Genome Mapping In The Gastony Lab

With funding from the National Science Foundation, the Gastony lab is constructing a high-resolution genetic linkage map of the model homosporous fern species Ceratopteris richardii. Our mapping population consists of ~500 Doubled Haploid Lines (DHLs) generated from an initial cross between highly diverged geographic races of diploid inbred lines of Ceratopteris richardii (Hickok et al. 1995): phiN8 (derived from a Nicaraguan collection) and and pq45, a mutant of Hn-n (derived from a Cuban collection). We produced and analyzed Restriction Fragment Length Polymorphism (RFLP), Amplified Fragment Length Polymorphism (AFLP) and allozyme markers for construction of the map. Our RFLP probes are generated from sequenced Ceratopteris ESTs (provided by Dr. Jo Ann Banks of Purdue University). The advantages of using sequenced ESTs as probes are threefold: 1) the uniqueness of each probe is ensured by comparing their sequences, 2) the biochemical function of the probes can be inferred if the functions of their homologues are known in other organisms, and 3) the distribution of homologous loci in the genome can be compared between Ceratopteris and organisms with complete genome sequences such as Arabidopsis, and the evolution of chromosomal structure can be studied. We chose only ESTs with significant BLAST matches (E < 1e^-10) with Arabidopsis genomic sequences. We screened the selected ESTs by Southern hybridization and selected ones that are highly polymorphic and that generate clear banding patterns for actual RFLP genotyping.

Our mapping work is still in progress, but even at this unfinished stage our map permits a general description of the Ceratopteris genome, such as the distribution of duplicated loci in the genome and possible locations of sterility factors functioning as reproductive barriers between the parents. The map not only gives us a general picture of the Ceratopteris genome, but also provides a foundation for further studies of plant genome evolution, including the roles of gene and genome duplication. In addition to generating the DHLs through intragametophytic selfing of F₁-derived gametophytes, a set of ~200 F2s was generated through intergametophytic selfing of F1-derived gametophytes for use in assessing selection for/against heterozygotes at the sporophytic stage of the life cycle.

This project has been the primary focus of the doctoral dissertation research of Takuya Nakazato, with assistance from laboratory technicians Joanna Johnson and Valena Fiscus and from undergraduate researchers Elizabeth Eakin and Ellison Tipton. and with guidance by co-Principal Investigators Gerald J. Gastony and Loren H. Rieseberg.

1) Map description

To date we have mapped a moderate number of AFLP and RFLP markers covering 40 linkage groups that partially correspond to the 39 chromosomes in this species. From our preliminary genetic linkage map, linkage group (LG) 15 of Ceratopteris richardii is shown here as an example.

Each linkage group includes percent recombination and genetic distance (in cM) between two markers using the Kosambi mapping function, arbitrary marker number, and marker names. AFLP marker names begin with "*a", arbitrary AFLP marker number, and approximate fragment size in base pairs on AFLP gels. RFLP marker names consist of the restriction enzyme used: "*E" for EcoRI and "*H" for HindIII, GenBank gi number of the EST used as a probe, and arbitrary paralogue number.

Across the 40 linkage groups identified thus far, there seems to be clustering of markers in some regions of the genome, suggesting the presence of recombination hotspots and coldspots.

AFLP markers, mostly representing non-coding sequences, are known to cluster in low-recombination regions (Vuylsteke et al. 1999), but in our study, the marker clustering does not seem to depend on marker types, based on visual assessments. There are many large gaps within linkage groups (>30cM), and it is difficult to determine at present whether these represent false linkages or recombination hotspots. Subsequent analyses need to be performed with caution. We are continuing to map additional RFLP markers in order to increase the map's marker density and the evenness of marker distribution. It is hoped that this will help us to detect duplicated chromosomal regions and to test the hypothesis of paleopolyploidy in homosporous ferns (Gastony 1991)

2) Segregation distortion (SD) and reproductive barriers between the parents

A large proportion of markers (41.5% of RFLPs and 19.6% of AFLPs) showed significant SD (p<0.05). The magnitude of SD ranged up to 74.3% for RFLPs and 54.3% for AFLPs. A higher proportion of RFLP markers showed SD and and showed SD of a greater magnitude than did AFLP markers, consistent with the fact that AFLP markers represent mostly non-coding sequences, which are under weaker selection. Among the markers showing SD, regardless of marker types, the proportion of markers favoring the pq45 parental allele is not significantly different from the proportion favoring phiN8 parent (p >0.5), indicating that there is no overall preference of alleles from one parent over the other, unlike the findings of some other studies (Devicente and Tanksley 1993; Fishman et al. 2001), where alleles from one parent are preferentially transmitted. However, within linkage groups, linked markers typically show SD toward one parent or the other, suggesting that multiple adaptive loci are distributed throughout the genome, and markers around these loci have "hitchhiked" (Fig. 2_1 and 2_2, open symbols). While SD among homozygous individuals can be caused by selection at the gametophytic and/or sporophytic phase of the life cycle, an excess or deficiency of heterozygotes relative to Hardy-Weinberg expectation among heterozygous F2s can be caused only by selection on the sporophytic phase (Harushima et al. 2001; Harushima et al. 2002). Only seven out of 40 co-dominant RFLP markers (17.5%) scored among heterozygous F2s showed significant (p<0.05) deficiency of heterozygotes, relative to Hardy-Weinberg expectations (Fig. 2_1 and 2_2, closed symbols). This suggests that selection at the sporophytic phase is not frequent, although at some loci it may be strong, and that overdominance is not important for the fitness of hybrids. More co-dominant markers among heterozygous F2s need to be mapped to determine the contribution of selection at the sporophytic phase.

