Mitochondrial Genetic Analysis by Zygote Cell Lineages in SACCHAROMYCES CEREVISIAE

Yeast strains were constructed carrying multiple mitochondrial markers conferring resistance to the inhibitors erythromycin, chloramphenicol, paromomycin and oligomycin. A pedigree analysis of two crosses was made by micromanipulating buds from zygotes. The first few daughter buds isolated from the zygotes sometimes gave rise to diploid clones which had a mixture of mitochondrial types. All possible classes of mitochondrial parental and recombinant types were found although they never appeared all together as the progeny from a single zygote. It was inferred that multiple recombination events took place in zygotes and in some of the buds derived from them. After removal of the first four or so daughter buds, subsequent buds from the zygote carried one mitochondrial type only. In cross I in which three markers were analyzed this was most frequently one of the parental types. In cross II (involving four mitochondrial markers) the later buds from the zygotes were frequently of recombinant mitochondrial type.     

The effect of zygotic bud position on the transmission of mitochondrial genes in Saccharomyces cerevisiae.

Summary Segregation of mitochondrial genomes in yeast zygotes has been investigated by partial pedigree analysis of crosses involving the markers cap, ery, oli1 and par. The results demonstrate that the segregation pattern of markers is non-random during the first zygote generation and is directly related to slow mixing of the zygote cytoplasm. We have observed that a first bud may be formed at the center or either end of the dumbbell-shaped zygote. Cytoplasmic mixing is particularly slow in those zygotes producing first end buds.Clones derived from first end buds are usually pure (or nearly so) for a parental genotype and so detectable recombination of mitochondrial markers is reduced in these zygotes. Cells derived from a zygote after removal of a first end bud are predominantly of the other parental genotype. This observation suggests that a large fraction of the available segregating units enters each first bud and illustrates one means of obtaining complete segregation (even in multi-factor crosses) at the first generation. First center buds generally receive mitochondrial markers from both parents and the recombination frequency in such clones (and the clones derived from isolated first center buds) is significantly higher than in similar clones from zygotes with first end buds. Therefore, the distribution of first bud positions within a population of zygotes can influence the recombination frequency between mitochondrial loci. The delay in cytoplasmic mixing in combination with certain patterns of zygotic budding can distort the relationship between input of mitochondrial genomes and the output of a cross.The phage analogy model of yeast mitochondrial genetics has been re-examined in light of these data. The assumption of rapid panmixis is not supported by the data from any of the crosses analyzed here. Since panmixis is most closely approximated in zygotes with first center buds, crosses with predominantly zygotes of that type may be the ones where the model is most applicable.     

The mitochondrial genome of wild-type yeast cells. VII. Recombination in crosses.

Two Saccharomyces cerevisiae wild-type strains were crossed, and 26 diploid clones were obtained from (1) mass mating; (2) individual buds in zygote lineages; (3) individual zygotes. The mitochondrial DNAs from these diploids were investigated in their recombination and segregation by analyzing their restriction fragment patterns.Recombinant mitochondrial genomes were present in 75% of the diploid clones. Such recombinant genomes had unit sizes different from, yet within ± 5% of, the parental ones and showed EcoRI and HindII + III fragment patterns of parental types, two strong indications that both the gene complement and the gene order were very largely preserved in the progeny.Fragment patterns produced by HpaII and HaeIII were characterized by (1) fragments originating from the DNAs of both parents; and (2) new fragments, namely fragments absent in either parent. The new fragments appear to arise from unequal crossing-over events occurring in the spacers of allelic parental genetic units and usually have preferential localizations in the genome.These results provide the first evidence for physical recombinations of mitochondrial DNA in crosses of wild-type yeast cells, indicate that recombination is very frequent in crosses, and shed some light on mitochondrial segregation. They also have interesting implications for recombination phenomena in interspersed systems of unique and repetitive nucleotide sequences.     

