible because 36 of the 49 individuals used for the
DNA fingerprint study were also used in the microsatellite
analysis. Microsatellites revealed fewer clones (six) than did
multilocus fingerprints (12). Some microsatellite clones di-
rectly represented a distinct clone (also found using multi-
locus fingerprints), though in some cases multilocus finger-
prints differentiated more than one clone in a microsatellite
line (Table 1).
The genetic identities calculated after the method of Tom-
iuk and Loeschke (1991) revealed low to medium levels of
relatedness between the sexual P. mexicana and the asexual
P. formosa genotypes, whereas the asexual groups (diploid
and triploids) showed a much higher genetic similarity (ge-
netic similarity: diploid/triploid ? 0.590; mexicana/diploid
? 0.088; mexicana/triploid ? 0.232).
The neighbor-joining trees created from the DAS and the
CSE genetic distances (data not shown) were very similar
and revealed three clearly distinct groups: P. mexicana ge-
notypes were separated from all asexual P. formosa geno-
types, while within the P. formosa clones, triploids were sep-
arated from diploids (bootstrapping values: 90%). The mi-
crosatellites, however, did not further resolve genotype re-
latedness within these major groups (all bootstrapping values
?70%). The results from the Wagner parsimony analysis
supported these findings. A consensus tree calculated from
three alleles per individual revealed strong support for a
monophyletic origin of the triploids (76%; data not shown).
The consensus tree calculated from two alleles per genotype
also found a monophyletic origin of the two triploid geno-
types to be the most likely solution, although bootstrapping
support for this solution was only 35% (Fig. 2).
A detailed look at allele sizes revealed that the diploid
HBB genotype and the triploid genotypes YYY shared a high
number of alleles. The triploid clone ZZZ still had one allele
in common with YYY and HBB in the Sat1 and KonD15
locus. In the PR39 locus it even shared two alleles with both
other genotypes. Although ZZZ and YYY shared only one
allele in the Sat1 locus, their allelic pattern was very similar.
YYY alleles that were not found in the HBB genotype were
very close (Sat1) or even similar in size (KonD15, PR39) to
alleles found in P. mexicana (Table 1).
Are microsatellites adequate tools to describe clonal
Since microsatellites have so far not been used to determine
genetic variability in asexual vertebrate species, we estimated
their value by comparing microsatellite results with the re-
sults of the multilocus DNA fingerprints. The three micro-
satellites developed for this study proved to be excellent tools
to investigate clonal diversity in P. formosa. Even though in
a direct comparison they identified fewer clones than did
multilocus fingerprints in the asexuals, their resolution was
good enough to differentiate between almost every sexual
individual. In addition, they allowed analysis of a large group
MONOPHYLY OF POECILIA FORMOSA TRIPLOIDS
mexicana (C). n, number of animals. Shared alleles between triploid and diploid genotypes are bold; shared alleles between P. formosa
and P. mexicana (with either diploids or triploids or both) are italic. In the last column multilocus DNA fingerprint genotypes (letters)
for 36 P. formosa that were analyzed with both methods are given (numbers of individuals).
Number of microsatellite genotypes and allele sizes found in 337 diploid (A) and 55 triploid (B) Poecilia formosa and in P.
Allele sizes at locus
24f(7); h(1); i(1)
A(1); C(4); C?(1)
1Significant heterozygote deficiency.
KATHRIN P. LAMPERT ET AL.
investigated, numbers of clones, and statistical analysis of data following Menken et al. (1995) are given. (B) Allele numbers, genotypes,
and statistical description of microsatellite data on genotypic variation in diploid and triploid P. formosa and P. mexicana. Numbers of
alleles represent the total number of different alleles found in the always heterozygous diploids and in the triploids that also differed in
all three alleles. All results for the three variable microsatellites and for their combined genotypes are given.
(A) Results from the multilocus DNA fingerprint analysis for diploid and triploid Poecilia formosa. Total number of fish
A. Multilocus DNA fingerprintsDiploidTriploid
No. of genotypes
Effective no. of genotypes (ENC)1
Proportion distinguishable (PDC)2
No. of unique genotypes3
Genotypic diversity (CD)4
Genotypic evenness (CE)5
B. MicrosatellitesDiploid Triploid
number of alleles
number of genotypes
number of alleles
number of genotypes
number of alleles
number of genotypes
no. of genotypes
effective no. of genotypes (ENC)1
proportion distinguishable (PDC)2
no. of unique genotypes3
genotypic diversity (CD)4
genotypic evenness (CE)5
All loci 26
1Calculated as 1/? p , pi being the frequency of the ith genotype.
