Molecular Biology and Evolution

MBE Advance Access originally published online on August 22, 2006
Molecular Biology and Evolution 2006 23(11):2245-2258; doi:10.1093/molbev/msl095

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 23/11/2245    most recent msl095v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Hileman, L. C. Articles by Irish, V. F. Search for Related Content PubMed PubMed Citation Articles by Hileman, L. C. Articles by Irish, V. F. Social Bookmarking          What's this?

© The Author 2006. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org

Research Articles

Molecular and Phylogenetic Analyses of the MADS-Box Gene Family in Tomato

Lena C. Hileman^*,1, Jens F. Sundstrom^*,2, Amy Litt^*,3, Meiqin Chen^*,4, Takudzwa Shumba^* and Vivian F. Irish^*,

^* Department of Molecular, Cellular and Developmental Biology, Yale University
Department of Ecology and Evolutionary Biology, Yale University

E-mail: vivian.irish{at}yale.edu.

	Abstract

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
MIKC^c-type MADS-box genes encode key transcriptional regulators of a variety of developmental processes in Arabidopsis thaliana. However, there has been relatively little effort to systematically carry out comparative genomic or functional analyses of these genes across flowering plants. Here we describe a strategy to identify members of the MIKC^c-type MADS-box gene family from any angiosperm species of interest. Using this approach, we have identified 24 MIKC^c-type MADS-box genes in tomato, including 17 that have not previously been characterized. Using these sequences, we have performed phylogenetic analyses that indicate that there have been a number of gene duplication and loss events in tomato relative to Arabidopsis. We also describe the expression domains of these genes and compare these results with their cognates in Arabidopsis. These analyses demonstrate the utility of this approach for characterizing a large number of MIKC^c-type MADS-box genes from any flowering plant species of interest and provide a framework for evolutionary comparisons of this important gene family across angiosperms.

Key Words: tomato • Solanaceae • homeotic gene • MADS-box gene • gene duplication

	Introduction

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
Evolution of the MIKC^c subfamily of MADS-box genes likely played a central role in the diversification of flowering plants (Becker and Theissen 2003; Irish 2003). Recent studies have focused on the evolution and diversification of specific MIKC^c-type MADS-box gene lineages (Kramer et al. 1998, 2004; Litt and Irish 2003; Stellari et al. 2004; Zahn et al. 2005). These studies have contributed significantly to our understanding of MADS-box gene family diversification, but comprehensive approaches to study the entire MIKC^c-type MADS-box gene family across flowering plants remain in their infancy (Baum et al. 2002; Soltis et al. 2002).

MADS-box genes encode DNA-binding transcription factors and have been identified in the genomes of plants, animals, and fungi (Messenguy and Dubois 2003). Before the divergence of plants from fungi and animals, a duplication occurred in the MADS-box lineage, resulting in type I and type II (MIKC-type) MADS-box genes (Alvarez-Buylla, Pelaz, et al. 2000; Svensson et al. 2000). Type I and type II MADS-box genes in animals and fungi are responsible for transcriptional regulation of a variety of processes, including the regulation of growth and development (reviewed in Messenguy and Dubois 2003). In plants, only very limited analyses of the type I MADS-box genes have been carried out to date (De Bodt et al. 2003; Kohler et al. 2003; Nam et al. 2004; Kohler et al. 2005). On the other hand, the plant type II (MIKC-type) MADS-box genes have been the subject of extensive investigation. Before the radiation of land plants, a duplication in the plant MIKC-type MADS-box gene lineage gave rise to the MIKC* (MADS in Parenicova et al. 2003) and MIKC^c-type MADS-box genes (Henschel et al. 2002; Kaufmann et al. 2005). It is the MIKC^c-type MADS-box genes that have diversified extensively in land plants, with 39 paralogous genes present in the Arabidopsis thaliana (Arabidopsis) genome (Kofuji et al. 2003; Parenicova et al. 2003) and an estimated 47 copies in the Oryza sativa (rice) genome (Nam et al. 2004). Functional diversification of the MIKC^c-type MADS domain proteins has resulted in their critical roles during many aspects of flowering plant development, including the transition from vegetative to reproductive development, establishment of floral organ identity, and differentiation of roots, leaves, fruits, and ovules (Becker and Theissen 2003).

MIKC-type MADS-box genes encode transcription factors that are characterized by their distinctive M–I–K–C protein structure. The MADS domain (M) is the most conserved region of the protein (the only region that is shared by all MADS-box genes) and is composed of a 58–60 amino acid motif at or near the N-terminus of the protein that forms a novel alpha-helical DNA-binding structure (Pellegrini et al. 1995; Shore and Sharrocks 1995; Tan and Richmond 1998; Santelli and Richmond 2000; Mo et al. 2001). The I domain exhibits low levels of amino acid sequence conservation but is predicted to form an alpha helix that influences DNA-binding specificity and dimerization (Riechmann et al. 1996). The K domain is structurally conserved among MIKC-type MADS-box genes and consists of regularly arranged hydrophobic residues that are predicted to form amphipathic helices required for protein–protein interactions (Ma et al. 1991; Pnueli et al. 1991; Pellegrini et al. 1995; Riechmann et al. 1996). The C-terminal domain of MIKC-type MADS-box genes is not highly conserved but can contain specific motifs including amino acid sequences important for transcriptional activation, posttranslational modification, or protein–protein interaction (Fan et al. 1997; Kramer et al. 1998; Cho et al. 1999; Kramer and Irish 1999; Yalovsky et al. 2000; Honma and Goto 2001; Pelaz et al. 2001; Lamb and Irish 2003; Litt and Irish 2003; Vandenbussche et al. 2003).

From work in Arabidopsis, it is becoming clear that many of the MIKC^c-type MADS-box genes and their protein products function as part of a tightly integrated genetic network. These genes act in auto- and cross-regulatory circuits (Jack 2004), their protein products are thought to function as components of larger multimeric complexes (Davies et al. 1996; Krizek and Meyerowitz 1996; Riechmann et al. 1996; Riechmann and Meyerowitz 1997; Egea-Cortines et al. 1999; Krizek et al. 1999; Pelaz et al. 2000; Honma and Goto 2001; Favaro et al. 2003), and the coordinate expression and function of these genes in a particular developmental time and place is necessary for specifying discrete aspects of development. However, the extent that these networks are conserved and act to specify analogous developmental outcomes across angiosperms is far from clear. Furthermore, it is apparent that orthologous MADS-box genes do not necessarily have analogous developmental roles (Zachgo et al. 1997; Davies et al. 1999; Kramer et al. 2004; Causier et al. 2005); conversely there are examples of paralogous MADS-box genes that appear to have taken on equivalent developmental functions (Martienssen and Irish 1999; Irish 2003; Causier et al. 2005). From these data, it is clear that assessment of orthology in this gene lineage is not adequate to establish functional similarity (Irish and Litt 2005; Moore and Purugganan 2005).

