- Split View
-
Views
-
Cite
Cite
Katarina Andreasen, Bruce G. Baldwin, Unequal Evolutionary Rates Between Annual and Perennial Lineages of Checker Mallows (Sidalcea, Malvaceae): Evidence from 18S–26S rDNA Internal and External Transcribed Spacers, Molecular Biology and Evolution, Volume 18, Issue 6, June 2001, Pages 936–944, https://doi.org/10.1093/oxfordjournals.molbev.a003894
- Share Icon Share
Abstract
Heterogeneous DNA substitution rates were found in the 18S–26S nuclear ribosomal DNA internal transcribed spacer (ITS) and external transcribed spacer (ETS) regions of Sidalcea (Malvaceae), a putatively young genus of annuals and perennials. The majority of comparisons revealed that the annual species had significantly higher molecular evolutionary rates than the perennials, whereas rates were consistently homogenous between obligate annual species. These findings led us to conclude that generation time or possibly another biological factor distinguishing annuals and perennials has influenced rates of molecular evolution in Sidalcea. The congruence of relative-rate test results across both spacer regions reinforced the association between life history and rate of rDNA evolution across lineages of checker mallows. Evolutionary rate variation within perennials mainly involved three basally divergent lineages. The faster rate in one lineage, Sidalcea stipularis, compared with other perennials may be the result of genetic drift in the only known, small, population. The other two basally divergent lineages had slower evolutionary rates compared with the remaining perennials; possible explanations for these differences include rate-reducing effects of a suffrutescent (rather than herbaceous) habit and seed dormancy.
Introduction
Heterogeneous evolutionary rates in the genomes of both plant and animal groups have been demonstrated in several studies (e.g., Wu and Li 1985 ; Britten 1986 ; Bousquet et al. 1992 ; Martin, Naylor, and Palumbi 1992 ; Nickrent and Starr 1994 ; Eyre-Walker and Gaut 1997 ; Laroche and Bousquet 1999 ), although the causes of rate heterogeneity are still unclear. Hypothetical explanations include generation time and speciation rate effects, as well as fidelity of DNA replication, frequency of DNA damage, efficiency of DNA repair, and changes in population size (Britten 1986 ; Martin and Palumbi 1993 ; Eyre-Walker and Gaut 1997 ; Gaut et al. 1997 ; see also Sanderson [1998] for a review of estimation of evolutionary rates).
Comparisons of molecular evolutionary rates among plant lineages have been made mostly for relatively conserved genes. Substitution rates have been compared between different plant taxa using a single gene, e.g., the chloroplast gene rbcL (Bousquet et al. 1992 ; Gaut et al. 1992 ) and nuclear 18S rDNA (Nickrent and Starr 1994 ). Other studies have also tested for molecular evolutionary rate correlations in comparisons between different plant lineages but have examined multiple genes (see Muse [2000] for a review). Genes from all three plant genomes have been compared, e.g., the chloroplast genes rbcL and ndhF (Gaut et al. 1997 ); rbcL and the nuclear gene Adh (Gaut et al. 1996 ); Adh, rbcL, and the mitochondrial gene atpA (Eyre-Walker and Gaut 1997 ); and nuclear 18S rDNA, mitochondrial 19S rDNA, and the three plastid genes rbcL, rps2, and 16S rDNA (Nickrent et al. 1998 ).
Comparisons of more than one DNA region can yield significant insights into processes affecting molecular evolutionary rates and may allow discrimination among different possible explanations for rate heterogeneity (Eyre-Walker and Gaut 1997 ; Gaut et al. 1997 ; Nickrent et al. 1998 ; Muse 2000 ). To identify evolutionary forces that are expected to affect substitution rates at the organismal level, e.g., the impact of generation time or speciation rate, investigation of functionally independent DNA regions from more than one genome is ideal. Organismal-level factors should cause correlated patterns of rate heterogeneity between DNA regions from different genomes within a lineage. Comparisons between DNA regions from the same organelle can be useful to distinguish between gene-specific factors (e.g., selection) and genome-specific factors (e.g., DNA replication rate).