2No. of genotypes divided by sample size.
3A unique genotype is represented by one individual.
4Calculated as 1 ? ? p , pi being the frequency of the ith genotype.
5Effective number of genotypes divided by number of genotypes.
of several hundred individuals and gave valuable additional
information by identifying all alleles of an individual.
Clonal variability of triploids
From the multilocus fingerprint data we had expected high
levels of clonal variability in triploids. However, microsat-
ellites revealed very low levels of genetic differentiation in
triploid P. formosa. These findings contradict a number of
earlier studies investigating clonal diversity in the asexual
fish Poeciliopsis (e.g., Vrijenhoek et al. 1978). Also, Turner
et al. (1983), using allozymes, found no difference in genetic
variability between diploid and triploid P. formosa. We pur-
posefully sampled all habitat types (see Ia–d) to exclude a
possible sampling bias due to differences in ecological niche
adaptations in different clones as proposed by Vrijenhoek
(1998). However, a more recent study of the Poeciliopsis
complex (Quattro et al. 1992) revealed low diversity in trip-
loids compared to diploids and therefore supports our finding.
How frequent are successful introgressions?
The neighbor-joining trees and the Wagner parsimony
analysis clearly point toward a single triploidization event.
All asexual lineages were separated from the sexual lineages,
and within the asexuals triploids formed their very own clade.
Overall bootstrap values were relatively low, most likely due
to the low number of informative microsatellite loci that are
available in this parthenogenetic species. However, the boot-
strap values for the branches separating the sexuals from the
nonsexuals, and the triploids from the diploids, were rea-
sonably high to be considered valid (90%). Constructing the
Wagner parsimony tree was somewhat difficult as diploid
and triploid genotypes could not be analyzed in the same
dataset. The result of the diploid-only consensus tree was
therefore very conservative, but still revealed the monophy-
letic origin as the most parsimonious solution, even though
the bootstrapping support was not overwhelming. Since both
ways of creating phylogenetic trees showed the same tree,
which was additionally supported by the Tomiuk and Loesch-
ke (1991) genetic identity, we are confident that those trees
give a reliable picture of the actual relatedness of clones.
More support for a monophyletic origin of the triploid
clones came from the allelic patterns observed in the two
triploid clones. Both triploid clones shared at least one allele
at every locus. Alleles that were not identical showed a very
similar pattern that could be explained by subsequent mu-
tations. It seems highly unlikely that two independent intro-
gression events converged by mutation to such a similar al-
lelic pattern. Furthermore, in the whole geographic range of
P. formosa, triploid clones are found in only one river system
(Schlupp et al. 2002) and new triploidizations in laboratory
strains are always sterile (Nanda et al. 1995). A random com-
bination of genotypes is obviously not sufficient to produce
fertile new clones. Considering the rarity of hybridizations
leading to successful unisexual lineages (Dawley and Bogart
1989), most likely due to the genetic (Moritz et al. 1989)
and ecological (Vrijenhoek 1989) constraints on asexuality,
it seems likely that the origin of triploids in P. formosa was
restricted to a single introgression event. Similar arguments
have been used by Quattro et al. (1992) to explain the origin
MONOPHYLY OF POECILIA FORMOSA TRIPLOIDS
allele combination; see text).
Wagner parsimony tree for all genotypes. Bootstrapping values are derived from 1350 randomizations (50 trees per triploid
of triploids in the Poeciliopsis complex. Turner et al. (1990)
interpreted the extraordinarily high level of clonal diversity
found in a DNA fingerprint study as evidence of mutation as
a driving force in variation processes in P. formosa. Other
authors also proposed that mutations are more likely to be
the source of variation rather than multiple hybridizations
(Kallman 1962; Schartl et al. 1995a). The rarity of successful
introgressions and the lack of tetraploids or higher ploids
makes the hypothesis of low selective pressure against po-
lyploidization in asexuals rather unlikely.