Establishing a robust gene tree for the MIKC^c-type MADS-box genes from representative angiosperms is key to understanding the evolution of the regulatory network in which their gene products function. Poor sampling from gene families may lead to incorrect assessment of orthology, failure to obtain robust estimates of the relationships among genes, poor inference of character evolution, and difficulty in assessing rates of molecular evolutionary change, including gene duplication and gene loss events. Therefore, the full compliment from multiple flowering plant genomes and relationships among MIKC^c-type MADS-box genes is critical for assessing functional and network diversification. In addition, by generating robust phylogenies, patterns of molecular evolution can be better correlated with functional data.

In order to specifically sample the diversity of MIKC^c-type MADS-box genes from a target species, we have developed a strategy for isolating these genes from any angiosperm species of interest. Using this strategy, we have isolated 24 MADS-box genes from Solanum lycopersicon (tomato; formerly Lycopersicon esculentum), 17 of which have not previously been characterized, and 3 of which would not have been identified from tomato expressed sequence tags (EST) data sets. Characterization of the expression patterns of these tomato genes and comparisons with cognate expression domains from Arabidopsis have pinpointed where evolutionary shifts in expression, and likely function, of these genes have occurred. Tomato is an excellent taxon for comparative work because it is distantly related to Arabidopsis, it is amenable to genetic and transgenic analyses, and significant genomic resources are available. As such, the studies presented here establish a foundation for comparative studies of the genetic networks in which MIKC^c-type MADS-box gene products function to control aspects of plant development.

	Methods

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
QVT Primer Design
In order to develop a set of degenerate forward primers that could be used with an oligo-dT primer to isolate MIKC^c-type MADS-box genes from cDNA, we first determined the most conserved region of the MIKC^c-type MADS domain across angiosperms. Using BlastP (Altschul et al. 1997) in September 2002, we sequentially searched the protein sequence databases against the 60 amino acid MADS domain from each of the 39 Arabidopsis MIKC^c-type MADS-box proteins (Parenicova et al. 2003), keeping sequences with a minimum similarity (E value) of 1 x 10^–10. We imposed an expect value of 10, word size of 3, the blosum62 matrix, an insert gap cost of –11, and an extend gap cost of –1, and we limited our search to "angiospermae." The MADS domains from the resulting sequences were aligned and scanned by eye to determine the most conserved short stretch of amino acid residues. The region we identified spans 13 amino acids and was named the QVT region after the first 3 amino acids in the sequence (glutamine, valine, and threonine; fig. 1).

View larger version (14K): [in this window] [in a new window] [Download PowerPoint slide]   FIG. 1.— Primers used to amplify angiosperm-type MIKCc-type MADS-box genes. See also supplementary table 1, Supplementary Material online. Letters in boxes represent amino acids found at positions in the QVT region using standard code. Letters below boxes show the sequence of degenerate QVT primers 1–7. Inosine was incorporated at any position designated "N."

 
Seven degenerate forward primers (QVT1–7) were designed against the QVT region (fig. 1) by back translating from the amino acid sequence to all possible nucleotide sequences using the standard genetic code. Sequences for the QVT primers are QVT1: 5'-CARGTNACNTTYNSNAARMGNMGNNNNGGNYTNYTNAA-3', QVT2: 5'-CARGTNACNTWYWSNAARMGNMRNWNNGGNATHYTNAA-3', QVT3: 5'-CARGTNMSNTWYWSNAARMGNMGNNSNGGNNTNWTNAA-3', QVT4: 5'-CARGTNACNTWYWSNRARMGNMGNRNNGGNYTNATNAA-3', QVT5: 5'-CARGTNACNTWYWSNAARMGNNWNAANGGNATNATNAA-3', QVT6: 5'-CARGTNACNTWYWSNAARMGNMGNRSNGGNYTNGTNAA-3', and QVT7: 5'-CARGTNACNTWYWSNAARMGNMGNAANGGNYTNATNGA-3'. To optimize primer design, we grouped the QVT sequences by codon similarity across the C-terminal end (fig. 1, supplementary table 1, Supplementary Material online). We allowed for 2-fold degenerate sites, but at 3- and 4-fold degenerate sites (designated "N" above and in fig. 1) inosine was incorporated into the primer oligonucleotide sequence. We allowed for up to 2 nt mismatches between the primers and the back-translations (supplementary table 1, Supplementary Material online), although in 3 cases additional mismatches were required (supplementary table 1, Supplementary Material online). Rice (AAB71822 [GenBank] ) contains 3 mismatches from QVT1, although there is not enough data from this sequence to determine if it is an MIKC-type MADS-box gene. Arabidopsis (AAN52807 [GenBank] ) has 4 mismatches from QVT4; this accession corresponds to AGL63, a MIKC^c-type MADS-box gene. Arabidopsis (AAD51984 [GenBank] ) has 4 mismatches from QVT6; this accession corresponds to PISTILLATA (PI). PI and AGL63 in Arabidopsis appear to be quite divergent from other QVT sequences. To determine how robust the QVT1–7 primers are to the addition of novel MADS-box genes to GenBank, we repeated the similarity search (above) on 5 September 2005 and determined the similarity of the QVT regions in GenBank at this time to the QVT1–7 primers (supplementary table 1, Supplementary Material online).

View this table: [in this window] [in a new window]   Table 1 Recovery of Arabidopsis thaliana MADS-Box Genes Using Different QVT Primers

 
Utility of QVT Primers in Arabidopsis
To determine if the full compliment of Arabidopsis MIKC^c-type MADS-box genes could be amplified using QVT1–7 primers, each of the primers was used in combination with an oligo-dT primer (5'-CAGTCGAGTCGACATCGA(T)¹⁷V-3') to generate polymerase chain reaction (PCR) products from cDNA derived from various Arabidopsis tissue types. RNA was extracted from Arabidopsis roots, seedlings, rosette leaves, cauline leaves, early flowers (approximately stages 1–9), late flowers (approximately stages 10–14), siliques, and seeds using Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. cDNA was generated from total RNA using Superscript III (Invitrogen) according to the manufacturer's instructions. PCR products using each of the 7 forward QVT (1–7) primers with an oligo-dT reverse primer were generated from the cDNAs using 40 cycles of PCR and a gradient of annealing temperatures between 45 °C and 55 °C. PCR products were separated on a 1% low melt agarose gel, and bands between 900 bp and 1 kb were excised. Aliquots of the excised bands were used for an additional round of PCR with Arabidopsis MIKC^c–specific primers (supplementary table 2, Supplementary Material online). The resulting PCR products were cleaned using the QIAquick PCR Purification Kit (QIAGEN, Valencia, CA) and sequenced directly, or cloned into TOPO-TA 4.0 (Invitrogen) according to the manufacturer's instructions, and 10 clones per PCR product were sequenced with M13F(-20) and M13R primers.