Kinetic and segregational analysis of mitochondrial DNA recombination in yeast

A pair of yeast strains of opposite mating type was constructed to contain polymorphisms at three loci on the mitochondrial genome--the 21 S rRNA gene, var1, and cob--such that parental and recombinant forms of these genes could be easily detected by Southern blot analysis. These polymorphisms were used to measure in a single cross gene conversions at the 21 S rRNA and var1 loci and a reciprocal recombination at cob. For all three loci, recombination initiates at about the same time, 4 to 6 h after mixing cells, and increases with similar kinetics over a 24-h period. The segregation of parental and recombinant forms of these genes was then followed by pedigree analysis. The results, which show a high variance in the distribution of parental and recombinant forms of all three genes in cells derived from both the first bud and the mother zygote, are consistent with the segregation of a small number of mitochondrial DNA molecules from the zygote to diploid buds. Based on these results and previous experiments of this type, a limited "zone of mixing" of parental mitochondrial DNA molecules probably exists in the zygote. The extent of sampling from this zone, together with the intrinsic properties of the recombination events themselves, is likely to determine the observed pattern of recombination of mitochondrial DNA sequences at the population level.     

Patterns of mitochondrial sorting in yeast zygotes.

Inheritance of mitochondrial DNA (mtDNA) in Saccharomyces cerevisiae is usually biparental. Pedigree studies of zygotic first buds indicate limited mixing of wild-type (p+) parental mtDNAs: end buds are frequently homoplasmic for one parental mtDNA, while heteroplasmic and recombinant progeny usually arise from medial buds. In crosses involving certain petites, however, mitochondrial inheritance can be uniparental. In this study we show that mitochondrial sorting can be influenced by the parental mtDNAs and have identified intermediates in the process. In crosses where mtDNA mixing is limited and one parent is prelabeled with the matrix enzyme citrate synthase 1 (CS1), the protein freely equilibrates throughout the zygote before the first bud has matured. Furthermore, if one parent is p0 (lacking mtDNA), mtDNA from the p+ parent can also equilibrate; intracellular movement of mtDNA is unhindered in this case. Surprisingly, in zygotes from a p0 CS1+ x p+ CS1- cross, CS1 is quantitatively translocated to the p+ end of the zygote before mtDNA movement; subsequently, both components equilibrate throughout the cell. This initial vectorial transfer does not require respiratory function in the p+ parent, although it does not occur if that parent is p-. Mouse dihydrofolate reductase (DHFR) present in the mitochondrial matrix can also be vectorially translocated, indicating that the process is general. Our data suggest that in zygotes mtDNA movement may be separately controlled from the movement of bulk matrix constituents.     

Evidence of animal mtDNA recombination between divergent populations of the potato cyst nematode Globodera pallida

Recombination is typically assumed to be absent in animal mitochondrial genomes (mtDNA). However, the maternal mode of inheritance means that recombinant products are indistinguishable from their progenitor molecules. The majority of studies of mtDNA recombination assess past recombination events, where patterns of recombination are inferred by comparing the mtDNA of different individuals. Few studies assess contemporary mtDNA recombination, where recombinant molecules are observed as direct mosaics of known progenitor molecules. Here we use the potato cyst nematode, Globodera pallida, to investigate past and contemporary recombination. Past recombination was assessed within and between populations of G. pallida, and contemporary recombination was assessed in the progeny of experimental crosses of these populations. Breeding of genetically divergent organisms may cause paternal mtDNA leakage, resulting in heteroplasmy and facilitating the detection of recombination. To assess contemporary recombination we looked for evidence of recombination between the mtDNA of the parental populations within the mtDNA of progeny. Past recombination was detected between a South American population and several UK populations of G. pallida, as well as between two South American populations. This suggests that these populations may have interbred, paternal mtDNA leakage occurred, and the mtDNA of these populations subsequently recombined. This evidence challenges two dogmas of animal mtDNA evolution; no recombination and maternal inheritance. No contemporary recombination between the parental populations was detected in the progeny of the experimental crosses. This supports current arguments that mtDNA recombination events are rare. More sensitive detection methods may be required to adequately assess contemporary mtDNA recombination in animals.     