Is triploidization an ancient or recent event in Poecilia
The high proportion of shared alleles in triploids and dip-
loids (71%) points to a rather recent origin of the triploid
clones in P. formosa (Birky 1996). Diploids, however,
showed a high number of allelic divergences and are therefore
likely to be more ancient. We identified HBB as a clone
closely similar to the likely diploid ancestor with the triploid
clone YYY. This assumption was based on the striking sim-
ilarity of alleles. Also, the close proximity of the ‘‘third’’
alleles in triploid P. formosa to alleles found in P. mexicana
lends support to the assumption of a rather recent introgres-
sion of a P. mexicana genome. Triploids seem to be old
enough to be slightly different in one of the three loci but
still rather young considering the close proximity to each
other and the diploid HBB clone.
Triploidization is often interpreted as being advantageous
for asexually reproducing organisms, as it might increase the
ecological and evolutionary potential by providing additional
genetic variability. It was generally assumed that diploid P.
formosa arose from a single or very few hybridization events
whereas subsequent triploidization was a more common event
(Turner et al. 1990). Our data, however, point in a different
direction: the low amount of variation in all of the micro-
satellite loci in triploids provides evidence for a relatively
recent monophyletic origin. Thus, it is reasonable to assume
that mutations rather than multiple hybridization events are
the major source of genetic variation and divergence in the
triploid clones of P. formosa. Triploidization seems unlikely
to be P. formosa’s main ‘‘strategy’’ for overcoming the mu-
tational meltdown predicted for its species (Lynch et al.
1993). However, triploidization definitely contributes to ge-
netic diversification and might, therefore, have a key function
by creating multiple starting points for mutations. On an in-
dividual basis, being triploid might still be advantageous,
even though on a population scale diploids seem to be more
KATHRIN P. LAMPERT ET AL.
diverse than triploids. Even rare diversification events may
be sufficient to slow down Muller’s ratchet (Muller 1964).
We are grateful to the Mexican government for issuing
permit no. 210696-213-03 to collect P. formosa, and to J.
Parzefall, M. Do ¨bler, I. Schlupp, U. Hornung, and A. Fros-
chauer for help in the field. For help with the laboratory work
we thank S. Schories. B. Wilde developed the microsatellite
Sat1 used in this study. We thank H. Schwind, G. Schneider,
and P. Weber for breeding the fish in the laboratory, and C.
West Wheat and N. Pongratz for critically reading the man-
uscript. Financial support for this study was granted by the
Deutsche Forschungsgemeinschaft (SFB 567 Mechanismen
der interspezifischen Interaktion von Organismen) and Fonds
der Chemischen Industrie. We thank M. Schmid for the pos-
sibility to use the flow cytometer.
Altschmied, J., U. Hornung, I. Schlupp, J. Gadau, R. Kolb, and M.
Schartl. 1997. Isolation of DNA suitable for PCR for field and
laboratory work. Biotechniques 23:228–229.
Avise, J. C., J. C. Trexler, J. Travis, and W. S. Nelson. 1991.
Poecilia mexicana is the recent female parent of the unisexual
fish Poecilia formosa. Evolution 45:1530–1533.
Balsano, J. S., E. M. Rasch, and P. J. Monaco. 1989. The evolu-
tionary ecology of Poecilia formosa and its triploid associate.
Pp. 277–297 in G. K. Meffe and F. F. Snelson, Jr., eds. Ecology
and evolution of livebearing fishes (Poeciliidae). Prentice Hall,
Englewood Cliffs, NJ.
Becher, S. A., S. T. Russell, and A. E. Magurran. 2002. Isolation
and characterization of polymorphic microsatellites in the Trin-
idadian guppy (Poecilia reticulata). Mol. Ecol. Notes 2:456–458.
Birky, W. C., Jr. 1996. Heterozygosity, heteromorphy, and phy-
logenetic trees in asexual eukaryotes. Genetics 144:427–437.
Blin, N., and D. W. Stafford. 1976. A general method for isolation
of high molecular weight DNA from eukaryotes. Nucleic Acids
Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic
analysis: Models and estimation procedures. Am. J. Hum. Genet.
Chakraborty, R., and L. Jin. 1993. A unified approach to study
hypervariable polymorphisms: Statistical considerations of de-
termining relatedness and population distances. Pp.153–175 in
S. D. J. Pena, R. Chakraborty, J. T. Epplen, and A. J. Jeffreys,
eds. DNA fingerprinting: State of the science. Birkhauser, Basel,
Dawley, R. M. 1989. An introduction to unisexual vertebrates. Pp.