View this table: [in this window] [in a new window]   Table 2 Recovery of Arabidopsis thaliana MADS-Box Genes from Various Tissues

 
Isolation of MADS-Box Genes from Tomato Using QVT Primers
To isolate MIKC^c-type MADS-box genes from tomato, each of the QVT1–7 primers in combination with an oligo-dT primer was used to amplify PCR products from cDNA isolated from various tomato tissues. RNA was extracted from tomato roots, seedlings (including young leaves), inflorescences bearing early to late-stage flowers, sepals, petals, stamens, carpels (sepals, petals, stamens, and carpels were dissected from early- to late-stage flowers), green fruit, breaker fruit, yellow fruit, and ripe (red) fruit (including seeds) using Trizol according to the manufacturer's instructions. cDNA was generated from total RNA using Superscript III according to the manufacturer's instructions. PCR products using each of the 7 forward QVT (1–7) primers with an oligo-dT reverse primer were generated from cDNA using 35–40 cycles of PCR and a gradient of annealing temperatures between 45 °C and 55 °C. PCR products were separated on a 1% low melt agarose gel, and bands between 900 bp and 1 kb were excised. Excised PCR products were cloned into TOPO-TA 4.0 (Invitrogen) using the in-gel ligation protocol provided by the manufacturer.

As an initial survey for MADS-box genes, 48 PCR clones from each successful amplification for a given tissue type/QVT primer combination were sequenced using M13F(-20) and M13R primers, and distinct MADS-box genes were identified. For each amplification, approximately 95% of clones contained MADS-box sequences. Based on the recovered MADS-box genes from the initial sequencing survey, gene-specific primers were designed for use in subsequent rounds of screening (supplementary table 3, Supplementary Material online). Subsequent rounds of PCR clone screening were performed in order to isolate MADS-box genes that might be present in the sets of cloned PCR products at lower frequency. Subsequent PCR screening was performed using 96-well, high-throughput format. For each set of cloned PCR products, between 96 and 384 colonies were PCR screened with the gene-specific primers corresponding to those MADS-box genes that had been identified in that set of cloned PCR product during the initial survey. PCR clones that did not screen as positive for one of the initially surveyed MADS-box genes were sequenced, and additional, distinct MADS-box genes were identified. Confirmation of the sequence of distinct MADS-box genes was based on a minimum recovery of at least 10 independently isolated and sequenced clones.

View this table: [in this window] [in a new window]   Table 3 Complete List of Known Tomato MADS-Box Genes

 
Identification of Tomato ESTs
To identify sequenced tomato MIKC^c-type MADS-box genes from EST projects, the tomato gene index EST database at The Institute for Genomic Research (TIGR) was sequentially queried in September 2005, using BlastN (Altschul et al. 1997) with the 147-bp region encoding the MADS domain for all 39 Arabidopsis MIKC^c-type MADS-box genes (table 1). We retained sequences with a minimum similarity (E value) of 1 x 10^–9, imposing an expect value of 10, and setting all other options to the TIGR database default values. This EST database query revealed sequenced tomato MIKC^c-type MADS-box genes containing a full-length or nearly full-length MADS box. The identification of these ESTs is presented in table 4. To identify sequenced tomato MIKC^c-type MADS-box genes from EST projects that do not contain a full-length MADS box but that correspond to MADS-box genes recovered in this study, we again queried the TIGR tomato EST database using BlastN. We queried the EST database with the full-length MADS-box gene sequences recovered in our study that had not been identified in the tomato EST database by querying with the MADS-box region from Arabidopsis MIKC^c-type MADS-box genes. Only exact or nearly exact matches were considered to represent the same tomato MADS-box gene. The identification of the additionally recovered ESTs is presented in table 4.

View this table: [in this window] [in a new window]   Table 4 Recovery of Tomato MADS-Box Transcripts from Various Tissue Types

 
Phylogenetic Analyses
The matrix for phylogenetic analysis included the 39 MIKC^c-type MADS-box genes from Arabidopsis (table 1) and 36 MADS-box genes from tomato (table 4). Four MIKC*-type MADS-box genes from Arabidopsis (AGL30: At2g03060, AGL65: At1g18750, AGL66: At1g77980, and AGL67: At1g77950) were included in the matrix and used to root the MIKC^c-type MADS-box genes. Nucleotide sequences from the MADS-box and K domains were aligned manually with reference to the corresponding amino acid alignment using MacClade 4.0 (Maddison WP and Maddison DR 2000). The alignment is presented in supplementary figure 1, Supplementary Material online. The MADS + K nucleotide matrix was used for phylogenetic analysis under parsimony and Bayesian criteria.

Parsimony analysis was conducted using PAUP* 4.0 (Swofford 2002). Maximum parsimony trees were generated with heuristic searches (100 random stepwise taxon additions and Tree Bisection-Reconnection branch-swapping algorithm) with gaps treated as missing data. Bootstrap support for nodes (Felsenstein 1985) was estimated with 1,000 heuristic search replicates using the same settings as the original search, but only one random stepwise addition for each bootstrap replicate. Bayesian analysis was performed using Metropolis-coupled Markov chain Monte Carlo methods as implemented in MrBayes V. 3.1.1 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). The chain was run for 5 million generations without enforcing a molecular clock. A user-defined starting tree with 4 random rearrangements was implemented. This tree corresponded to one of the most parsimonious trees. The chain was sampled every 100 generations for a total of 50,000 trees sampled from the posterior distribution of trees and used to calculate posterior probabilities for clades (clade credibility values; CVs). The first 12,500 trees were discarded as "burn-in" (trees generated before likelihood values reached stationary). The general time reversible model with rates specified as gamma distributed across sites was used with priors set to default values. Four chains were run, with 1 chain heated at the default setting of 0.2.

Expression of Tomato MIKC^c-Type MADS-Box Genes
To define the general pattern of mRNA expression for the 36 tomato MIKC^c-type MADS-box genes, we performed gene- and tissue-specific reverse transcriptase (RT)–PCR. RNA was extracted from tomato roots, seedlings (with 2–3 true leaves), leaves, inflorescences, sepals, petals, stamens, and carpels (sepals, petals, stamens, and carpels were dissected from early- to late-stage flowers), green fruit, breaker fruit, yellow fruit, and ripe (red) fruit (including seeds) using Trizol according to the manufacturer's instructions. cDNA was generated from total RNA using Superscript III according to the manufacturer's instructions. Gene-specific primers were designed for each of the 36 tomato MIKC^c-type MADS-box genes (supplementary table 3, Supplementary Material online), and "ACTIN" (Prasad et al. 2001) was used as a positive control for each tissue type. The optimal annealing temperature and linear range of amplification was determined for each primer pair. Based on these determinations, PCR was performed with 24–32 cycles at an annealing temperature of 55–60 °C.

	Results

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
QVT Primer Design
Degenerate forward primers (QVT1–7) were designed by identifying a highly conserved stretch of 13 amino acids in the MADS box of MIKC^c-type MADS-box genes from across angiosperms. This region is dubbed the QVT region after the first 3 amino acids in the sequence (glutamine, valine, and threonine; fig. 1). The QVT region was identified by comparing the amino acid sequence of the MADS domain from 308 full-length angiosperm MADS domain sequences present in GenBank in September 2002. The back translation of groups of similar QVT sequences resulted in a total of 7 degenerate QVT primers (fig. 1) that can be used with an oligo-dT primer to amplify MIKC^c-type MADS-box genes.

