| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
National Evolutionary Synthesis Center, 2024 West Main Street Suite A200, Durham, North Carolina 27705 USA
Received for publication 20 August 2007. Accepted for publication 9 January 2008.
ABSTRACT
Desiccation tolerance is a complex trait that is broadly but infrequently present throughout the evolutionary tree of life. Desiccation tolerance has played a significant role in land plant evolution, in both the vegetative and reproductive life history stages. In the land plants, the late embryogenesis abundant (LEA) gene families are involved in both abiotic stress tolerance and the development of reproductive propagules. They are also a major component of vegetative desiccation tolerance. Phylogenies were estimated for four families of LEA genes from Arabidopsis, Physcomitrella, and the desiccation tolerant plants Tortula ruralis, Craterostigma plantagineum, and Xerophyta humilis. Microarray expression data from Arabidopsis and a subset of the Physcomitrella LEAs were used to estimate ancestral expression patterns in the LEA families and to evaluate alternative hypotheses for the origins of vegetative desiccation tolerance in the flowering plants. The results contradict the idea that vegetative desiccation tolerance in the resurrection angiosperms Craterostigma and Xerophyta arose through the co-option of genes exclusively related to stress tolerance, and support the propagule-derived origin of vegetative desiccation tolerance in the resurrection plants.
Key Words: ancestral state • angiosperms • Bayesian • Bryophytes • desiccation tolerance • late embryogenesis abundant protein
Desiccation tolerance (DT), the ability to survive and recover from the near complete loss of water in the protoplasm, is a complex physiological trait that is thought to have played a critical role in the evolution of land plants. In a phylogenetic context, it is evident that DT has been both retained and recruited repeatedly in different developmental stages and structures throughout the history of the land plants (Fig. 1), an observation that underscores its ecological and physiological significance at various stages in the plant life cycle.
|
). The relatively frequent occurrence of vegetative DT in the modern day representatives of these lineages has supported the hypothesis that vegetative DT was the common ancestral condition in the early land plants. However, many modern bryophyte gametophytes are desiccation sensitive, and those DT bryophytes (particularly mosses) that have been extensively studied are known to employ dramatically different mechanisms for achieving tolerance (Oliver et al., 1998, 2000b; Rascio and La Rocca, 2005; Wood, 2007). Therefore, the possibility remains that this complex trait may have arisen independently multiple times in the bryophytes, possibly through the co-option of genes responsible for the DT of spores.
As land plants evolved more elaborate architectures and began to internalize their water relationships with structures such as roots and vascular tissue, DT in the vegetative tissues was lost, probably due to a trade-off between growth rates and the high energetic costs of maintaining vegetative DT (Oliver et al., 2000a; Alpert, 2006). However, in the tracheophytes, some of the genes involved in the original vegetative DT and spore DT of the bryophytes were co-opted for functions related to stress tolerance and the maturation of reproductive propagules such as seeds and pollen (Oliver et al., 2000a). The vast majority of modern seed plants produce seeds with DT embryos (i.e., "orthodox" seeds), but desiccation sensitive ("recalcitrant") seeds occur in lineages throughout the angiosperm clade (Tweddle et al., 2003). The widespread occurrence of recalcitrant seeds throughout the extant seed plants (gymnosperms and angiosperms) has led to considerable controversy as to whether the earliest seeds were DT (Tweddle et al., 2003) or desiccation sensitive (Pammenter and Berjak, 2000). While several early-diverging angiosperms have recalcitrant seeds, many other early angiosperm lineages have orthodox seeds. In the gymnosperms, almost all the coniferous taxa that have been studied have orthodox seeds, while there are conflicting reports on the DT status of seeds in Ginkgo and the cycads (Forsyth and Vanstaden, 1983; Attree and Fowke, 1993; Pammenter and Berjak, 2000; Tweddle et al., 2003). Therefore, the possibility remains that DT (orthodox) seeds evolved more than once early in the history of the seed plant clade. If the earliest seeds were desiccation sensitive, then seed DT could have arisen through the co-option of genes and pathways related stress tolerance or, potentially, the DT of the male microgametophyte (pollen).
Perhaps the most dramatic instances in the evolution of DT have occurred in the angiosperms. As mentioned, DT is common and widespread in the reproductive propagules of the vascular plants, but DT in the vegetative tissues (i.e., shoots, roots, and leaves) was lost early in the evolution of the tracheophytes (or possibly was not present in the most recent common ancestor (MRCA) of the tracheophytes and their sister bryophyte lineage). However, vegetative DT has arisen at least once in the lycophyte genus Selaginella P. Beauv. and once in the fern clade. Remarkably, this complex trait has independently evolved at least eight times in the angiosperms (Oliver et al., 2000a). Forms of vegetative DT have been reported for approximately 70 species of flowering plants (Porembski and Barthlott, 2000), which are often collectively referred to as "resurrection plants."
The mechanisms conferring vegetative DT in the vascular plants differ fundamentally from those involved in the bryophyte DT response, thus the vegetative DT of resurrection plants has been coined "modified DT" (Oliver et al., 1998, 2000a). Modified DT in the vascular plants (exemplified by the response of model DT angiosperms such as Craterostigma plantagineum Hochst. and members of the genus Xerophyta Juss.) relies more on protective mechanisms induced during slow drying (Bewley and Oliver, 1992; Bartels et al., 1996, 1997; Bernacchia et al., 1996; Oliver et al., 2000a; Hoekstra et al., 2001; Bartels, 2005). In contrast, vegetative DT in bryophytes such as the moss Syntrichia ruralis (Hedw.) F. Weber & D. Mohr. [referred to throughout this paper by its older synonym, Tortula ruralis (Hedw.) P. Gaertn., B. Mey. & Scherb.] involves proteins that are constitutively expressed. Because DT-related transcripts are relatively abundant at all times, the DT response in this moss is not as sensitive to the rate of drying and is based more heavily on cellular repair mechanisms that become active upon rehydration (Oliver et al., 1993, 2000a, b; Wood et al., 2000).