1–18 in Evolution and ecology of unisexual vertebrates. R. M.
Dawley and J. P. Bogart, eds. New York State Museum, Albany,
Dawley, R. M., and J. P. Bogart, eds. 1989. Evolution and ecology
of unisexual vertebrates. New York State Museum, Albany, NY.
Delmotte, F., B. Sabater-Mun ˜oz, N. Prunier-Leterme, A. Latorre,
P. Sunnucks, C. Rispe, and J.-C. Simon. 2003. Phylogenetic
evidence for hybrid origins of asexual lineages in an aphid spe-
cies. Evolution 57:1291–1301.
Felsenstein, J. 1989. PHYLIP: Phylogeny inference package (Ver-
sion 3.2). Cladistics 5:164–166.
Goldstein, D. B., A. R. Linares, L. L. Cavalli-Sforza, and M. W.
Feldman. 1995. An evaluation of genetic distances for use with
microsatellite loci. Genetics 139:463–471.
Guldbrandtsen, B., J. Tomiuk, and V. Loeschcke. 2000. POPDIST,
Version 1.1.1: A program to calculate population genetic dis-
tance and identity measures. J. Hered. 91:178–179.
Hubbs, C. L., and L. C. Hubbs. 1932. Apparent parthenogenesis in
nature, in a form of fish of hybrid origin. Science 76:628–630.
Kallman, K. D. 1962. Population genetics of the gynogenetic tel-
eost, Mollienesia formosa (Girard). Evolution 16:497–504.
Lamatsch, D. K., I. Nanda, J. T. Epplen, M. Schmid, and M. Schartl.
2000a. Unusual triploid males in a microchromosome-carrying
clone of the Amazon molly, Poecilia formosa. Cytogenet. Cell
Lamatsch, D. K., C. Steinlein, M. Schmid, and M. Schartl. 2000b.
Noninvasive determination of genome size and ploidy level in
fishes by flow cytometry: Detection of triploid Poecilia formosa.
Langella, O. 1999. Populations 1.2.28. Available via http://
Lynch, M., R. Bu ¨rger, D. Butcher, and W. Gabriel. 1993. The mu-
tational meltdown in asexual populations. J. Hered. 84:339–344.
Menken, S. B., E. Smit, and H.(J.) C. M. Den Nijs. 1995. Genetical
population structure in plants: Gene flow between diploid sexual
and triploid asexual dandelions (Taraxacum section Ruderalia).
Mogie, M. 1986. On the relationship between asexual reproduction
and polyploidy. J. Theor. Biol. 122:493–498.
Monaco, P. J., E. M. Rasch, and J. S. Balsano. 1984. Apomictic
reproduction in the Amazon molly, Poecilia formosa, and its
triploid hybrids. Pp. 311–318 in B. J. Turner, ed. Evolutionary
genetics of fishes. Plenum Press, New York.
Moritz, C., W. M. Brown, L. D. Densmore, J. W. Wright, D. Vyas,
S. Donnellan, M. Adams, P. Braverstock. 1989. Genetic diversity
and the dynamics of hybrid parthenogenesis in Cnemidophorus
(Teiidae) and Heterodontia (Gekkonidae). Pp. 87–112 in R. M.
Dawley and J. P. Bogart, eds. Evolution and ecology of unisexual
vertebrates. New York State Museum, Albany, NY.
Muller, H. J. 1964. The relation of recombination to mutational
advance. Mut. Res. 1:2–9.
Nanda, I., H. Neitzel, K. Sperling, R. Studer, and J. T. Epplen.
1988. Simple GATCA repeats characterize the X chromosomal
heterochromatin of Microtus agrestis, European field vole (Ro-
dentia, Cricetidae). Chromosoma 96:213–219.
Nanda, I., M. Schartl, W. Feichtinger, I. Schlupp, J. Parzefall, and
M. Schmid. 1995. Chromosomal evidence for laboratory syn-
thesis of a triploid hybrid between the gynogenetic teleost Poe-
cilia formosa and its host species. J. Fish Biol. 47:619–623.
Quattro, J. M., J. C. Avise, and R. C. Vrijenhoek. 1992. Mode of
origin and sources of genotypic diversity in triploid gynogenetic
fish clones (Poeciliopsis: Poeciliidae). Genetics 130:621–628.