A pattern found in some of the comparative studies cited above was that annual plants often showed an elevated rate of molecular evolution compared with perennials. These differences in DNA substitution rates between plants of contrasting life histories have been explained in various ways, e.g., by generation time or speciation rate effects or by varying efficiency of DNA replication or repair in combination with selection against heterozygosity and differences in population sizes between annuals and perennials (Bousquet et al. 1992 ; Eyre-Walker and Gaut 1997 ; Gaut et al. 1997 ; Laroche and Bousquet 1999 ), but none of the hypotheses explain the empirical data completely. In contrast to animals, which have determinate germ cell lines, the correlation between generation time and evolutionary rate is not straightforward in plants. The frequency of DNA replication should depend on both generation time and the number of cell divisions per generation (Eyre-Walker and Gaut 1997 ).
Most studies to date have included comparisons between relatively distantly related annual and perennial taxa (e.g., Bousquet et al. 1992 ; Laroche et al. 1997 ; Laroche and Bousquet 1999 ). In contrast, the plants compared herein (checker mallows) belong to a putatively young western North American genus, Sidalcea (Malvaceae), comprising ca. 25 closely related species of annuals and perennials. The existence of both annual and perennial lineages in Sidalcea enables us to make comparisons of evolutionary rates between plants of different life histories that share highly similar genetic backgrounds. We presume that closely related taxa are less likely to differ in intrinsic metabolic properties such as DNA polymerase fidelity, frequency of DNA damage, and efficiency of DNA repair than are more distantly related groups (Thorne, Kishino, and Painter 1998 ). Another advantage to using Sidalcea for comparisons of evolutionary rates between annuals and perennials is that most of the annuals are more closely related to perennial species than to other annual taxa. Thus, changes in habit must have occurred more than once, enabling us to make more independent comparisons of evolutionary rates between groups of similar or different habits than if the annuals and perennials each had constituted only a single monophyletic group.
Recently, unequal evolutionary rates in plants have been detected using noncoding DNA regions, e.g., the trnL intron in chloroplast DNA (Gielly and Taberlet 1996 ), the internal transcribed spacer (ITS) in 18S–26S nuclear ribosomal DNA (Wendel, Schnabel, and Seelanan 1995 ; Baldwin and Sanderson 1998 ; Aïnouche and Bayer 1999 ), and the mitochondrial intron rps3 (Laroche and Bousquet 1999 ). However, most published studies on molecular evolutionary rates in lineages have involved evolutionarily conservative coding regions in the chloroplast genome; the dynamics of evolutionary rates in the mitochondria and the nucleus are relatively poorly known (Muse 2000 ). To evaluate evolutionary rates in Sidalcea, we studied sequences from two 18S–26S nuclear ribosomal DNA (rDNA) spacer regions: the ITS region and the external transcribed spacer (ETS) region. The ITS region has been widely used in phylogenetic studies at lower taxonomic levels (see Baldwin et al. 1995 ), but the phylogenetic utility of the ETS has been explored in relatively few plant groups (Baldwin and Markos 1998 ; Bena et al. 1998 ; Linder et al. 2000 ; Markos and Baldwin 2001 ). In this paper, we investigate the rates of molecular evolution in the closely related annual and perennial lineages of Sidalcea and evaluate whether differences in rates are correlated to plant habit and whether relative rate patterns are similar in the two noncoding rDNA regions, ITS and ETS.