Considerable attention has been directed at understanding the proteins involved in the vegetative DT of resurrection plants (Piatkowski et al., 1990; Bartels et al., 1996; Ingram and Bartels, 1996; Bockel et al., 1998; Ditzer et al., 2001; Smith-Espinoza et al., 2003, 2005; Collett et al., 2004; Bartels, 2005; Rascio and La Rocca, 2005). One the most significant components of the DT response characterized in these studies is the group known as the late embryogenesis abundant (LEA) proteins. In addition to vegetative DT, the LEAs are expressed under a variety of abiotic stress conditions and during seed development (Wise and Tunnacliffe, 2004; Tunnacliffe and Wise, 2007). Despite their common expression patterns and typical hydrophilicity (the D95 LEA family is an exception, e.g., Galau et al., 1993), the LEA proteins constitute a phylogenetically heterogeneous group; that is, they are unlikely to share common ancestry (but see Tunnacliffe and Wise, 2007 for an alternative interpretation). The LEA proteins identified to date have been assigned to several families on the basis of conserved peptide motifs (Dure, 1993b; Wise, 2003); Table 1 summarizes the different nomenclatures that have been applied to the putative LEA families to date (but note that this is a relatively inclusive list of possible LEA families; for a more exclusive treatment of LEAs, see Tunnacliffe and Wise, 2007). Various functions for LEAs have been proposed, including roles as antioxidants and membrane or protein stabilizers, but the precise mechanism of their function remains unknown.
|
; Baker et al., 1988; Dure et al., 1989, 1993a; Boudet et al., 2006). Because of their prevalence in the vegetative DT responses of Craterostigma (Piatkowski et al., 1990; Bockel et al., 1998; Velasco et al., 1998; Bartels, 2005), Xerophyta (Collett et al., 2004) and T. ruralis (Oliver et al., 2004), the LEA families have become the most conspicuous element in the literature on the molecular aspects of DT, despite their relatively enigmatic functional status.
In those cases where angiosperm lineages have independently evolved vegetative DT, it is frequently hypothesized that the key genes involved (such as the LEAs) were recruited from processes related to seed dormancy in desiccation sensitive ancestors (Oliver et al., 2000a; Bartels and Salamini, 2001; Illing et al., 2005; Rascio and La Rocca, 2005; Tunnacliffe and Wise, 2007). However, many of the genes that are expressed during drying in resurrection plants have homologs that are upregulated in response to abiotic stresses (e.g., salt, freezing, and drying) in desiccation-sensitive plants such as Arabidopsis (Appendix 1). For instance, in desiccation-sensitive plants, LEAs have been shown to be an important component of the drought and cold stress responses (Kiyosue et al., 1994; Close, 1997; Rinne et al., 1998; Romo et al., 2001; Ali-Benali et al., 2005; Park et al., 2005). Therefore, an alternative scenario for the origin of modified vegetative DT in resurrection plants is one in which this trait evolved through adaptation of the abiotic stress response pathways present in a desiccation sensitive ancestor (Illing et al., 2005). Of course, these two scenarios for the origins of vegetative DT in the resurrection plants are not mutually exclusive, and some overlap in function (i.e., expression in seed development and in stress response) of the multiple gene families involved in DT could be expected.
The origins of vegetative DT in resurrection plants have important implications for agriculture and crop improvement, particularly in the context of global climate change. Understanding the orthology of DT-related genes in flowering plants aids in identifying pathways and genes that might be manipulated to effect desiccation tolerance in drought-sensitive, agriculturally important species (Bartels and Salamini, 2001).
To date, no studies have addressed the question of the origins of vegetative DT in the angiosperms from a phylogenetic perspective. The relationship of genes expressed in the desiccation responses of DT plants to those expressed in non-DT angiosperms (e.g., Arabidopsis) has so far been established through sequence similarity alone (Collett et al., 2004; Oliver et al., 2004; Illing et al., 2005). While putative orthology assessments conducted via similarity searches (e.g., BLAST; Altschul et al., 1990) are frequently employed in functional genomic studies, sequence similarity alone is not always an accurate indicator of evolutionary relationships (Swofford et al., 1996; Koski and Golding, 2001; Sicheritz-Ponten and Andersson, 2001). Similarity-based approaches fail to distinguish between processes such as lateral gene transfer, gene duplication, gene loss, and differential substitution rates, all of which are important factors influencing genome evolution (Eisen, 1998; Eisen and Hanawalt, 1999).
The current study employs an alternative phylogenomic approach to understanding the functional evolution of gene families involved in the vegetative DT of resurrection plants. As outlined, the complex trait of DT has potentially undergone multiple independent origins in three main contexts throughout land plant evolution: the origins of vegetative DT in the gametophytes of the bryophyte lineages, the origin(s) of seed DT, and the multiple origins of modified vegetative DT in the angiosperm clade. The focus of the current study concentrates on this last phenomenon, the convergent evolution of vegetative DT in the resurrection plants, and attempts to discriminate between two alternative hypotheses for the independent origins of vegetative DT in the angiosperms (i.e., co-option from processes related to either seed maturation or stress responses) by reconstructing the expression patterns of the genes ancestral to those that are now expressed in the DT response of resurrection plants.