Except in 3 cases (see methods), the original 308 sequences used to design the QVT primers do not differ by more than 2 nt from at least one of the QVT primer sequences (supplemental table 1, Supplementary Material online). To determine how robust our primer design is to the addition of new MADS-box genes, we repeated the search used to develop the QVT primers using all MADS-box sequences present in GenBank in September 2005. At this time, our search resulted in an additional 568 angiosperm MADS-box genes, for a total of 876. Comparison of the QVT region from these 568 new MADS-box genes with our QVT primers revealed that, as in the original design, nearly all sequences differed from one of the QVT primers by no more than 2 nt, and the majority of sequences corresponded exactly to at least one QVT primer sequence (supplemental table 1, Supplementary Material online). There are 5 exceptions to this. One rice accession (CAE05606 [GenBank] ) differs by 5 nt from QVT1. However, this accession contains a MADS box but no K domain, suggesting that it is not an MIKC-type MADS-box gene. One Arabidopsis accession (AAN52796 [GenBank] ) differs by 3 nt from QVT2. This accession corresponds to AGL30 which is a MIKC*-type MADS-box gene. One Asparagus officinalis accession (AAA18768 [GenBank] ) differs by 3 nt from QVT3. This accession does not contain enough sequence data to determine if there is a K domain in addition to the MADS box. One accession of Helianthus annuus (AAO18230 [GenBank] ) differs by 4 nt from QVT5. This accession does correspond to MIKC^c-type MADS-box genes. Lastly, in addition to the Arabidopsis PISTILLATA sequence recovered in the initial search (see Methods), there were 2 additional accessions of PISTILLATA from Arabidopsis lyrata and Brassica juncea (AAF25591 [GenBank] andAAY63867, respectively) that contain 4–5 nt differences from QVT6.

Proof of Concept: MIKC^c-Type MADS-Box Genes from Arabidopsis
To determine the efficiency of QVT1–7 primers, in combination with oligo-dT primers, for amplifying MIKC^c-type MADS-box genes, we first applied the degenerate primer approach to Arabidopsis. Based on previous work and the Arabidopsis genome sequence, there are 39 MIKC^c-type MADS-box genes in the Arabidopsis genome (table 1, Parenicova et al. 2003; Kaufmann et al. 2005). Nearly all Arabidopsis MIKC^c-type MADS-box genes are expected to amplify with the QVT1 primer. One gene, AGL24, is expected to amplify with QVT2. Four genes correspond exactly to QVT3. Five Arabidopsis MIKC^c-type MADS-box genes are expected to amplify with QVT4 (TT16, AGL17, AGL18, AGL21, and AGL63), none with QVT5, 2 (PI and AGL79) with QVT6, and 6 (FLM, FLC, and MAF2-5) with QVT7 (supplementary table 2, Supplementary Material online).

Thirty-seven of 39 Arabidopsis MIKC^c-type MADS-box genes were recovered with our QVT primers (tables 1 and 2). AGL13 and AGL71 were not recovered. AGL71 has a very restricted pattern of expression. Its expression has only been detected in germinating seedlings (Ma et al. 2005), which may explain our inability to recover it given our tissue-sampling strategy. AGL13, on the other hand, is not known to have a pattern of expression that might limit its successful amplification under our strategy (Rounsley et al. 1995; Parenicova et al. 2003), and neither does the QVT region sequence from AGL13 differ in any significant way from our QVT primers (supplementary table 2, Supplementary Material online). In fact, reconstruction experiments using QVT primers to amplify from a plasmid containing a cloned AGL13 sequence were successful, indicating that QVT primers are effective at recovering the AGL13 gene as well.

The QVT primers were each successful in amplifying more MIKC^c-type MADS-box genes than overall similarity would predict (table 1 and supplementary table 2, Supplementary Material online). The observed broader utility of QVT primers in amplifying given MIKC^c-type MADS-box genes is likely due to the high degree of nucleotide sequence similarity among the QVT primers (fig. 1).

In most cases, we were able to recover MIKC^c-type MADS-box genes from the predicted Arabidopsis tissue types (table 2). Parenicova et al. (2003) reported on Arabidopsis MADS-box gene expression profiles from roots, leaves, inflorescences, and siliques (fruit). We sampled more broadly and at a finer developmental scale (table 2), but by combining the results of our sampling from rosette and cauline leaves, and from early and late flowers, we can compare our MADS-box gene recovery data with expression profiles from leaves and inflorescences reported in the Parenicova study. In general, Arabidopsis MIKC^c-type MADS-box genes were recovered from all the expected tissue types (table 2). There were, however, a few exceptions, including AGL24, AG, SEP4, AGL18, AGL24, AGL63, and AGL72, that were recovered only in a subset of the predicted tissues (table 2). In most cases, though, the Arabidopsis MIKC^c-type MADS-box genes were recovered from additional tissues not reported in Parenicova et al. (2003) (table 2). These observations suggest that the methodologies we used may be more effective in identifying genes expressed at very low levels.

Isolation of MIKC^c-Type MADS-Box Genes from Tomato
A degenerate primer, RT–PCR approach was used to isolate novel MIKC^c-type MADS-box genes from tomato. The QVT1–7 primers, in combination with an oligo-dT reverse primer, were used to amplify MIKC^c-type MADS-box genes from cDNA derived from various tomato tissue types. These PCR products were cloned and sequenced and MIKC^c-type MADS-box genes were identified. On the basis of initial rounds of MADS-box gene screening, cloned PCR pools were subsequently screened in a high-throughput fashion for the presence of additional MIKC^c-type MADS-box genes potentially present at lower frequency. Using this approach, we recovered 24 tomato MIKC^c-type MADS-box genes (table 3), 7 of which have previously been reported in the literature, and 17 that are previously uncharacterized. A total of 12 tomato MIKC^c-type MADS-box genes are known, either from the literature or from EST collections, which were not recovered in this study. Together, there are 36 unique tomato MADS-box genes (table 3). This approaches the number that might be expected, given the 39 and 47 MIKC^c-type MADS-box genes in the Arabidopsis (Parenicova et al. 2003; Kaufmann et al. 2005) and rice (Nam et al. 2004) genomes, respectively.