Our crossing design using homozygous F2s (DHLs) proved to be much more sensitive for detecting SD than is the typical use of heterozygous F2s in mapping studies. Seventy six markers showed SD in homozygous F2s, heterozygous F2s, or both. Of those 76 markers, 42 (55.3%) showed significant SD only in homozygous F2s, whereas only 22 (28.9%) showed SD in both homozygous and heterozygous F2s. Therefore, by using DHLs, we increased the chance of detecting SD by nearly twofold. This reflects the fact that homozygous F2s are directly exposed to selection, whereas heterozygous F2s can buffer selection. Interestingly, 11 markers (14.5%) showed SD only in heterozygous F2s for unknown reasons.

3) Distribution of paralogous gene copies in the genome

Fig. 3 shows the distribution of paralogous gene pairs in the 40 linkage groups. Paralogues seem to be randomly distributed in the genome, and no "duplication hotspots" are apparent. To date we have detected 24 pairs of linkage groups that share two or more paralogous pairs (shaded blocks in Fig. 3), potentially representing homoeologous chromosomes or chromosomal regions. Figs. 4_1, 4_2, amd 4_3 show positional relationships of paralogous gene pairs for each of these 24 pairs of linkage groups, superimposed on the SD data (Fig. 2). Most of these paralogous gene pairs seem to be results of independent duplication events because a) shared paralogues in each chromosome pairs tend to be far apart (>10cM, Fig. 4), and b) there is usually no one-to-one correspondence with regard to which linkage groups are indicated as potentially homoeologous (Fig. 3). There is no question that gene duplication is abundant in the genome, but the modes of gene duplication are not clear from the current maps. It may be that gene-based duplications (e.g., tandem repeats, retrotransposition) are common. Alternatively, large-scale duplications (e.g., segmental duplication, polyploidization) may be common but evidence for them may be masked by extensive rearrangements. It is probably safe to conclude that there has been no large-scale duplication in recent history, and mapping more markers may not help to identify homoeologous chromosomal regions. Furthermore, the distribution of chromosome numbers across the fern rbcL phylogeny suggests that polyploid speciation is probably not important in fern phylogeny in general, although polyploidization readily occurs within lineages.

Although we have not detected large homoeologous chromosomal blocks, some paralogues shared by pairs of linkage groups are tightly linked (e.g., lg4 & lg9 in Fig. 4_1 and lg5 & lg34 in Fig. 4_2) and are likely to have structural synteny. The structure of duplicated chromosomal regions can be studied by focusing on the microsynteny of these regions in the future.

4) Divergence of paralogous and orthologous gene pairs

Despite the co-dominant nature of RFLP markers, a surprisingly large proportion of markers (72.0%) were dominant. Probably this is mostly because of divergence/deletion of alleles in the parents, although some instances may be caused by factors such as missing alternative alleles attributable to overlapping RFLP bands. As expected, a significant proportion of RFLP dominant markers (65.1%, p<0.001) came from parent pq45, from which the RFLP probes were made, whereas AFLP markers came equally from both parents (p>0.1), indicating that it is more difficult to detect alleles from parent phiN8 by RFLPs because of sequence divergence/deletion. However, for 34.9% of the dominant RFLP markers, probes are more similar to phiN8 alleles than to pq45 alleles. These are probably cases in which paralogues have diverged in pq45 but not in phiN8. Therefore, independent evolution of paralogous duplicated genes after speciation may be common. Sixty-six RFLP probes were scored into three types: 1) 11 probes consisting of only dominant markers, 2) 42 probes with one co-dominant and 0-11 dominant markers, and 3) 13 probes with two co-dominant markers and 0-3 dominant markers. Type 1 probes may represent genes that have diverged because their function is no longer shared by the parents of the cross. Type 2 probes may represent genes in which the function of one copy is still shared by the parents but in which homology for other gene copies has been lost. Type 3 probes may represent genes that may have been duplicated relatively recently and in which the parents still share homology at multiple gene copies. It may be interesting to investigate whether Type 3 probes show subfunctionalization and perhaps are involved in DDC-type (Duplication, Degeneration, Complementation) hybrid breakdown (Lynch and Force 2000). We can study the evolution of paralogues and orthologues by analyzing the RFLP patterns in more detail.

Although, we have not detected convincing homoeologous chromosomal regions so far (see section 3 above), homoeologous chromosomal regions may be more apparent if we focus on young paralogous gene pairs generated by Type 3 probes.