In this study, I used the complete genome sequences of the moss Physcomitrella patens and the angiosperm Arabidopsis thaliana and EST data from the DT moss Tortula ruralis and the DT angiosperms Craterostigma plantagineum and Xerophyta humilis (Baker) T. Durand & Schinz to estimate phylogenies for four LEA gene families. While this study focuses on questions related to the angiosperm clade, I included LEA homologs from non-angiosperm (i.e., moss) lineages that demonstrate varying degrees of vegetative DT, to investigate the possibility of orthology between genes expressed in these lineages and the angiosperm lineages in question. DNA microarray expression data from Arabidopsis were then used in a Bayesian framework to reconstruct ancestral expression patterns for the MRCAs of clades containing genes expressed by both DT and non-DT angiosperms.
MATERIALS AND METHODS
Sequences and alignments
To identify those LEA gene families and transcripts involved in the DT response, LEA transcripts from the desiccation expressed sequence tag (EST) libraries of Tortula, Craterostigma, and Xerophyta were identified via searches on sequence annotations. LEA transcripts from this first search were then translated into conceptual amino acid sequences and used for subsequent tBLASTn (Altschul et al., 1990) searches of the NCBI GenBank EST database (http://www.ncbi.nlm.nih.gov/dbEST/). The conceptual amino acid translations were then used to search the TAIR database (http://www.arabidopsis.org) to locate potential Arabidopsis homologs. This search was supplemented with an additional search of the Floral Genome Project's PlantTribes database (Wall et al., 2008) to confirm that all the potential Arabidopsis homologs for each gene family had been detected and included. In addition, Arabidopsis amino acid sequences from all potential LEA families (Table 1) were used to search the EST libraries of Tortula, Craterostigma, and Xerophyta to ensure that no LEA transcripts had been missed due to incomplete annotation. Finally, conceptual amino acid translations of sequences from each LEA gene family were used to perform tBLASTn searches of the Physcomitrella genome (http://www.jgi.doe.gov).
Alignments of the amino acid translations were performed for each gene family using T-COFFEE (Notredame et al., 2000). These alignments were subsequently reviewed and manipulated manually if necessary. Columns missing data for a majority of the sequences were manually masked to exclude them from later analyses. In all, four of the possible 10 named LEA groups were used in this study. Of those LEA families excluded from this analysis, the D19 LEAs are not currently known to be expressed in the vegetative tissues of any DT angiosperms and are expressed solely in the seeds of Arabidopsis; the D29 and Lea76 groups are thought to be a subset of the broader D7 (group 3) (Wise, 2003) LEAs (and are therefore subsumed in the D7 group in this study); the D73 LEAs are not present across all land plants (they are restricted to the vascular plants); and the D113 and D34 LEAs do not have stress-only expression in Arabidopsis.
Phylogenetic analyses
Aligned amino acid sequences were back-translated to nucleotide sequences using the program RevTrans (Wernersson and Pedersen, 2003), and the resulting alignments were imported into the program MrBayes 3.1.1 (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003) for analysis. Bayesian Markov chain Monte Carlo (MCMC) analyses were performed for each alignment under a general time reversible (GTR) model with gamma distributed across-site rate variation and a proportion of invariant sites. State frequencies, nucleotide substitution rates, and the shape parameter of the gamma distribution of rate variation were estimated from the data using the default prior settings. Three of the LEA groups, D7, D11, and D95, displayed considerable sequence divergence, so additional Bayesian analyses were performed on these groups using more conservative amino acid sequence alignments. For these analyses, the MCMC was allowed to explore all of the fixed-rate amino acid models available in MrBayes, and each model's relative contribution to the results was directly proportional to its posterior probability (Ronquist and Huelsenbeck, 2003). In all analyses, two chains were run for 2 000 000 generations with a burn-in of 5000 trees. Trees were sampled from each chain every 500 generations, and at the end of each run convergence of the two chains was confirmed (standard deviation of split frequencies 
0.01). Majority-rule consensus trees were created from the distribution of trees sampled during the MCMC runs, and these consensus trees were used to assess the posterior probabilities of subclades.
Expression data
Microarray expression data for Arabidopsis genes were obtained from the AtGenExpress project (http://web.uni-frankfurt.de/fb15/botanik/mcb/AFGN/atgenex.htm [in German]). Expression data for each locus were sorted by replicate hybridization set percentile, and signals exceeding the 50th percentile were examined to establish experimental variables (i.e., treatments and tissues) under which a gene was expressed. A gene was considered upregulated in a particular tissue (e.g., seed) or under a particular condition (e.g., stress) if the signal for that tissue or condition was at least double that of the expression signal for all of the nonstressed vegetative (shoot, leaf, or root) tissues. Expression that was upregulated under stress was coded as 1, in the seed or pollen as 2, and both under stress and in the seed and/or pollen as 3. If all control shoot and root expression levels were not at least two times lower than seed/pollen or stress treatment signals, then expression was coded as 4 for vegetative/housekeeping (Appendix 1).
A subset of the Physcomitrella LEA genes included in this study was shown to be upregulated by stress in a recent DNA microarray experiment (Cuming et al., 2007). These genes were coded as being expressed under stress; all other genes from Physcomitrella were code as missing data. DT-expressed genes from Craterostigma and Xerophyta were left uncoded, as this study assumes that the gain of DT-associated expression is a derived condition that arose along the terminal branches of the gene phylogenies (i.e., the distribution of vegetative DT in the angiosperms supports the idea that this trait was not present in the MRCA of Arabidopsis + Craterostigma and/or Xerophyta). While the vegetative DT-expression of the LEA homologs from the resurrection plants (Craterostigma and Xerophyta) was not explicitly incorporated into the ancestral character state reconstructions, the placement of these homologs in the gene phylogenies was used to identify which nodes (i.e., MRCAs) would be the focus of the ancestral state reconstructions.