Materials and Methods
The plants sampled in this study were from 28 annual and perennial populations of Sidalcea and three outgroup taxa in tribe Malveae (see table 1 ). The leaf material used for DNA extraction was either fresh, dried in silica gel, or from herbarium specimens. Total DNAs were isolated using the CTAB procedure (Saghai-Maroof et al. 1984 ; Doyle and Doyle 1987 ) with phenol extraction and ethanol precipitation. The ITS region was amplified with ITS leu.1 and ITS4 and sequenced using standard procedures (see Andreasen, Baldwin, and Bremer [1999] and modifications below). To obtain ETS sequences, the intergenic spacer (IGS) was amplified using long-distance PCR with KlenTaq LA (Sigma) and universal primers 18S-IGS and 26S-IGS (fig. 1 ; Baldwin and Markos 1998 ). Primer 18S-E (Baldwin and Markos 1998 ) was used to sequence upstream into the ETS region of the IGS product. “Primer walking” was performed by constructing a reverse primer, Sid-R (5′-CCTGCAAATACACACAACCAAATT-3′, 423 bp upstream of 18S). Finally, a forward primer, Sid-F (5′-CGTGCCTATCGGTTGTGGTGT-3′, 550 bp upstream of 18S), was constructed and used with 18S-E to amplify a 3′ ETS fragment of ca. 550 bp. Sid-F and 18S-E also were used for ETS sequencing reactions. Primers (White et al. 1990 ) ITS4, ITS5-A (a modification of ITS5 to match angiosperm sequences more closely than fungal sequences: 5′-GGAAGgAgAAGTCGTAACAAGG-3′, where lowercase letters indicate changed nucleotides [Downie and Katz-Downie 1996]), and, for some DNAs, ITS2 and/or ITS3 were used to obtain sequences of ITS-1 and ITS-2 separately (fig. 1 ). The ITS and ETS amplification products were cycle-sequenced with BigDye (Perkin Elmer) or Thermo Sequenase (Amersham Life Sciences) terminator cycle sequencing kits and resolved on polyacrylamide gels using an ABI Prism 377 Automated Sequencer (Perkin-Elmer).
Resolved sequences were aligned using the Clustal algorithm or Gotoh-Myers' comparative alignment option as implemented in the software Sequence Navigator (Perkin Elmer) and adjusted by eye. Gaps were coded as missing data and recoded as binary characters. PAUP*, version 4.0b2a (Swofford 1998 ), was used for the parsimony analyses and likelihood calculations. Cladistic analyses of the combined and separate ITS and ETS aligned-sequence matrices were conducted using heuristic searches with 100 random-taxon-addition replicates, tree bisection-reconnection (TBR) branch swapping, and the MULPARS option in effect. To estimate support for clades, jackknife (Farris et al. 1996 ) and Bremer/decay support (Bremer 1994 ) analyses were carried out using PAUP* in combination with the program Autodecay (Eriksson 1998 ) for determining Bremer support.
To test for heterogeneous evolutionary rates, a treewide likelihood ratio (LR) rate test was carried out using the Hasegawa-Kishino-Yano two-parameter substitution model with empirically determined transition/transversion ratio and base frequencies, estimated among-sites variation, gamma distribution for the variable sites (with four rate categories and shape parameter estimated), and the Rogers-Swofford method for initializing starting branch lengths. The outgroups were pruned off the trees before the tests were carried out.
Relative-rate tests for pairwise comparisons were conducted for the separate ITS and ETS matrices and for the combined data with the program K2WuLi (Jermiin 1996 ), which executes the test of Wu and Li (1985) as implemented by Muse and Weir (1992) using the Kimura two-parameter model (1980) for correction of multiple substitutions. The results of pairwise comparisons for the separate ITS and ETS matrices were used in the Mantel test to determine whether the ITS and ETS relative-rate matrices exhibited the same pattern of evolutionary rates. For this purpose, the program Mantel, version 2 (Liedloff 1999 ), was employed with the number of random permutations set to 5,000.
To address the problem of nonindependence of multiple relative-rate tests, we conducted sister group comparisons using each of the four annual lineages and a perennial sister group chosen at random. The individual P values were then combined using a G-test (see, e.g., Sokal and Rohlf 1995 , pp. 685–793). Another approach that we chose in order to circumvent the problem of nonindependence was a relative-rate test with phylogenetic weighting (Robinson et al. 1998 ). The program RRTree (available at http://pbil.univ-lyon1.fr/software/rrtree.html) was used to implement this test with Kimura 1980 two-parameter distances and 10 defined lineages (4 annual and 6 perennial).