Previously reported tomato MIKC^c-type MADS-box genes recovered in this study include TAG1, TAGL1, TAGL11, TAGL12 (Pnueli, Harevan, Broday, et al. 1994; Busi et al. 2003), TM4 (Pnueli et al. 1991; Busi et al. 2003), TM5 (Pnueli et al. 1991), and SlMBP7 (Litt and Irish 2003). SlMBP7 corresponds to LeFUL2 (Litt and Irish 2003), but contains a 1-bp indel in the C-terminal end of the coding sequence. The C-terminal amino acid sequence of SlMBP7 is highly similar to NsMADS1 from Nicotiana sylvestris (AF068725), and therefore, likely represents the actual C-terminal sequence for SlMBP7/LeFUL2. The additional 17 MIKC^c-type MADS-box genes have been given SlMBP numbers between 1 and 23 (S. lycopersicon MADS-box Protein; table 3). SlMBP23 is identical to TM3 at the 5' end of the gene but is highly divergent at the 3' end. Given that there is an EST available that is identical to SlMBP23, SlMBP23 likely does not represent a PCR artifact but may represent an alternatively spliced form or recent derivative of TM3.

Of the 17 previously undescribed tomato MIKC^c-type MADS-box genes recovered in this study, 12 were identified in the TIGR tomato EST database by searching for ESTs with similarity to the 39 MIKC^c-type MADS-box genes from Arabidopsis (table 3). Two additional tomato MIKC^c-type MADS-box gene ESTs were identified by this similarity search that were not recovered in our analysis, SlMBP24 and SlMBP25 (table 3). Because ESTs may lack full-length 5' sequence data and so lack the MADS domain, it is possible that the additional 5 novel genes recovered in this study may be present in the tomato EST database, but lack sufficient sequence necessary for their recognition as MADS-box genes. To determine if this is the case, the sequences we obtained for SlMBP9, SlMBP10, SlMBP12, SlMBP13, and SlMBP19 were compared with the TIGR tomato EST database. This database contains ESTs that correspond to 3' sequences of SlMBP9 and SlMBP13. We also recovered a truncated sequence corresponding to TAGL12 in this manner. Therefore, there are 3 tomato MADS-box genes—SlMBP10, SlMBP12, and SlMBP19—recovered in this analysis that have not been characterized in the literature nor have any corresponding sequences been recovered in the over 160,000 ESTs identified for tomato (http://www.tigr.org/tigr-scripts/tgi/T_index.cgi?species=tomato).

As we observed in Arabidopsis, multiple QVT primers were capable of amplifying the same tomato MIKC^c-type MADS-box gene (supplementary table 5, Supplementary Material online), likely due to the high degree of similarity among the QVT primers. Also similar to the results from Arabidopsis, tomato MIKC^c-type MADS-box genes were recovered from tissues where no expression was detected by semiquantitative RT–PCR (compare table 4 with fig. 2). For example, TM5, TM4, TAG1, SlMBP6, SlMBP7, SlMPB8, SlMPB11, and SlMBP18 were recovered from tomato root cDNA (table 4) but were not detected in root tissue by semiquantitative RT–PCR (fig. 2). Again, this is likely to be the result of relatively high numbers of PCR cycles used in the initial cloning procedure, resulting in the recovery of genes expressed at very low levels.

View larger version (30K): [in this window] [in a new window] [Download PowerPoint slide]   FIG. 2.— Semiquantitative RT–PCR expression of tomato MIKCc-type MADS-box genes in various tissues.

 
Phylogeny of Tomato and Arabidopsis MIKC^c-Type MADS-Box Genes
The 36 tomato MIKC^c-type MADS-box genes were aligned in a matrix with 39 Arabidopsis MIKC^c-type MADS-box genes and 4 Arabidopsis MIKC*-type MADS-box genes (supplementary fig. 1, Supplementary Material online). The MIKC* MADS-box genes were used to root the MIKC^c MADS-box gene tree. The matrix was reduced to just the MADS-box and K domain and used for phylogenetic analysis under maximum parsimony and Bayesian criteria. Maximum parsimony and Bayesian searches resulted in similar tree topologies. Bayesian analysis resulted in a tree with likelihood of –ln L = –24,417.91 (fig. 3). Parsimony analysis resulted in 5 most parsimonious trees of length 6,471 steps. The parsimony bootstrap consensus tree only differed in topology from the Bayesian tree at 1 position. In the parsimony bootstrap consensus tree, FUL is resolved as sister to AP1 + CAL with 58% support. This relationship remains unresolved in the Bayesian analysis. As expected, given previous phylogenetic work on the MADS-box gene family, there is not a one-to-one relationship of orthology between the tomato and Arabidopsis MADS-box genes. Multiple duplication and potential gene loss events have shaped the diversification of the MIKC^c MADS-box gene family (fig. 3).

View larger version (19K): [in this window] [in a new window] [Download PowerPoint slide]   FIG. 3.— Bayesian estimate of phylogenetic relationships among Arabidopsis and tomato MIKCc-type MADS-box genes. Numbers above branches indicate Bayesian clade CVs, and numbers below branches represent % parsimony bootstrap support. To the right of the gene tree, gene expression profiles for Arabidopsis from Parenicova et al. (2003) are compared with summarized profiles from tomato (fig. 2). Major clades are indicated to the right of the gene expression profiles. Asterisk: In the parsimony bootstrap consensus tree, this relationship is resolved with 58% support as FUL sister to AP1 + CAL. Genes in black from Arabidopsis, genes in gray from tomato. R, root; L, leaf; I, inflorescence; F, flower.

 
Eleven major lineages within the MIKC^c MADS-box gene tree can be defined with relatively strong clade CVs (Bayesian posterior probability support) and parsimony bootstrap percentages (BS). These include an AP3/PI clade with a CV of 100 and 59% BS support, an SVP clade with CV of 100 and 100% BS support, an ANR1 clade with a CV of 100 and 99% BS support, an AGL15 clade with a CV of 100 and 82% BS support, a TT16 clade that is not well supported with a BS percentage less than 50% and a CV of only 88, a SEP clade with a CV of 100 and 81% BS support, an AGL6 clade with a CV of 100 and 98% BS support, an AP1 clade with a CV of 100 and 99% BS support, a FLC/MAF clade with a CV of 100 and 92% BS support, a SOC1 clade with a CV of 100 and 98% BS support, and an AG clade with a CV of 100 and 100% BS support (fig. 3). The tomato MADS-box gene, TM8, does not fall within any of the above-defined clades (fig. 3). Its position at the base of the tree is only weakly supported (CV of 88 and 51% BS support); however, if this relationship holds, it suggests that TM8 may be more closely allied to the MIKC*-type MADS-box genes, than to the 11 defined clades of MIKC^c-type MADS-box genes.

Expression of MIKC^c-Type MADS-Box Genes in Tomato
Semiquantitative RT–PCR was used to determine the spatial patterns of tomato MIKC^c-type MADS-box mRNA expression. Several genes showed a broad pattern of expression, with mRNA detected in all, or nearly all, tissues including both vegetative and floral tissues. These included TM8, TAGL12, SlMBP15, SlMBP25, TM3, SlMBP23, SlMBP13, TM4, SlMBP7, SlMBP8, SlMBP10, SlMBP18, SlMBP19, SlMBP20, and SlMBP24. For several other genes, expression was restricted to vegetative tissues: SlMBP9 and SlMBP14 had various patterns of vegetative and/or root expression. For most of the tomato MIKC^c-type MADS-box genes, expression was limited to reproductive tissues. These included TM29, SlMBP21, SlMBP6, TM5, TM6, LePI, LeAP3, MADS-MC, TAGL1, TAG1, TAGL11, SlMBP11, SlMBP12, SlMBP22, SlMBP3, MADS1, and MADS-RIN. JOINTLESS expression was not detected (fig. 2).