Ancestral character state reconstruction
Ancestral character state reconstruction requires rooted trees, so the program NOTUNG (Durand et al., 2006) was used to locate the best branch on which to root each gene tree. NOTUNG reconciles a gene tree with a species tree, and then identifies the root(s) that minimizes duplications and losses on the gene tree (Durand et al., 2006). If multiple rootings yielded identical numbers of duplications and losses, ancestral character state reconstruction analyses were performed for each alternative rooting.
A Bayesian MCMC approach to ancestral character state estimation was chosen for this study. In contrast to reconstructing ancestral states under an optimality criterion (e.g., parsimony) on a single tree, the Bayesian MCMC method has the advantage of integrating uncertainty in tree topology and character state assignment into the posterior probability for a given state reconstruction (Pagel et al., 2004). Ancestral state estimations were performed using BayesMultiState (Pagel et al., 2004; Pagel and Meade, 2007), part of the BayesTraits package (http://www.evolution.rdg.ac.uk).
Procedures for the ancestral state analyses were performed following the recommended procedures outlined in the BayesTraits documentation (Pagel and Meade, 2007). For each gene family, 3000 post burn-in trees from the MrBayes analyses were imported into BayesTraits. A uniform prior was used for models of trait evolution and a gamma prior was used for the rate coefficients. The mean and variance of the gamma prior were seeded from a uniform hyperprior, which allowed the program to estimate the gamma prior from the data (Pagel and Meade, 2006). The ranges for the uniform distribution used to seed the prior were obtained using an empirical approach: MultiState (Pagel and Meade, 2007) was used to estimate the rate coefficients on each of the 3000 trees under maximum likelihood. The mean and variance ranges for the hyperprior were then specified based on the maximum likelihood estimates of the rate coefficients. For each LEA family, several preliminary exploratory chains were run, iteratively adjusting the value of the rate coefficient proposals (the ratedev parameter) until an acceptance rate of 20–40% was achieved (Pagel and Meade, 2007). Each chain was run for 5 050 000 iterations and was sampled every 1000 generations (burn-in 50 000 generations). During the first MCMC analysis for a particular node of interest, the posterior probability of each alternative character state reconstruction was estimated using the AddMRCA command. For each subsequent run, a node of interest was fixed at one of the alternative character states, and the harmonic mean of the likelihood of the trees (over the sampling interval) was recorded after 5 000 000 iterations. Due to the instability of the harmonic mean, this procedure was repeated three times for each character state to ensure that an appropriate value had been estimated. A Bayes factor test was performed using harmonic means obtained from runs fixing alternative character states (BF = twice the difference between the two numbers) to gauge the support for one character state reconstruction over its alternative(s) (Pagel and Meade, 2007).
RESULTS
The four gene families analyzed in this study and their constituent gene members are detailed in Appendix 1. All the potential LEA genes from the Arabidopsis and Physcomitrella genomes were included in the alignments for the D11, D95, and LEA 10 families. In Physcomitrella, 17 members of the D7 LEA family were identified as belonging to a large paralogous cluster (as indicated by preliminary phylogenetic analysis) lacking any close relationship to angiosperm genes. To improve the efficiency of the Bayesian ancestral state reconstructions, I omitted this paralogous cluster of Physcomitrella D7 genes from further analyses. NEXUS files containing LEA family alignments and trees are available as online supplements, Appendices S1–S6 (see Supplemental Data with online version of this article).
Of the LEA families analyzed, one, D11, yielded phylogenies with multiple equally parsimonious rootings (as determined by NOTUNG; Durand et al., 2006). Phylogenies for each of the four LEA families are presented in Figs. 2–5. One notable pattern emerges from these LEA gene family phylogenies: in most of the LEA gene family phylogenies that include multiple homologs from DT angiosperms, the DT angiosperm sequences form fairly exclusive, monophyletic groups. A Xerophyta sequence in the D7 family presents the only exception to this pattern (Fig. 5).
|
|
The D95 LEA phylogeny (Fig. 3) resolves into two well-supported subclades containing both moss and angiosperm homologs. One subclade contains several moss sequences and the constitutively expressed At2g4460 from Arabidopsis. The other subclade contains the CDeT27-45 gene from Craterostigma, sister to two Arabidopsis genes expressed in both reproductive propagules and stressed vegetative tissues. The CK906400 transcript from Xerophyta also falls into this subclade, but its exact placement is ambiguous due to weak support (Fig. 3).
Only one DT-expressed transcript (from Xerophyta) is present in the LEA 10 phylogeny (Fig. 4 Its sister relationship to the one Arabidopsis gene with seed/pollen expression is well supported, and their MRCA is estimated as having seed/pollen expression (Table 2). The remaining Arabidopsis homologs with combined seed/pollen + stress expression form a separate clade (Fig. 4) with high posterior probability. The Arabidopsis gene locus At5g46530, which appears homologous to the other LEA 10 genes, is not represented in the AtGenExpress experiments. With respect to the relative abundance of Physcomitrella homologs, the LEA 10 family is unique among the LEA families presented in this study. The Arabidopsis genome tends to contain more genes per family than the Physcomitrella genome in the LEA groups investigated here. However, Physcomitrella has twice the number of LEA 10 homologs (several of which are expressed during abiotic stress responses) as Arabidopsis (8 vs. 4, respectively).
DT-expressed angiosperm sequences (except for the Xerophyta transcript CK906406) form a well-supported clade in the D7 LEA gene phylogeny (Fig. 5). Arabidopsis genes display a range of expression patterns, but the phylogenetic resolution of the Arabidopsis homologs with stress-only expression is particularly notable. One of the loci with stress-only expression, At3g62580, is nested with high posterior probability within a clade of Physcomitrella D7 genes (Fig. 5). In contrast, the other two exclusively stress-expressed Arabidopsis loci (At2g42530 and At2g42540) form what appears to be a paralogous pair, nested well within a clade of Arabidopsis genes with seed/pollen expression. This placement suggests that these two stress-expressed genes in Arabidopsis may have undergone a more recent duplication and possibly represent a case in which stress expression was derived from seed/pollen expression.