Recombination Occurs Mainly Between Parental Genomes and Precedes DNA Replication in Pseudorabies Virus-Infected Cells

The experiments described in this paper were part of an attempt to determine the mechanisms involved in the isomerization of the pseudorabies virus genome. To this end, [(14)C]thymidine-labeled parental virus DNA that was transferred to progeny virions produced by cells incubated in medium containing bromodeoxy-uridine was analyzed in neutral and alkaline CsCl density gradients. The buoyant density of the (14)C-labeled DNA indicated that the parental DNA strands had retained their integrity and had not undergone breakage and reunion with progeny DNA strands; neither massive intermolecular nor intramolecular recombination had occurred after replication of the DNA. Whereas breakage and reunion between parental and progeny virus DNA strands were not detectable, these processes were observed between differentially density-labeled parental DNAs. Furthermore, the frequency of recombination between progeny DNAs accumulating in the cells was low. These results indicate that in pseudorabies virus-infected rabbit kidney cells recombination occurs mainly between parental genomes and precedes DNA replication. An analysis of the kinetics of appearance of recombinants between pairwise combinations of temperature-sensitive mutants also indicated that recombination is an early event. The ratio between the number of recombinant virions and the number of temperature-sensitive mutant virions produced by the cells remained the same throughout infection. Since the relative amounts of viral DNAs synthesized early and late during the infective process that were integrated into virions were approximately the same, it appears that late viral DNA did not experience an increased number of recombinational events compared with early viral DNA. These results, which reinforce the conclusion reached from the results of the analysis of the behavior of the parental DNA molecules in density shift experiments, indicate that recombination is an early event.     

Vegetative segregation of mitochondria in yeast: Estimating parameters using a random model

Summary Yeast zygotes which are heteroplasmic for mitochondrial genes reproduce vegetatively to form clones of diploid progeny which are homoplasmic. This vegetative segregation of mitochondrial genes has been interpreted in terms of a random distribution of mitochondria or mitochondrial genomes between mother and bud at cell division. We have developed equations which permit calculation of the number of segregating units in the zygote and the number of those units which enter the bud, assuming that segregation of the units is genetically random and numerically variable or equal. Use of the equations requires data from partial pedigree analyses: we isolate zygotes, separate the first bud, then determine the frequency of mitochondrial alleles among the progeny of mother cells whose first buds were homoplasmic. Application of this method to data from five crosses suggests that most zygotes have a small number of segregating units (usually less than a dozen) and only one or two enter the first bud. Analysis of the frequency of buds which are nearly but not quite homoplasmic indicates that the segregating units may be mitochondria or portions thereof which include many mitochondrial genomes, all the genomes in a unit being genetically identical in most but not all cases. These results are compatible with, but do not prove, the hypothesis of random vegetative segregation of mitochondria.     

Mitochondrial fusion and inheritance of the mitochondrial genome

Although maternal or uniparental inheritance of mitochondrial genomes is a general rule, biparental inheritance is sometimes observed in protists and fungi, including yeasts. In yeast, recombination occurs between the mitochondrial genomes inherited from both parents. Mitochondrial fusion observed in yeast zygotes is thought to set up a space for DNA recombination. In the last decade, a universal mitochondrial fusion mechanism has been uncovered, using yeast as a model. On the other hand, an alternative mitochondrial fusion mechanism has been identified in the true slime mold Physarum polycephalum. A specific mitochondrial plasmid, mF, has been detected as the genetic material that causes mitochondrial fusion in P. polycephalum. Without mF, fusion of the mitochondria is not observed throughout the life cycle, suggesting that Physarum has no constitutive mitochondrial fusion mechanism. Conversely, mitochondria fuse in zygotes and during sporulation with mF. The complete mF sequence suggests that one gene, ORF640, encodes a fusogen for Physarum mitochondria. Although in general, mitochondria are inherited uniparentally, biparental inheritance occurs with specific sexual crossing in P. polycephalum. An analysis of the transmission of mitochondrial genomes has shown that recombinations between two parental mitochondrial genomes require mitochondrial fusion, mediated by mF. Physarum is a unique organism for studying mitochondrial fusion.     

Inheritance of Mitochondrial DNA and Plasmids in the Ascomycetous Fungus, Epichloe Typhina

We analyzed the inheritance of mitochondrial DNA (mtDNA) species in matings of the grass symbiont Epichloe typhina. Eighty progeny were analyzed from a cross in which the maternal (stromal) parent possessed three linear plasmids, designated Callan-a (7.5 kb), Aubonne-a (2.1 kb) and Bergell (2.0 kb), and the paternal parent had one plasmid, Aubonne-b (2.1 kb). Maternal transmission of all plasmids was observed in 76 progeny; two progeny possessed Bergell and Callan-a, but had the maternal Aubonne-a replaced with the related paternal plasmid Aubonne-b; two progeny lacked Callan-a, but had the other two maternal plasmids. A total of 34 progeny were analyzed from four other matings, including a reciprocal pair, and in each progeny the plasmid transmission was maternal. The inheritance of mitochondrial genomes in all progeny was analyzed by profiles of restriction endonuclease-cleaved mtDNA. In most progeny the profiles closely resembled those of the maternal parents, but some progeny had nonparental mtDNA profiles that suggested recombination of mitochondrial genomes. These results indicate that the fertilized stroma of E. typhina is initially heteroplasmic, permitting parental mitochondria to fuse and their genomes to recombine.     