	Discussion

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
A major obstacle to understanding how gene family diversification contributes to phenotypic evolution is the adequate assessment of gene orthology and paralogy (Irish and Litt 2005; Moore and Purugganan 2005). Assessment of gene orthology can be achieved through phylogenetic analysis of genes isolated from genome-sequencing projects, extensive EST projects, or by directly isolating members of a gene family of interest. Although the first 2 approaches are powerful methods, it is unlikely that all genomes of interest will be studied so intensely in the near future. Therefore, we need rapid and convenient methods to recover members of a gene family for robust comparative studies.

We have developed a reiterative procedure, based on a set of degenerate primers targeted against a wide array of MADS-box sequences, which allows for high-throughput cloning of these important developmental regulators in the absence of EST or genomic-level data. The QVT primers were designed based on alignment of all available angiosperm MIKC^c-type MADS-box gene sequences, in order to avoid any species-specific bias (supplementary table 1, Supplementary Material online). One major advantage of this approach is that MIKC^c-type MADS-box genes can be obtained from any angiosperm species of interest and potentially from a variety of nonangiosperm species. We carried out a reconstruction experiment in Arabidopsis, in order to demonstrate a proof of principle that this strategy would be effective. Furthermore, 80% of MIKC^c-type MADS-box gene sequences in Arabidopsis were recovered using just the QVT1 primer (table 1), illustrating that a simplified version of our strategy can be extremely effective in recovering the majority of these genes from a target species.

We also used this strategy in a test species, tomato, where there is considerable EST and genomic data available, yet full genome sequence is not yet available. Using this degenerate RT–PCR method, we recovered a number of previously unknown MADS-box genes from tomato. Placing these genes in a phylogenetic framework allows us to begin making comparative assessments about how diversification of MIKC^c-type MADS-box genes has contributed to plant phenotypic evolution. These comparisons also highlight the fact that there have been multiple gene duplication events that have taken place after the diversification of the lineages leading to tomato or Arabidopsis, resulting in numerous "in-paralogs" (Sonnhammer and Koonin 2002).

Phylogenetic and Functional Classification of Tomato MADS-Box Genes
The phylogenetic analyses and expression studies that we have carried out can serve as predictors of possible gene function in tomato. Similarities or differences in the expression patterns of tomato and Arabidopsis orthologs can point to conservation or diversification of gene function. In fact, there is a surprising degree of diversification in expression patterns among orthologs (fig. 3). In several cases, functional analyses have already been carried out for tomato MADS-box genes, and a comparison of these genes with their Arabidopsis counterparts is instructive in assessing the degree to which orthology corresponds to gene function.

Functional analyses have been carried out for SEP clade genes in both Arabidopsis and tomato, as well as in several other species. The tomato TM5, TM29, MADS-RIN, MADS1, and SlMBP21 genes are well-supported members of the SEP clade, although there is not a one-to-one orthology correspondence of tomato genes with those of Arabidopsis. Functional analyses of TM5 support the idea that this gene has a role similar to that of the Arabidopsis SEP genes in specifying floral organ identities (Pnueli, Harevan, Broday, et al. 1994). However, in addition to a role in specifying floral organ identities, the tomato TM29 gene appears to have an additional role in the maintenance of floral meristem identity (Ampomah-Dwamena et al. 2002). This is reminiscent of the proposed role of GRCD2, a SEP gene in Gerbera hybrida, that has been shown to have roles not only in floral development but also in the maintenance of inflorescence architecture and meristem identity (Uimari et al. 2004). Furthermore, the Arabidopsis SEP4 gene has also been shown to have roles in both floral organ and floral meristem identity (Ditta et al. 2004). Together, SlMBP21, MADS1, and MADS-RIN form a weakly supported clade within the SEP lineage, which in turn suggests that these genes arose from duplication events that postdate the diversification of Arabidopsis and tomato (fig. 3). A frameshift mutation in MADS-RIN affects tomato fruit ripening, but it is difficult to determine if this reflects a loss of function or novel role of the altered protein product (Vrebalov et al. 2002). It is difficult to suggest roles for MADS1 and SlMBP21 during tomato development given that there do not appear to be close orthologs of these genes in Arabidopsis (fig. 3). Based on their observed expression in flowers and fruit, one might suggest that SlMBP21 and MADS1 are involved in aspects of tomato flower and/or fruit development. One possibility is that they retain SEP-like functionality and interact with other MADS-box genes to direct floral organ identity (Jack 2001). In general, then, the SEP clade genes appear to have conserved roles in floral organ identity specification, but different gene duplicates in different species may have taken on other, or altered, roles in development.

The AP3/PI clade consists of 2 paralogs from Arabidopsis (AP3 and PI) and 4 from tomato (LeAP3, TM6, LePI, and LePI-B). Based on their observed patterns of expression, mainly restricted to petals and stamens (fig. 2), it is likely that all of the tomato AP3/PI genes, like AP3 and PI in Arabidopsis (Jack et al. 1992; Goto and Meyerowitz 1994), are involved in establishing petal and stamen identity during flower development. There has been an apparent loss of the TM6 ortholog along the lineage leading to Arabidopsis (Lamb and Irish 2003), and either a similar loss of a PI-like gene in Arabidopsis, or a duplication in the PI lineage leading to tomato. Genetic analyses of LeAP3 and TM6 indicate that these tomato genes have acquired distinct roles in petal and stamen development, underscoring the importance of gene duplication in diversification of developmental functions (de Martino et al. 2006).

The Arabidopsis SVP and AGL24 genes are involved in the transition to flowering (Hartmann et al. 2000; Yu et al. 2002; Michaels et al. 2003). JOINTLESS, the tomato ortholog of SVP is, by contrast, required to specify the pedicel abscission zone (Mao et al. 2000). We could not detect expression of JOINTLESS in our studies, likely due to it either being expressed in only a few cells or at low levels. As such, it appears that the functions of SVP and JOINTLESS have diversified considerably. However, we have found that SlMBP24, a well-supported member of the SVP clade (fig. 3) is expressed in vegetative and inflorescence tissue but not during later stages of flower and fruit development (fig. 2). Therefore, SlMBP24 appears to be a good candidate for flowering-time regulation in tomato.

The well-supported AG clade includes 4 Arabidopsis genes (AG, SHP1, SHP2, and STK) involved in carpel identity, floral determinacy, and aspects of fruit development (Favaro et al. 2003). The tomato MADS-box genes that fall into this clade are TAG1, TAGL1, TAGL11, and SlMBP3. Functional analyses of TAG1 suggest that it has a similar role to Arabidopsis AG (Pnueli, Haraven, Rounsley, et al. 1994). It is likely though, that in general there has been significant diversification in the roles of the tomato AG clade genes in comparison to their Arabidopsis counterparts in regulating carpel and fruit development, given the high level of morphological divergence between Arabidopsis and tomato fruit.