DISCUSSION
The most compelling result from these analyses is that none of the ancestral nodes were estimated to have an expression pattern that was exclusively stress related. All the significant reconstructions included seed and/or pollen expression either exclusively or in combination with stress expression. Thus, the ancestral character state reconstructions presented here contradict the hypothesis that modified vegetative DT arose via a modified stress tolerance pathway and support the hypothesis that modified vegetative DT in angiosperms arose through the adoption of mechanisms related to the onset of DT in reproductive propagules.
The one possible exception is present in the D7 LEA family. One gene member from this family (GB CK906406), present in the dehydration EST libraries of the DT angiosperm Xerophyta humilis, is resolved as sister to a large and diverse clade of D7 LEA genes. This topology (and the overall lack of support for this placement of CK906406 in the posterior distribution of trees) made ancestral state estimation for this node tenuous. However, the fact that the large clade that includes CK906406 contains members from Physcomitrella indicates that this particular gene may be derived from a MRCA with nonseed (i.e., stress or spore) expression.
Gene family phylogenies are widely employed in the annotation and functional prediction of unknown proteins (Eisen and Fraser, 2003; Sjolander, 2004; Engelhardt et al., 2005; Brown and Sjolander, 2006). The rationale underpinning these techniques is that there is functional conservation in genes sharing recent common ancestry. The current study borrowed from this approach to understanding gene function, but instead of comparing contemporary orthologous gene pairs, the study is an attempt to estimate ancestral gene expression based on the phylogenetic distribution of expression states in extant genes.
Some problematic issues could arise with this approach and should be addressed here. First, a major concern with this approach is that the relative abundance of a particular character state in the genes included in the analysis might bias the ancestral state estimates. For this reason, the D34 and D113 LEA families were excluded from this analysis, and only those LEA families containing members with exclusive stress expression or constitutive expression patterns were analyzed (Appendix 1). Second, the Bayesian posterior support for some of the reconstructed nodes in Figs. 2–5 is low, which might raise legitimate concerns about the reliability of the ancestral state estimations. However, it should be noted that the ancestral state reconstructions were iterated over the entire posterior distribution of trees (not just the tree selected for display), so the final ancestral state estimates incorporate character states estimated on topologies with more inclusive clades containing these sequences plus additional sequences (Pagel et al., 2004). For instance, in the case of the expression patterns estimated for the MRCA of the six D11 genes in Fig. 2, the character state estimations for the MRCA of the clade with 98% posterior probability in Fig. 2 is likely to have had a significant influence on the final state estimations for the clade containing the six sequences in question. Finally, a clear weakness of the current study is the large amount of missing expression data (essentially, this information was only available for Arabidopsis homologs). This uncertainty, coupled with the uncertainty inherent in the phylogenetic reconstructions of deeply divergent gene families (the MRCA of Physcomitrella and Arabidopsis existed ca. 450 million years ago) was the main motivation for using a method that could estimate ancestral states in a probabilistic framework (Pagel et al., 2004). Because only the observed data are fixed in this method (the probabilities of each character state in the ancestral nodes are summed over all potential histories for the nodes without assigning fixed states to the internal nodes), the posterior distribution of the ancestral character states will never be inconsistent with respect to the observed character states (Pagel et al., 2004). In other words, a lack of character data should be reflected in weaker support for a given character state reconstruction, but should not result in a reconstruction that does not agree with the available character data. As rich genomic resources continue to expand in the future, studies utilizing an approach similar to the one described in this paper will be able to incorporate additional data and should benefit from the added breadth that key taxa and their expression patterns will provide. Furthermore, with the availability of genomic data for other important land plant groups such as liverworts, hornworts, ferns, lycophytes, and gymnosperms, this approach should be applicable to the investigation of other important transitions in the evolution of expression patterns in gene families.
The results support the conclusions of Illing et al. (2005) and are consistent with the evolutionary scenario for DT proposed by Oliver et al. (2000b). Throughout the history of plants, the same core LEA gene families have been repeatedly involved in the evolution of DT. In the angiosperms, these same LEA families were recruited for related functions in abiotic stress tolerance and the production of reproductive propagules. The implication of the results presented here is that in those instances where vegetative DT independently evolved in the angiosperms, evolution favored the use of LEA genes most closely related to those functioning in the maturation of reproductive propagules.
The question this raises is why vegetative DT evolution should have favored the recruitment of a propagule-related subset of LEA genes over LEAs from stress response pathways. One possibility is that LEA genes involved in seed and pollen development have been preferentially appropriated for vegetative DT due to some aspect of their regulation. A natural direction for future research in this system would therefore be a comparison of the cis-regulatory structures of Arabidopsis orthologs of DT-expressed genes and closely related Arabidopsis genes that are not sister to DT-functioning genes. Repeating this procedure for many gene families could potentially identify common features in the regulatory architecture of clades that contain DT genes.
The onset of DT in seeds is a strictly regulated component of plant development (Goupil et al., 1992; Delseny et al., 2001). If vegetative DT in the resurrection plants is primarily derived from seed/pollen DT, then the emergence of this trait might best be considered an instance of developmental evolution, as suggested by Bartels and Salamini (2001). Developmentally regulated genetic pathways ancestrally restricted to angiosperm seeds may have been imported into the vegetative tissues, where they evolved to respond to environmental rather than developmental cues (Bartels and Salamini, 2001). This scenario predicts that vegetative DT could evolve through the co-option of a coordinated gene network related to seed development, a possibility that can be investigated in the future.