The non-reciprocality of organelle gene recombination in Chlamydomonas reinhardtii and Saccharomyces cerevisiae: some new observations and a restatement of some old problems.

Summary Organelle recombinant genotype frequencies, derived from analysis of individual mitotic zygote clones of Chlamydomonas reinhardtii and Saccharomyces cerevisiae, were subjected to two types of statistical tests in an attempt to detect the occurrence of reciprocal recombination: (i) calculation of correlation coefficients for the frequencies of two recombinant genotypes (reciprocal or non-reciprocal pairs) within individual zygote clones, and (ii) application of the chi-square test for independence to the frequencies of zygotes yielding one or the other, neither, or both of a given recombinant pair.Applying test (i), the strongest correlations are found for non-reciprocal rather than reciprocal pairs. When the data are analyzed by method (ii), some reciprocal as well as non-reciprocal pairs appear to be produced concurrently in zygote clones. However, such deviations from independence are greatest for non-reciprocal pairs. These tests yield comparable results for yeast mitochondrial and Chlamydomonas chloroplast gene recombination, and provide no convincing evidence for reciprocal genetic exchange. Explanations for the observed lack of reciprocality are discussed with reference both to our present understanding of the molecular events responsible for genetic recombination, and to the problems which may be unique to the analysis of organelle gene recombination.     

Genetic evidence for nonrandom sorting of mitochondria in the basidiomycete Agrocybe aegerita.

We studied mitochondrial transmission in the homobasidiomycete Agrocybe aegerita during plasmogamy, vegetative growth, and basidiocarp differentiation. Plasmogamy between homokaryons from progeny of three wild-type strains resulted in bidirectional nuclear migration, and the dikaryotization speed was dependent on the nuclear genotype of the recipient homokaryon. Little mitochondrial migration accompanied the nuclear migration. A total of 75% of the dikaryons from the fusion lines had both parental mitochondrial haplotypes (mixed dikaryons), and 25% had only a single haplotype (homoplasmic dikaryons); with some matings, there was a strong bias in favor of one parental haplotype. We demonstrated the heteroplasmic nature of mixed dikaryons by (i) isolating and subculturing apical cells in micromanipulation experiments and (ii) identifying recombinant mitochondrial genomes. This heteroplasmy is consistent with the previously reported suggestion that there is recombination between mitochondrial alleles in A. aegerita. Conversion of heteroplasmons into homoplasmons occurred (i) during long-term storage, (ii) in mycelia regenerated from isolated apical cells, and (iii) during basidiocarp differentiation. Homokaryons that readily accepted foreign nuclei were the most efficient homokaryons in maintaining their mitochondrial haplotype during plasmogamy, long-term storage, and basidiocarp differentiation. This suggests that the mechanism responsible for the nonrandom retention or elimination of a given haplotype may be related to the nuclear genotype or the mitochondrial haplotype or both.     

Uniparental Inheritance of Mitochondrial Genes in Yeast: Dependence on Input Bias of Mitochondrial DNA and Preliminary Investigations of the Mechanism