Predictions of Tomato MADS-Box Gene Function
We can predict likely functions of other tomato MADS-box genes based on the phylogenetic and expression analyses we have carried out; these data also provide a reference point for comparisons with MIKC^c genes from other plant species. SlMBP9 and SlMBP12 are close paralogs of one another and are strongly supported members of the ANR1 clade (fig. 3). ANR1, AGL17, and AGL21 are involved in root-cell patterning during Arabidopsis development (Zhang and Forde 1998; Burgeff et al. 2002). AGL16 seems to have evolved a novel role in guard-cell and trichome patterning (Alvarez-Buylla, Liljegren, et al. 2000). Similar to ANR1, AGL21, and AGL17, SlMBP9 and SlMBP12 are expressed in roots, suggesting that they may retain a similar role in root development as their Arabidopsis counterparts. However, in addition to root expression, SlMBP12 expression is detected during early-stage carpel development, suggesting that SlMPB12 may have been recruited for a novel role in carpel and/or ovule development.

SlMBP11 is a well-supported member of the AGL15 clade (fig. 3). AGL15 and AGL18 are involved in Arabidopsis embryogenesis (Alvarez-Buylla, Liljegren, et al. 2000). SlMBP11 expression is detected in stamens and weakly in early-stage carpels; therefore, SlMBP11 may be involved in embryo development like the closely related Arabidopsis gene but is also a good candidate for playing a more general role during the development of reproductive organs, including stamens, in tomato.

Although the TT16 clade lacks robust support, SlMBP22 is fairly well supported as the ortholog to the Arabidopsis genes TT16 and AGL63 within the TT16 lineage. TT16 is involved in the specification of the seed coat during Arabidopsis ovule development (Nesi et al. 2002). No specific role has been identified for AGL63. Given that SlMBP22 expression is highly restricted to early-stage carpel development (fig. 2), SlMBP22 is a good candidate for having a role in ovule development.

Within the well-supported SOC1 clade, there are 5 newly described tomato MADS-box genes (SlMBP23, SlMBP13, SlMBP14, SlMBP18, and SlMBP19). SlMPB23 is a recent paralog of TM3 and may be the result of alternative splicing (see Results). Together, TM3 and SlMBP23 are orthologous to SOC1, which promotes flowering in Arabidopsis (Moon et al. 2003). Expression of TM3 was previously reported to be restricted to vegetative tissue (Pnueli et al. 1991). However, we find that TM3 and SlMBP23 are expressed in vegetative and floral tissues (fig. 2). Either TM3 or SlMBP23 (or both) may have similar developmental roles as SOC1, regulating the transition to flowering in tomato. SlMBP13 and SlMBP14 are orthologous to AGL14 and AGL19 from Arabidopsis. AGL14 and AGL19 function in root development (Rounsley et al. 1995; Alvarez-Buylla, Liljegren et al. 2000). SlMBP13 and SlMBP14 are expressed in tomato roots and may function during their development, but the broader expression of SlMPB13 and SlMBP14 suggests that these genes may have evolved additional novel roles during vegetative and floral development. SlMPB18 and SlMBP19 are orthologous to an Arabidopsis clade consisting of AGL42, AGL71, and AGL72. Of these, only AGL42 has been characterized and has been shown to play a role in root development (Nawy et al. 2005). Given that expression of SlMBP18 and SlMBP19 was not detected in tomato roots but mainly in tomato floral organs, SlMBP18 and SlMBP19 likely have diversified in terms of function as compared with their cognates in Arabidopsis.

Origins of MIKC^c Subclades
Comparisons among fully sequenced plant genomes provide a good estimate of the numbers and complexity of the MIKC^c-type MADS-box genes across angiosperms. Complete or draft genome sequence information is now available for 3 Rosids, A. thaliana (Initiative 2000), Medicago truncatula (Cannon et al. 2005), and Populus trichocarpa (available through the Department of Energy Joint Genome Institute at http://genome.jgi-psf.org/Poptr1/Poptr1.home.html) as well as for a monocot, O. sativa (Project 2005; Yu et al. 2005). Through our analyses, we have identified an extensive set of MIKC^c-type MADS-box genes from tomato, an Asterid. Most of the subclades of genes that we have identified through phylogenetic analyses (including the AP3/PI, SVP, ANR1, AGL15, TT16, SEP, AGL6, AP1, SOC1, and AG subclades) are also present in the genomes of the sequenced species (Nam et al. 2004; Hecht et al. 2005; De Bodt et al. 2006) (L Hileman and VF Irish, unpublished data). However, there are several exceptions. Members of the AGL15 and TT16 clades have not yet been identified in Medicago (Hecht et al. 2005). However, we have identified members of both of these gene clades in tomato and they are present in Arabidopsis (fig. 3), suggesting that either these genes have been lost fairly recently from the lineage leading to Medicago or they have not been recovered due to incomplete genome annotation. These observations suggest that although gene duplication and loss have impacted the representation of these genes in different species, most MIKC^c-type MADS-box gene families likely predate the diversification of the angiosperms.

One dramatic exception appears to be the FLC/MAF subfamily. There seems to have been a recent radiation of FLC/MAF genes specific to the lineage leading to Arabidopsis (fig. 3, Becker and Theissen 2003; Hecht et al. 2005), and this radiation has resulted in a set of genes that differentially regulate the transition from vegetative to reproductive development in Arabidopsis (Boss et al. 2004). FLC transcription is downregulated in response to vernalization, which in turn induces flowering (Michaels and Amasino 1999). Given that the molecular mechanisms controlling the response to vernalization appear to have diversified (Sung and Amasino 2005), as well as the recent radiation of the FLC/MAF genes apparently within the Brassicaceae, it is difficult to extrapolate from Arabidopsis to species outside the Brassicaceae. We have defined an FLC/MAF-related clade of tomato genes (SlMBP8, SlMBP15, and SlMBP25) that may also be the result of recent gene duplication events in the lineage leading to tomato. These genes are broadly expressed in vegetative and reproductive tissues, and so may have roles in the vernalization response or other, completely distinct developmental processes.

The apparent loss of orthologs from a given species could have occurred by the degeneration of a gene to a pseudogene, or by a deletion event, or potentially by epigenetic silencing. Because we have assayed for the presence of MIKC^c-type MADS-box genes in tomato through cloning expressed genes, we cannot distinguish among these different types of "loss." The completion of the tomato genome project, which will provide genomic sequence, can help to ascertain the basis by which orthologs have been apparently lost in tomato.