Finally, we applied Wilcoxon's signed-ranks test (see, e.g., Sokal and Rohlf 1995 , pp. 440–444; the test is available at http://www.fon.hum.uva.nl/Service/Statistics.html) for assessing rate heterogeneity between annuals and perennials using the results of pairwise comparisons from the relative-rate tests. Comparisons included all five annual species versus a subset of perennial taxa (n = 30).
Results
The aligned sequences from the ITS region (693 bp) and the 3′ end of the ETS region (545 bp) constitute a matrix of 1,238 bp (the EMBL accession numbers for the aligned sequences are ALIGN_000016 for the ITS region and ALIGN_000017 for the 3′ ETS region; accession numbers for the sequences are given in table 1 ). In total, the sequences provided 201 potentially informative characters for parsimony-based phylogenetic analysis; 104 from the ITS region and 97 from the ETS region. In addition, 16 potentially informative gap characters were found, 11 in the ITS region and 5 in the ETS.
Separate phylogenetic analyses resulted in 162 most-parsimonious trees for the ITS matrix (not shown; consistency index [CI] = 0.61 excluding noninformative characters; retention index [RI] = 0.72) and 144 most-parsimonious trees for the ETS matrix (not shown; CI = 0.65 excluding noninformative characters; RI = 0.78). Simultaneous phylogenetic analysis of the two combined data sets yielded six most-parsimonious trees (fig. 2 ; CI = 0.62 excluding noninformative characters; RI = 0.74). The phylogram in figure 2 shows that the annual species Sidalcea diploscypha, S. keckii, S. calycosa, S. hartwegii, and S. hirsuta have longer branches than all of the perennial species except S. stipularis.
The treewide LR test for rate constancy demonstrated that rates of nuclear DNA evolution have been significantly heterogeneous in Sidalcea (P < 0.001 for both the separate ITS and ETS data sets and for the combined matrix). Results of the Mantel test, which compares two dissimilarity matrices to test for association between entries in the matrices, led us to reject the null hypothesis (H0 = no association between elements in the matrices) for the ITS and ETS matrices (P < 0.001). Rejection of H0 leads us to conclude that patterns of relative rates found in the separate ITS and ETS matrices are correlated.
The relative-rate tests of the combined sequence matrix showed that rates of rDNA evolution mostly have not been significantly heterogeneous among annuals or among perennials (see table 2 ). A lack of rate heterogeneity among plants of similar habits was found when the relative-rate test was performed for the ITS region and for the ETS but was more pronounced for the combined data. For the combined matrix, rate constancy could be rejected for only 4% (7/190) of the comparisons between perennials, excluding the most basally divergent lineages (if they were included, 19% [47/253] of the comparisons were significantly heterogeneous), and for none of the comparisons between obligate annuals (i.e., excluding S. hirsuta; see Discussion). In contrast, rate constancy could be rejected in 54% (74/138) of the comparisons between annuals and perennials. In these pairwise comparisons, evolutionary rates in the annuals were up to seven times as fast as those in the perennials (for the comparison between the annual S. calycosa with the perennial S. oregana).
The major exceptions to homogenous rates in comparisons between perennial taxa were for the three basally divergent taxa S. malachroides, S. hickmanii, and S. stipularis. Sidalcea malachroides and S. hickmanii appeared to have evolved significantly more slowly than most of the other perennials; S. stipularis exhibited a faster evolutionary rate than the other perennials. Among the annuals, comparisons of S. hirsuta with either S. calycosa or S. hartwegii were significantly heterogeneous (P < 0.001 and P < 0.05, respectively) and the evolutionary rate was slower for S. hirsuta in both cases. Results from Wilcoxon's signed-ranks test, which was used to test for significant evolutionary rate differences between annuals versus perennials, suggested that the rDNA spacers of the annuals had evolved at a significantly higher rate than those of the perennials (n = 30, P < 0.0001). Both the results from the relative-rate test with phylogenetic weighting (see table 3 ) and the results from the four annual-perennial sister group comparisons (a combined P value of <0.001) also indicated that molecular evolutionary rates had been significantly higher in the annuals than in the perennials, whereas rates had been consistently homogenous between obligate annual species.