In Saccharomyces cerevisiae, previous studies on the inheritance of mitochondrial genes controlling antibiotic resistance have shown that some crosses produce a substantial number of uniparental zygotes, which transmit to their diploid progeny mitochondrial alleles from only one parent. In this paper, we show that uniparental zygotes are formed especially when one parent (majority parent) contributes substantially more mitochondrial DNA molecules to the zygote than does the other (minority) parent. Cellular contents of mitochondrial DNA (mtDNA) are increased in these experiments by treatment with cycloheximide, alpha-factor, or the uvsp5 nuclear mutation. In such a biased cross, some zygotes are uniparental for mitochondrial alleles from the majority parent, and the frequency of such zygotes increases with increasing bias. In two- and three-factor crosses the cap1, ery1, and oli1 loci behave coordinately, rather than independently; minority markers tend to be transmitted or lost as a unit, suggesting that the uniparental mechanism acts on entire mtDNA molecules rather than on individual loci. This rules out the possibility that uniparental inheritance can be explained by the conversion of minority markers to the majority alleles during recombination. Exceptions to the coordinate behavior of different loci can be explained by marker rescue via recombination. Uniparental inheritance is largely independent of the position of buds on the zygote. We conclude that it is due to the failure of minority markers to replicate in some zygotes, possibly involving the rapid enzymatic destruction of such markers. We have considered two general classes of mechanisms: (1) random selection of molecules for replication, as for example by competition for replicating sites on a membrane; and (2) differential marking of mtDNA molecules in the two parents, possibly by modification enzymes, followed by a mechanism that "counts" molecules and replicates only the majority type. These classes of models are distinguished genetically by the fact that the first predicts that the output frequency of a given allele among the progeny of a large number of zygotes will approximately equal the average input frequency of that allele, while the second class predicts that any input bias will be amplified in the output. The data suggest that bias amplification does occur. We hypothesize that maternal inheritance of mitochondrial or chloroplast genes in many organisms may depend upon a biased input of organelle DNA molecules, which usually favors the maternal parent, followed by failure of the minority (paternal) molecules to replicate in many or all zygotes.     

Recombination in Mitochondrial DNA of European Mussels <Emphasis Type="Italic">Mytilus</Emphasis>

Mitochondrial DNA was long believed to be purely clonal and free from recombination. Major phylogenetic studies still depend on such assumptions. The peculiar genetic system of marine mussels Mytilus in which two divergent mitochondrial genomes exist provides a unique opportunity to study mtDNA recombination. Previous reports showed the existence of a few haplotypes having very strong recombination signal in the control region of mtDNA. Those recombinant variants have been found in a Baltic Sea population of Mytilus trossulus as well as in Mytilus galloprovincialis from the Black Sea. In both cases the mosaic genomes switched their transmission route and have been inherited paternally. In the present study rearranged mtDNA genomes found in all three European Mytilus species are described. The structure of their control region is a result of intra- and intermolecular recombination between mitochondrial genomes. Together with the phylogenetic reconstruction and geographic distribution, this suggests that two interlineage recombination events have occurred in the control region of mtDNA of European mussels Mytilus. Contrary to earlier observations, some of the mosaic genomes do not show any gender bias, which has important implications regarding the transmission and evolution of blue mussel mitochondrial genomes.     

Biparental inheritance of organelles in Pelargonium: evidence for intergenomic recombination of mitochondrial DNA

While uniparental transmission of mtDNA is widespread and dominating in eukaryotes leaving mutation as the major source of genotypic diversity, recently, biparental inheritance of mitochondrial genes has been demonstrated in reciprocal crosses of Pelargonium zonale and P. inquinans. The thereby arising heteroplasmy carries the potential for recombination between mtDNAs of different descent, i.e. between the parental mitochondrial genomes. We have analyzed these Pelargonium hybrids for mitochondrial intergenomic recombination events by examining differences in DNA blot hybridization patterns of the mitochondrial genes atp1 and cob. Further investigation of these genes and their flanking regions using nucleotide sequence polymorphisms and PCR revealed DNA segments in the progeny, which contained both P. zonale and P. inquinans sequences suggesting an intergenomic recombination in hybrids of Pelargonium. This turns Pelargonium into an interesting subject for studies of recombination and evolutionary dynamics of mitochondrial genomes.     

Transmission of yeast mitochondrial loci to progeny is reduced when nearby intergenic regions containing ori sequences are deleted

Summary Mitochondrial DNAs (mtDNA) from four stable revertant strains generated from high frequency petite forming strains of Saccharomyces cerevisiae have been shown to contain deletions which have eliminated intergenic sequences encompassing ori1, ori2 and ori7. The deleted sequences are dispensable for expression of the respiratory phenotype and mutant strains exhibit the same relative amount of mtDNA per cell as the wild-type (wt) parental strain. These deletion mutants were also used to study the influence of particular intergenic sequences on the transmission of closely linked mitochondrial loci. When the mutant strains were crossed with the parental wt strains, there was a strong bias towards the transmission into the progeny of mitochondrial genomes lacking the intergenic deletions. The deficiency in the transmission of the mutant regions was not a simple function of deletion length and varied between different loci. In crosses between mutant strains which had non-overlapping deletions, wt mtDNA molecules were formed by recombination. The wt recombinants were present at high frequencies among the progeny of such crosses, but recombinants containing both deletions were not detected at all. The results indicate that mitochondrial genomes can be selectively transmitted to progeny and that two particular intergenic regions positively influence transmission. Within these regions other sequences in addition to ori/rep affect transmission.This paper is dedicated to colleagues J. Jana, D. Tasi, I. Bortner, and F. Zavrl     