Appendix 1. LEA gene families, their constituent gene members and gene expression patterns. Where applicable, corresponding protein names and/or gene locus IDs are provided in addition to GenBank accession numbers. Expression codes are as follows: (1) stress only, (2) seed/pollen + stress, (3) seed/pollen only, (4) constitutive.
|
FOOTNOTES
1 The author thanks B. Mishler and M. Oliver, whose insights and suggestions provided the initial motivation for this project. Thanks also to M. Oliver for generously sharing T. ruralis EST results. The manuscript and methods further benefited from the thoughtful input of T. Vision, S. Otto, and B. Sidlauskas. This project was completed during a postdoctoral fellowship at the NSF-sponsored National Evolutionary Synthesis Center (NESCent). 
2 Author for correspondence (e-mail: kirstenfisher{at}nescent.org)
LITERATURE CITED
Ali-Benali, M. A., R. Alary, P. Joudrier, AND M. F. Gautier. 2005. Comparative expression of five Lea genes during wheat seed development and in response to abiotic stresses by real-time quantitative RT-PCR. Biochimica et Biophysica Acta—Gene Structure and Expression 1730: 56–65.[CrossRef]
Alpert, P. 2006. Constraints of tolerance: Why are desiccation-tolerant organisms so small or rare? The Journal of Experimental Biology 9: 1575–1584.
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, AND D. J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology 215: 403–410.[CrossRef][Web of Science][Medline]
Attree, S. M., AND L. C. Fowke. 1993. Embryogeny of gymnosperms: Advances in synthetic seed technology of conifers. Plant Cell, Tissue and Organ Culture 35: 1–35.[Medline]
Baker, J., C. Steele, AND L. Dure. 1988. Sequence and characterization of 6 Lea proteins and their genes from cotton. Plant Molecular Biology 11: 277–291.[CrossRef][Web of Science]
Bartels, D. 2005. Desiccation tolerance studied in the resurrection plant Craterostigma plantagineum. Integrative and Comparative Biology 45: 696–701.
Bartels, D., J. Chandler, C. Bockel, W. Frank, M. Kleines, M. J. Rodrigo, J. Phillips, J.-B. Mariaux, A. Furini, AND F. Salamini. 1997. Investigating the molecular basis of desiccation tolerance using the resurrection plant Craterostigma plantagineum as an experimental system. Acta Physiologiae Plantarum 19: 399–403.[CrossRef][Web of Science]
Bartels, D., A. Furini, J. Ingram, AND F. Salamini. 1996. Responses of plants to dehydration stress: A molecular analysis. Plant Growth Regulation 20: 111–118.[CrossRef][Web of Science]
Bartels, D., AND F. Salamini. 2001. Desiccation tolerance in the resurrection plant Craterostigma plantagineum. A contribution to the study of drought tolerance at the molecular level. Plant Physiology 127: 1346–1353.
Bernacchia, G., F. Salamini, AND D. Bartels. 1996. Molecular characterization of the rehydration process in the resurrection plant Craterostigma plantagineum. Plant Physiology 111: 1043–1050.[Abstract]
Bewley, J. D., AND M. J. Oliver. 1992. Desiccation tolerance in vegetative plant tissues and seeds: Protein synthesis in relation to desiccation and a potential role for protection. In G. N. Somero, C. B. Osmond, and C. L. Bolis [eds.], Water and life: Comparative analysis of water relationships at the organismic, cellular, and molecular levels, 141–160. Springer-Verlag, New York, New York, USA.
Bockel, C., F. Salamini, AND D. Bartels. 1998. Isolation and characterization of genes expressed during early events of the dehydration process in the resurrection plant Craterostigma plantagineum. Journal of Plant Physiology 152: 158–166.[Web of Science]
Boudet, J., J. Buitink, F. A. Hoekstra, H. Rogniaux, C. Larre, P. Satour, AND O. Leprince. 2006. Comparative analysis of the heat stable proteome of radicles of Medicago truncatula seeds during germination identifies late embryogenesis abundant proteins associated with desiccation tolerance. Plant Physiology 140: 1418–1436.
Bray, E. A. 1994. Alterations in gene expression in response to water deficit. In A. S. Basra [ed.], Stress-induced gene expression in plants, 1–23. Harwood Academic, Newark, New Jersey, USA.
Brown, D., AND K. Sjolander. 2006. Functional classification using phylogenomic inference. PLoS Computational Biology 2: 479–483.[Web of Science]
Close, T. J. 1997. Dehydrins: A commonality in the response of plants to dehydration and low temperature. Physiologia Plantarum 100: 291–296.[CrossRef]
Collett, H., A. Shen, M. Gardner, J. M. Farrant, K. J. Denby, AND N. Illing. 2004. Towards transcript profiling of desiccation tolerance in Xerophyta humilis: Construction of a normalized 11 k X. humilis cDNA set and microarray expression analysis of 424 cDNAs in response to dehydration. Physiologia Plantarum 122: 39–53.[CrossRef]
Cuming, A. C., S. H. Cho, Y. Kamisugi, H. Graham, AND R. S. Quatrano. 2007. Microarray analysis identifies coordinated expression of stress responsive genes in the moss, Physcomitrella patens. New Phytologist 176: 275–287.[CrossRef][Web of Science][Medline]
Delseny, M., N. Bies-Etheve, C. Carles, G. Hull, C. Vicient, M. Raynal, F. Grellet, AND L. Aspart. 2001. Late Embryogenesis Abundant (LEA) protein gene regulation during Arabidopsis seed maturation. Journal of Plant Physiology 158: 419–427.[CrossRef][Web of Science]
Ditzer, A., H. H. Kirch, A. Nair, AND D. Bartels. 2001. Molecular characterization of two alanine-rich Lea genes abundantly expressed in the resurrection plant C. plantagineum in response to osmotic stress and ABA. Journal of Plant Physiology 158: 623–633.[CrossRef][Web of Science]
Durand, D., B. V. Halldorsson, AND B. Vernot. 2006. A hybrid micro-macroevolutionary approach to gene tree reconstruction. Journal of Computational Biology 13: 320–335.[CrossRef][Web of Science][Medline]
Dure, L. 1993a. A repeating 11-mer amino-acid motif and plant desiccation. Plant Journal 3: 363–369.[CrossRef][Web of Science][Medline]
Dure, I. L. 1993b. Structural motifs in LEA proteins. In T. J. Close, and E. A. Bray [ed.], Plant responses to cellular dehydration during environmental stress, 91–103. American Society of Plant Physiologists, Rockville, Maryland, USA.