5) Spore viability

Fig. 5 shows the distribution of germination rates among homozygous F2s (generated through intragametophytic selfing) and heterozygous F2s (generated through intergametophytic selfing). The average germination rate for heterozygotes was 35.2%, and a substantial number of individuals had lower than a 10% germination rate. The broad continuous distribution of germination rates suggests that multiple sterility factors are segregating in the mapping population. To estimate the magnitude of environmental and other non-genetic effects on germination rate, we also scored germination rates of parents (arrows in Fig. 5), and of homozygous F2s. The fact that completely homozygous parents did not have 100% germination rates indicates that 10-20%, but not all, of the variation in germination rates of heterozygotes is explained by environmental factors, such as mortality due to prolonged storage and fungal contamination during germination, and possibly by spontaneous mutations. If the haplotypes of spores are the sole determinant of germination rates, homozygotes should have 100% germination rates. The fact that homozygotes did not have 100% germination rates and had a wide range of germination rates indicates that environmental as well as maternal effects, such as poor regulation of spore production, may contribute to the variation. Furthermore, homoeologous or other abnormal pairing (Klekowski 1973; Hickok 1978) and transposons may affect the variation. It should be verified that individuals in each DHL have identical genotypes. Comparison of spore viability among parents, F1, and heterozygous F2s allows us to investigate the possible causes of hybrid sterility. Two major hypotheses for the causes of hybrid sterility are the Bateson-Dobzansky-Muller (BDM) model of epistasis and chromosomal rearrangements. The BDM model of epistasis causes hybrid sterility by incompatible interaction between genes that have diverged between the parents, and have been brought together by recombination. Differences in chromosomal structure (e.g., inversions, translocations) between the parents can cause hybrid sterility because portions of F1 gametes lack chromosomal segments because of recombination within the rearranged regions. Theories predict that hybrid sterility caused by chromosomal rearrangements is most severe among F1 gametes and that fitness improves in subsequent generations. This is because heterozygosity in chromosomal structure is highest in the F1 generation, assuming random recombination within rearranged chromosomal regions and lower fitness of recombinant gametes. In contrast, hybrid sterility by the BDM model should persist for several generations because it is affected only by interaction among different genes, not by heterozygosity. Our data show a much reduced spore germination rate from the parents (ca. 90%, Fig. 5), to the F1 gametes (ca. 30%, Hickok et al. 1995), consistent with both hypotheses above. However, the germination rate did not improve substantially on average from the F1 (ca. 30%) to F2 (ca. 35%) generations. Based on this observation alone, chromosomal rearrangements are not the major causes of hybrid sterility, unlike the case in sunflower (Rieseberg et al. 1995a; Rieseberg et al. 1995b). Further crossing experiments will help to improve our understanding of the causes of hybrid sterility. For example, reciprocal crosses of the parents should reveal whether cyto-nuclear interaction is partly responsible for hybrid sterility.

References

Devicente, M. C., and S. D. Tanksley. 1993. QTL analysis of transgressive segregation in an interspecific tomato cross. Genetics 134:585-596.

Fishman, L., A. J. Kelly, E. Morgan, and J. H. Willis. 2001. A genetic map in the Mimulus guttatus species complex reveals transmission ratio distortion due to heterospecific interactions. Genetics 159:1701-1716.

Gastony, G. J. 1991. Gene silencing in a polyploid homosporous fern: Paleoployploidy revisited. PNAS 88:1602-1605.

Harushima, Y., M. Nakagahra, M. Yano, T. Sasaki, and N. Kurata. 2001. A genome-wide survey of reproductive barriers in an intraspecific hybrid. Genetics 159:883-892.

Harushima, Y., M. Nakagahra, M. Yano, T. Sasaki, and N. Kurata. 2002. Diverse variation of reproductive barriers in three intraspecific rice crosses. Genetics 160:313-322.

Hickok, L. G. 1978. Homoeologous chromosome pairing and restricted segregation in the fern Ceratopteris. American Journal of Botany 65:516-521.

Hickok, L. G., T. R. Warne, and R. S. Fribourg. 1995. The biology of the fern Ceratopteris and its use as a model system. International Journal of Plant Sciences 156:332-345.

Klekowski, E. J., Jr. 1973. Sexual and subsexual systems in homosporous pteridophytes: A new hypothesis. American Journal of Botany 60:535-544.

Lynch, M., and A. G. Force. 2000. The origin of interspecific genomic incompatibility via gene duplication. American Naturalist 156:590-605.

Rieseberg, L. H., C. R. Linder, and G. J. Seiler. 1995a. Chromosomal and genic barriers to introgression in Helianthus. Genetics 141:1163-1171.

Rieseberg, L. H., C. Vanfossen, and A. M. Desrochers. 1995b. Hybrid speciation accompanied by genomic reorganization in wild sunflowers. Nature 375:313-316.

Vuylsteke, M., R. Mank, R. Antonise, E. Bastiaans, M. L. Senior, C. W. Stuber, A. E. Melchinger et al. 1999. Two high-density AFLP (R) linkage maps of Zea mays L.: analysis of distribution of AFLP markers. Theoretical and Applied Genetics 99:921-935.