Rasch, E. M., and J. S. Balsano. 1974. Biochemical and cytogenetic
studies of Poecilia from eastern Mexico. II. Frequency, perpet-
uation, and probable origin of triploid genomes in females as-
sociated with Poecilia formosa. Rev. Biol. Trop. 21:351–381.
Saitou, N., and M. Nei. 1987. The neighbor-joining method: A new
method for reconstructing phylogenetic trees. Mol. Biol. Evol.
Schalkwyk, L. C., M. Jung, A. Daser, M. Weiher, J. Walter, H.
Himmelbauer, and H. Lehrach. 1999. Panel of microsatellite
markers for whole-genome scans and radiation hybrid mapping
and a mouse family tree. Genome Res. 9:878–887.
Schartl, M., I. Nanda, I. Schlupp, M. Schmid, and J. T. Epplen.
1990. Genetic variation in the clonal vertebrate Poecilia formosa
is limited to few truly hypervariable loci. Fingerprint News 2:
Schartl, M., B. Wilde, I. Schlupp, and J. Parzefall. 1995a. Evolu-
tionary origin of a parthenoform, the Amazon molly Poecilia
formosa, on the basis of a molecular genealogy. Evolution 49:
Schartl, M., I. Nanda, I. Schlupp, B. Wilde, J. T. Epplen, M. Schmid,
and J. Parzefall. 1995b. Incorporation of subgenomic amounts
of DNA as compensation for mutational load in a gynogenetic
fish. Nature 373:68–71.
Schlupp, I., J. Parzefall, and M. Schartl. 2002. Biogeography of the
Amazon molly, P. formosa. J. Biogeogr. 29:1–6.
Schneider, S., D. Roessli, and L. Excoffier. 2000. Arlequin. Ver.
2.000. Software for population genetic data analysis. Genetics
and Biometry Laboratory, University of Geneva, Switzerland.
Seckinger, J., H. Brinkmann, and A. Meyer. 2002. Microsatellites
MONOPHYLY OF POECILIA FORMOSA TRIPLOIDS
in the genus Xiphophorus, developed in Xiphophorus montezu-
mae. Mol. Ecol. Notes 2:4–6.
Simon, J.-C., S. Baumann, P. Sunnucks, P. D. N. Herbert, J.-S.
Pierre, J.-F. Le Gallic, and C.-A. Dedryver. 1999. Reproductive
mode and population genetic structure of the cereal aphid Si-
tobion avenae studied using phenotypic and microsatellite mark-
ers. Mol. Ecol. 8:531–545.
Takezaki, N., and M. Nei. 1996. Genetic distance and reconstruction
of phylogenetic trees from microsatellite DNA. Genetics 144:
Tomiuk, J., and V. Loeschke. 1991. A new measure for genetic
identity between populations of sexual and asexual species. Evo-
Turner, B. J., B. L. H. Brett, E. M. Rasch, and J. S. Balsano. 1980a.
Evolutionary genetics of gynogenetic fish, Poecilia formosa, the
Amazon molly. Evolution 34:246–258.
Turner, B. J., B. L. H. Brett, and R. R. Miller. 1980b. Interspecific
hybridization and the evolutionary origin of a gynogenetic fish,
Poecilia formosa. Evolution 34:917–922.
Turner, B. J., J. S. Balsano, P. J. Monaco, E. M. Rasch. 1983. Clonal
diversity and evolutionary dynamics in a diploid-triploid breed-
ing complex of unisexual fishes (Poecilia). Evolution 37:
Turner, B. J., J. F. Endler Jr., T. F. Laughlin, and W. P. Davis.
1990. Genetic variation in clonal vertebrates detected by simple-
sequence DNA fingerprinting. Proc. Natl. Acad. Sci. USA 87:
Vrijenhoek, R. C. 1989. Genetic and ecological constraints in the
origins and establishment of unisexual vertebrates. Pp. 24–31 in
R. M. Dowley and J. P. Bogart, eds. Evolution and ecology of
unisexual vertebrates. New York State Museum, Albany, NY.
Vrijenhoek, R. C. 1998. Animal clones and diversity. BioScience
Vrijenhoek, R. C., R. A. Angus, and R. J. Schultz. 1978. Variation
and clonal structure in a unisexual fish. Am. Natl. 112:41–55.
Corresponding Editor: G. Ortı ´