Dure, L., M. Crouch, J. Harada, T. H. D. Ho, J. Mundy, R. Quatrano, T. Thomas, AND Z. R. Sung. 1989. Common amino acid sequence domains among the Lea proteins of higher plants. Plant Molecular Biology 12: 475–486.[CrossRef][Web of Science]
Eisen, J. A. 1998. A phylogenomic study of the MutS family of proteins. Nucleic Acids Research 26: 4291–4300.
Eisen, J. A., AND C. M. Fraser. 2003. Phylogenomics: Intersection of evolution and genomics. Science 300: 1706–1707.
Eisen, J. A., AND P. C. Hanawalt. 1999. A phylogenomic study of DNA repair genes, proteins, and processes. Mutation Research—DNA Repair 435: 171–213.[CrossRef]
Engelhardt, B. E., M. I. Jordan, K. E. Muratore, AND S. E. Brenner. 2005. Protein molecular function prediction by Bayesian phylogenomics. PLoS Computational Biology 1: 432–445.[Web of Science]
Forsyth, C., AND J. Vanstaden. 1983. Germination of cycad seeds. South African Journal of Science 79: 8–9.[Web of Science]
Galau, G. A., D. W. Hughes, AND L. Dure. 1986. Abscisic acid induction of cloned cotton late embryogenesis-abundant (Lea) mRNAs. Plant Molecular Biology 7: 155–170.[CrossRef][Web of Science]
Galau, G. A., H. Y. C. Wang, AND D. W. Hughes. 1993. Cotton Lea5 and Lea14 encode atypical late embryogenesis-abundant proteins. Plant Physiology 101: 695–696.[CrossRef][Web of Science][Medline]
Goupil, P., P. Hatzopoulos, G. Franz, F. D. Hempel, R. L. You, AND Z. R. Sung. 1992. Transcriptional regulation of a seed-specific carrot gene, DC8. Plant Molecular Biology 18: 1049–1063.[CrossRef][Web of Science][Medline]
Harada, J. J., A. J. DeLisle, C. S. Baden, AND M. L. Crouch. 1989. Unusual sequence of an abscisic acid-inducible mRNA which accumulates late in Brassica napus seed development. Plant Molecular Biology 12: 395–401.[CrossRef][Web of Science]
Hoekstra, F. A., E. A. Golovina, AND J. Buitink. 2001. Mechanisms of plant desiccation tolerance. Trends in Plant Science 6: 431–438.[CrossRef][Web of Science][Medline]
Huelsenbeck, J. P., AND F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics (Oxford, England) 17: 754–755.[CrossRef]
Illing, N., K. J. Denby, H. Collett, A. Shen, AND J. M. Farrant. 2005. The signature of seeds in resurrection plants: A molecular and physiological comparison of desiccation tolerance in seeds and vegetative tissues. Integrative and Comparative Biology 45: 771–787.
Ingram, J., AND D. Bartels. 1996. The molecular basis of dehydration tolerance in plants. Annual Review of Plant Physiology and Plant Molecular Biology 47: 377–403.[CrossRef][Web of Science][Medline]
Kiyosue, T., K. Yamaguchishinozaki, AND K. Shinozaki. 1994. Characterization of 2 cDNAs (Erd10 and Erd14) corresponding to genes that respond rapidly to dehydration stress in Arabidopsisthaliana. Plant &Cell Physiology 35: 225–231.
Koski, L. B., AND G. B. Golding. 2001. The closest BLAST hit is often not the nearest neighbor. Journal of Molecular Evolution 52: 540–542.[Web of Science][Medline]
Notredame, C., D. Higgins, AND J. Heringa. 2000. T-Coffee: A novel method for multiple sequence alignments. Journal of Molecular Biology 302: 205–217.[CrossRef][Web of Science][Medline]
Oliver, M. J., S. E. Dowd, J. Zaragoza, S. A. Mauget, AND P. R. Payton. 2004. The rehydration transcriptome of the desiccation-tolerant bryophyte Tortula ruralis: Transcript classification and analysis. BMC Genomics 5: 89–107.[CrossRef][Medline]
Oliver, M. J., B. D. Mishler, AND J. E. Quisenberry. 1993. Comparative measures of desiccation-tolerance in the Tortula ruralis complex. I. Variation in damage control and repair. American Journal of Botany 80: 127–136.[CrossRef][Web of Science]
Oliver, M. J., Z. Tuba, AND B. D. Mishler. 2000a. The evolution of vegetative desiccation tolerance in land plants. Plant Ecology 151: 85–100.[CrossRef][Web of Science]
Oliver, M. J., J. Velten, AND A. J. Wood. 2000b. Bryophytes as experimental models for the study of environmental stress tolerance: Tortula ruralis and desiccation tolerance in mosses. Plant Ecology 151: 73–84.[CrossRef][Web of Science]
Oliver, M. J., A. J. Wood, AND P. O'Mahony. 1998. "To dryness and beyond"—Preparation for the dried state and rehydration in vegetative desiccation-tolerant plants. Plant Growth Regulation 24: 193–201.[CrossRef][Web of Science]
Pagel, M., AND A. Meade. 2006. Bayesian analysis of correlated evolution of discrete characters by reversible-jump Markov chain Monte Carlo. American Naturalist 167: 808–825.[CrossRef][Web of Science]
Pagel, M., AND A. Meade 2007. BayesTraits, version1.0. Website http://www.evolution.rdg.ac.uk
Pagel, M., A. Meade, AND D. Barker. 2004. Bayesian estimation of ancestral character states on phylogenies. Systematic Biology 53: 673–684.[CrossRef][Web of Science][Medline]
Pammenter, N. W., AND P. Berjak. 2000. Evolutionary and ecological aspects of recalcitrant seed biology. Seed Science Research 10: 301–306.[Web of Science]
Park, B. J., Z. C. Liu, A. Kanno, AND T. Kameya. 2005. Increased tolerance to salt- and water-deficit stress in transgenic lettuce (Lactuca sativa L.) by constitutive expression of LEA. Plant Growth Regulation 45: 165–171.[CrossRef][Web of Science]
Piatkowski, D., K. Schneider, F. Salamini, AND D. Bartels. 1990. Characterization of 5 abscisic acid-responsive cDNA clones isolated from the desiccation-tolerant plant Craterostigma plantagineum and their relationship to other water-stress genes. Plant Physiology 94: 1682–1688.