Unisexual hybrids break through an evolutionary dead end by two-way backcrossing

Unisexual vertebrates (i.e., those produced through clonal or hemiclonal reproduction) are typically incapable of purging deleterious mutations, and, as a result, are considered short-lived in evolutionary terms. In hemiclonal reproduction (hybridogenesis), one parental genome is eliminated during oogenesis, producing haploid eggs containing the genome of a single parent. Hemiclonal hybrids are usually produced by backcrossing hemiclonal hybrids with males of the paternal species. When hemiclonal hybrids from a genus of greenlings (Hexagrammos) are crossed with males of the maternal species, the progeny are phenotypically similar to the maternal species and produce recombinant gametes by regular meiosis. The present study was conducted to determine if the hemiclonal genome is returned to the gene pool of the maternal species in the wild. Using a specific cytogenetic marker to discriminate between such progeny and the maternal species, we observed that Hexagrammos hybrids mated with maternal and paternal ancestors at the same frequency. This two-way backcrossing in which clonal genomes are returned to the gene pool where they can undergo recombination plays an important role in increasing the genetic variability of the hemiclonal genome and reducing the extinction risk. In this way, hybrid lineages may have survived longer than predicted through occasional recombinant generation.     

Absence of mitochondrial and chloroplast DNA recombinations in Brassica napus plants regenerated from protoplasts,protoplast fusions and anther culture

Summary Over 400 Brassica napus plants regenerated from individual protoplasts, from protoplast fusions and from anther culture were analysed for chloroplast and mitochondrial genome rearrangements by restriction fragment length polymorphisms. None were detected, attesting to the fidelity of the tissue culture procedures employed. In the majority of protoplast fusion products, the cytoplasmic organelles had completely sorted out at the callus stage but three regenerated plants possessed mixed parental populations of mitochondrial genomes and one regenerant contained mixed chloroplast genomes. In all four examples, the cytoplasmic genome sorted out in planta in favor of one parental type which was faithfully maternally transmitted to progeny.     

Effect of antibiotics on the transmission of mitochondrial factors in Saccharomyces cerevisiae

Summary Transmission of mitochondrial factors was studied in zygote clones, random diploids, and zygote cell lineages. An asymmetrical distribution of both parental and recombinant classes was consistently seen. Genetic analysis indicated that a nuclear factor was controlling this pattern. Treatment of young zygotes with either cycloheximide or thiolutin for 90 minutes destroyed the pattern and randomized the distribution of parental and recombinant classes. Inhibition of mitochondrial protein synthesis with antibacterial antibiotics had no detectable effect on the pattern of distribution of mitochondrial markers.     

Strict paternal transmission of mitochondrial DNA of Chlamydomonas species is explained by selection against maternal nucleoids

Summary.  The non-Mendelian inheritance of organelle DNA is common in most plants and animals. Here we examined inheritance mechanisms involved in the transfer of mitochondrial DNA. We successively backcrossed (to F₅) two interfertile strains of the unicellular isogamous haploid algae Chlamydomonas reinhardtii and Chlamydomonas smithii to match nuclear backgrounds and examine transmission patterns of mitochondrial DNA by PCR analysis of cob gene sequences. Mitochondrial DNA was strictly transmitted paternally. To investigate the behavior of parental mitochondrial DNA, we used F₅ progeny to form zygotes and isolated single zygotes. The results showed selective disappearance of maternal mitochondrial nucleoids occurred between 3 and 6 h after zygote formation. Received July 11, 2002; accepted September 28, 2002; published online June 13, 2003 RID="*" ID="*" Correspondence and reprints: Laboratory of Cell and Functional Biology, Faculty of Science, University of the Ryukyus, Nishihara, Okinawa 903-0213, Japan.