Discussion
Different explanations have been proposed for unequal evolutionary rates between plants with different life histories. Some researchers (Gaut et al. 1996, 1997 ; Eyre-Walker and Gaut 1997 ; Aïnouche and Bayer 1999 ) have suggested a correlation between evolutionary rate and generation time, although not all of the data support that conclusion. Bousquet et al. (1992) rejected the hypothesis of a generation time effect altogether; herbaceous and woody perennials had similarly slow rates in their study. To account for evidence for more rapid molecular evolution in annuals than in perennials, Bousquet et al. (1992) invoked differences in speciation processes and a faster speciation rate due to putatively smaller population sizes in annuals than in perennials. Gaut et al. (1996) did not corroborate this conclusion in their study of palms and grasses; molecular evolutionary rates were faster and population sizes were larger in grasses than in palms. Gaut et al. (1996) also rejected the hypothesis that DNA polymerase fidelity influences mutation rates in palms and grasses based on their finding of correlated rate differences between nuclear and chloroplast genes (the two genomes use different polymerases).
Most of the data for Sidalcea support the hypothesis of a generation time effect on molecular evolutionary rates. Nuclear rDNA spacers generally evolve significantly faster in annual members of Sidalcea than in the perennials, and the rates within annuals are essentially homogenous (see tables 2 and 3 ). Because the taxa in this study are closely related, they are presumably less likely than more distantly related taxa to exhibit differences in fundamental biological processes, e.g., metabolic rates, a factor that has been proposed to affect molecular evolutionary rates. In two comparisons between annuals involving S. hirsuta, the evolutionary rates were significantly heterogeneous, with S. hirsuta having the slower rate in both cases. One possible explanation for the slow evolutionary rate in S. hirsuta is that the species is sometimes of perennial habit and may be a facultative annual (unpublished data). If generation time is affecting evolutionary rate, then facultative annuals may be expected to have slower substitution rates than obligate annuals.
The comparisons between perennials with significantly heterogeneous molecular evolutionary rates mainly involved the three most basally divergent lineages, i.e., S. stipularis, S. malachroides, and S. hickmanii (table 2 ). The first species had a faster molecular evolutionary rate in comparisons with other perennials, while the other two taxa had slower evolutionary rates than the other perennials. The faster evolutionary rate in S. stipularis may be caused by genetic drift in the only known, small, population (ca. 300 plants). Sidalcea malachroides, one of the two taxa with the slowest evolutionary rates, is the only subshrubby species in the genus and therefore may be expected to have a longer generation time and slower evolutionary rate than the herbaceous perennials. The other slowly evolving perennial, S. hickmanii, appears to respond positively to forest fires, with a large increase in the number of plants occurring shortly after a fire (unpublished data). This high recruitment suggests that the seed bank is probably substantial, which could extend the effective generation time and reduce the evolutionary rate relative to other perennials (assuming seed banks are smaller in the other taxa; data are lacking). Other factors that may complicate interpretation of patterns of heterogeneous evolutionary rates include probable inconstancy of generation time within lineages and uncertainty about when habit shifts occurred in lineages.
Correlated rate heterogeneity in two DNA regions, as found here using the Mantel test for the ITS region and the partial ETS sequences, is concordant with the hypothesis of a generation time effect on molecular evolutionary rate, although other explanations for the correlated rates are possible. The ETS and ITS regions are part of the same transcriptional unit and may have interdependent roles in the processing of mature rRNAs (e.g., Good, Intine, and Nazar 1997 ), so some degree of coevolution between the two spacer regions is likely. Whether selection could result in rates of molecular evolution in the two different spacer regions that are as strongly correlated as those observed here is open to question. Correlated patterns of rate heterogeneity from functionally distinct loci on different chromosomes or (preferably) in different genomes would provide stronger support for the effect of generation time on rates of molecular evolution. At a minimum, our findings provide an instructive example from a group of closely related plant species of faster rDNA evolution in annuals than in perennials and represent unequivocal evidence that life history differences in plants can be associated with significant differences in rates of molecular evolution.