Porembski, S., AND W. Barthlott. 2000. Granitic and gneissic outcrops (inselbergs) as centers of diversity for desiccation-tolerant vascular plants. Plant Ecology 151: 19–28.[CrossRef][Web of Science]
Rascio, N., AND N. La Rocca. 2005. Resurrection plants: The puzzle of surviving extreme vegetative desiccation. Critical Reviews in Plant Sciences 24: 209–225.[CrossRef][Web of Science]
Rinne, P., A. Welling, AND P. Kaikuranta. 1998. Onset of freezing tolerance in birch (Betula pubescens Ehrh.) involves LEA proteins and osmoregulation and is impaired in an ABA-deficient genotype. Plant, Cell &Environment 21: 601–611.[CrossRef]
Romo, S., E. Labrador, AND B. Dopico. 2001. Water stress-regulated gene expression in Cicer arietinum seedlings and plants. Plant Physiology and Biochemistry 39: 1017–1026.[CrossRef][Web of Science]
Ronquist, F., AND J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (Oxford, England) 19: 1572–1574.[CrossRef]
Sicheritz-Ponten, T., AND S. G. E. Andersson. 2001. A phylogenomic approach to microbial evolution. Nucleic Acids Research 29: 545–552.
Sjolander, K. 2004. Phylogenomic inference of protein molecular function: Advances and challenges. Bioinformatics (Oxford, England) 20: 170–179.[CrossRef]
Smith-Espinoza, C. J., A. Richter, F. Salamini, AND D. Bartels. 2003. Dissecting the response to dehydration and salt (NaCl) in the resurrection plant Craterostigma plantagineum. Plant, Cell &Environment 26: 1307–1315.[CrossRef]
Smith-Espinoza, C. J., J. R. Phillips, F. Salamini, AND D. Bartels. 2005. Identification of further Craterostigma plantagineum cdt mutants affected in abscisic acid mediated desiccation tolerance. Molecular Genetics and Genomics 274: 364–372.[CrossRef][Web of Science][Medline]
Swofford, D. L., G. J. Olsen, P. J. Waddell, AND D. M. Hillis. 1996. Phylogenetic inference. In D. M. Hillis, C. Moritz, and B. K. Mable [eds.], Molecular systematics, 2nd ed., 407–514. Sinauer, Sunderland, Massachusetts, USA.
Tunnacliffe, A., AND M. J. Wise. 2007. The continuing conundrum of the LEA proteins. Die Naturwissenschaften 94: 791–812.[CrossRef][Web of Science][Medline]
Tweddle, J. C., J. B. Dickie, C. C. Baskin, AND J. M. Baskin. 2003. Ecological aspects of seed desiccation sensitivity. Journal of Ecology 91: 294–304.[CrossRef][Web of Science]
Velasco, R., F. Salamini, AND D. Bartels. 1998. Gene structure and expression analysis of the drought- and abscisic acid-responsive CDeT11-24 gene family from the resurrection plant Craterostigma plantagineum Hochst. Planta 204: 459–471.[CrossRef][Web of Science][Medline]
Wall, P. K., J. Leebens-Mack, K. F. Müller, D. Field, N. S. Altman, AND C. W. dePamphilis. 2008. PlantTribes: A gene and gene family resource for comparative genomics in plants. Nucleic Acids Research 36: D970–D976.
Wernersson, R., AND A. G. Pedersen. 2003. RevTrans: Multiple alignment of coding DNA from aligned amino acid sequences. Nucleic Acids Research 31: 3537–3539.
Wise, M. J. 2003. LEAping to conclusions: A computational reanalysis of late embryogenesis abundant proteins and their possible roles. BMC Bioinformatics 4: 52–70.[CrossRef][Medline]
Wise, M. J., AND A. Tunnacliffe. 2004. POPP the question: What do LEA proteins do? Trends in Plant Science 9: 13–17.[CrossRef][Web of Science][Medline]
Wood, A. J. 2007. The nature and distribution of vegetative desiccation-tolerance in hornworts, liverworts and mosses. Bryologist 110: 163–177.[CrossRef]
Wood, A. J., R. J. Duff, AND M. J. Oliver. 2000. The translational apparatus of Tortula ruralis: Polysomal retention of transcripts encoding the ribosomal proteins RPS14, RPS16 and RPL23 in desiccated and rehydrated gametophytes. Journal of Experimental Botany 51: 1655–1662.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
2 are in boldface type.