Brandon S. Gaut, Reviewing Editor
Abbreviations: ETS, external transcribed spacer of 18S–26S nuclear ribosomal DNA; ITS, internal transcribed spacer of 18S–26S nuclear ribosomal DNA.
Keywords: Sidalcea internal/external transcribed spacers 18S–26S nuclear ribosomal DNA nucleotide substitution rates generation time
Address for correspondence and reprints: Katarina Andreasen, Laboratory of Molecular Systematics, Swedish Museum of Natural History, Box 50007, SE-104 05 Stockholm, Sweden. katarina.andreasen@ebc.uu.se.
We thank H. Forbes (and the University of California–Berkeley Botanical Garden), V. Oswald, R. Halse, E. Gruber McEvoy, and K. Uptain for plant material; B. Guggolz, B. Söderström, and P. Hubbard (LADWP) for help in the field; D. Hickson, E. Burres, R. Bittman (all of the California Department of Fish and Game), C. Brown (Cal Trans), J. Stebbins, S. Bainbridge, B. Ertter, and A. Bradley (USDA Forest Service) for help with permits and/or locality information and access; S. Hill for assistance with voucher determinations; R. Moe for computer help; M. Sanderson for statistical advice; and M. Sanderson and two anonymous reviewers for helpful reviews of the manuscript. This study was supported by a postdoctoral grant to K.A. from the Swedish Natural Science Research Council/Swedish Foundation for International Cooperation in Research and Higher Education and grants to K.A. from the Faculty of Science and Technology at Uppsala University, The Royal Swedish Academy of Sciences, The H. Ax:son Johnsson Foundation, and The Lawrence R. Heckard Endowment Fund of the Jepson Herbarium, University of California–Berkeley.
literature cited
Aïnouche, A.-K., and R. J. Bayer.
Andreasen, K., B. G. Baldwin, and B. Bremer.
Baldwin, B. G., and S. Markos.
Baldwin, B. G., and M. J. Sanderson.
Baldwin, B. G., M. J. Sanderson, J. M. Porter, M. F. Wojciechowski, C. S. Campbell, and M. J. Donoghue.
Bena, G., M.-F. Jubier, I. Olivieri, and B. Lejeune.
Bousquet, J., S. H. Strauss, A. H. Doerksen, and R. A. Price.
Britten, R. J..
Downie, S., and D. Katz-Downie.
Doyle, J. J., and J. L. Doyle.
Eriksson, T..
Eyre-Walker, A., and B. S. Gaut.
Farris, J. S., V. A. Albert, M. Källersjö, D. Lipscomb, and A. G. Kluge.
Gaut, B. S., L. G. Clark, J. F. Wendel, and S. V. Muse.
Gaut, B. S., B. R. Morton, B. C. McCaig, and M. T. Clegg.
Gaut, B. S., S. Muse, W. D. Clark, and M. T. Clegg.
Gielly, L., and P. Taberlet.
Good, L., R. V. A. Intine, and R. N. Nazar.
Hill, S. R..
Kimura, M..
Laroche, J., and J. Bousquet.
Laroche, J., P. Li, L. Maggia, and J. Bousquet.
Liedloff, A..
Linder, C. R., L. R. Goertzen, B. V. Heuvel, J. Francisco-Ortega, and R. K. Jansen.
Markos, S., and B. G. Baldwin.
Martin, A., G. Naylor, and S. Palumbi.
Martin, A., and S. Palumbi.
Muse, S..
Nickrent, D. L., and E. M. Starr.
Nickrent, D. L., R. J. Duff, A. E. Colwell, A. D. Wolfe, N. D. Young, K. E. Steiner, and C. W. dePamphilis.
Robinson, M., M. Gouy, C. Gautier, and D. Mouchiroud.
Saghai-Maroof, M. A., K. M. Soliman, R. A. Jorgensen, and R. W. Allard.
Sanderson, M. J..
Sokal, R. R., and F. J. Rohlf.
Swofford, D. L..
Thorne, J. L., H. Kishino, and I. S. Painter.
Wendel, J. F., A. Schnabel, and T. Seelanan.
White, T. J., T. Bruns, S. Lee, and J. Taylor.