The phylogenetic framework presented here contributes to the growing field of comparative developmental genetics in angiosperms. We have used the most conserved regions of MADS-box genes, the MADS-box and K domain, to generate a broad-level phylogenetic hypothesis for relationships among Arabidopsis and tomato MIKC^c-type MADS-box genes. Therefore, generating these estimates of relationship required ignoring much of the phylogenetic information that is present in comparisons among closely related MADS-box genes (i.e., the I and C-terminal domains). As detailed characterization of tomato MIKC^c-type MADS-box genes is undertaken, full-length sequences from within any of the well-defined clades described here will be an important consideration for more accurate assessments of gene orthology.

	Supplementary Material

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
Supplementary tables 1–5 and figure 1 are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).

	Acknowledgements

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 
We thank Gemma de Martino for the initial identification of LePI and LePI-B, and we thank members of the Irish lab for their constructive comments during the course of this work. We appreciate the help of Matthew Kost in compiling supplementary table 1, Supplementary Material online. We thank Dr. Michael Donoghue for his advice during the course of this work. We also would like to thank Drs. James Giovannoni and Julia Vrebalov for sharing unpublished information. This project was supported by grant #DBI-0110115 from the National Science Foundation to V.F.I., a Yale University Brown Fellowship to L.C.H., a fellowship from the Hellmuth Hertz Foundation to J.F.S., and a Yale University STARS undergraduate fellowship to T.S.

	Footnotes

 
¹ Present address: Department of Ecology and Evolutionary Biology, University of Kansas

² Present address: Dept of Plant Biology and Forest Genetics, Swedish University of Agricultural Sciences, Uppsala, Sweden

³ Present address: New York Botanic Garden, Bronx, New York

⁴ Present address: Department of Plant Developmental and Molecular Biology, College of Life Sciences, Peking University, Beijing, People's Republic of China

Douglas Crawford, Associate Editor

	References

TOP Abstract Introduction Methods Results Discussion Supplementary Material Acknowledgements References

 

Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, Lipman DJ. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389–402.[Abstract/Free Full Text]

Alvarez-Buylla ER, Liljegren SJ, Pelaz S, Gold SE, Burgeff C, Ditta GS, Vergara-Silva F, Yanofsky MF. (2000) MADS-box gene evolution beyond flowers: expression in pollen, endosperm, guard cells, roots and trichomes. Plant J 24:457–66.[CrossRef][Web of Science][Medline]

Alvarez-Buylla ER, Pelaz S, Liljegren SJ, Gold SE, Burgeff C, Ditta GS, de Pouplana LR, Martinez-Castilla L, Yanofsky MF. (2000) An ancestral MADS-box gene duplication occurred before the divergence of plants and animals. Proc Natl Acad Sci USA 97:5328–33.[Abstract/Free Full Text]

Ampomah-Dwamena C, Morris BA, Sutherland P, Veit B, Yao JL. (2002) Down-regulation of TM29, a tomato SEPALLATA homolog, causes parthenocarpic fruit development and floral reversion. Plant Physiol 130:605–17.[Abstract/Free Full Text]

Baum DA, Doebley J, Irish VF, Kramer EM. (2002) Response: missing links: the genetic architecture of flower and floral diversification. Trends Plant Sci 7:31–4.[Medline]

Becker A and Theissen G. (2003) The major clades of MADS-box genes and their role in the development and evolution of flowering plants. Mol Phylogenet Evol 29:464–89.[CrossRef][Web of Science][Medline]

Boss PK, Bastow RM, Mylne JS, Dean C. (2004) Multiple pathways in the decision to flower: enabling, promoting, and resetting. Plant Cell 16:S18–31.[Free Full Text]

Burgeff C, Liljegren SJ, Tapia-Lopez R, Yanofsky MF, Alvarez-Buylla ER. (2002) MADS-box gene expression in lateral primordia, meristems and differentiated tissues of Arabidopsis thaliana roots. Planta 214:365–72.[CrossRef][Web of Science][Medline]

Busi MV, Bustamante C, D'Angelo C, Hidalgo-Cuevas M, Boggio SB, Valle EM, Zabaleta E. (2003) MADS-box genes expressed during tomato seed and fruit development. Plant Mol Biol 52:801–15.[CrossRef][Web of Science][Medline]

Cannon SB, Crow JA, Heuer ML, et al. (22 co-authors). (2005) Databases and information integration for the Medicago truncatula genome and transcriptome. Plant Physiol 138:38–46.[Abstract/Free Full Text]

Causier B, Castillo R, Zhou JL, Ingram R, Xue YB, Schwarz-Sommer Z, Davies B. (2005) Evolution in action: following function in duplicated floral homeotic genes. Curr Biol 15:1508–12.[CrossRef][Web of Science][Medline]

Cho SC, Jang SH, Chae SJ, Chung KM, Moon YH, An GH, Jang SK. (1999) Analysis of the C-terminal region of Arabidopsis thaliana APETALA1 as a transcription activation domain. Plant Mol Biol 40:419–29.[CrossRef][Web of Science][Medline]

Davies B, EgeaCortines M, Silva ED, Saedler H, Sommer H. (1996) Multiple interactions amongst floral homeotic MADS box proteins. EMBO J 15:4330–43.[Web of Science][Medline]

Davies B, Motte P, Keck E, Saedler H, Sommer H, Schwarz-Sommer Z. (1999) PLENA and FARINELLI: redundancy and regulatory interactions between two Antirrhinum MADS-box factors controlling flower development. EMBO J 18:4023–34.[CrossRef][Web of Science][Medline]

De Bodt S, Raes J, Van de Peer YV, Theissen G. (2003) And then there were many: MADS goes genomic. Trends Plant Sci 8:475–83.[CrossRef][Web of Science][Medline]

De Bodt S, Theissen G, Van de Peer Y. (2006) Promoter analysis of MADS-box genes in Eudicots through phylogenetic footprinting. Mol Biol Evol 23:1293–1303.[Abstract/Free Full Text]

de Martino G, Pan I, Emmanuel E, Levy AA, Irish VF. (2006) Functional analyses of two tomato APETALA3 genes demonstrate diversification in their roles in regulating flowering. Plant Cell 18:1833–1845.[Abstract/Free Full Text]

Ditta G, Pinyopich A, Robles P, Pelaz S, Yanofsky MF. (2004) The SEP4 gene of Arabidopsis thaliana functions in floral organ and meristem identity. Curr Biol 14:1935–40.[CrossRef][Web of Science][Medline]

Egea-Cortines M, Saedler H, Sommer H. (1999) Ternary complex formation between the MADS-box proteins SQUAMOSA, DEFICIENS and GLOBOSA is involved in the control of floral architecture in Antirrhinum majus. EMBO J 18:5370–9.[CrossRef][Web of Science][Medline]

Fan HY, Hu Y, Tudor M, Ma H. (1997) Specific interactions between the K domains of AG and AGLs, members of the MADS domain family of DNA binding proteins. Plant J 12:999–1010.[CrossRef][Web of Science][Medline]

Favaro R, Pinyopich A, Battaglia R, Kooiker M, Borghi L, Ditta G, Yanofsky MF, Kater MM, Colombo L. (2003) MADS-box protein complexes control carpel and ovule development in Arabidopsis. Plant Cell 15:2603–11.[Abstract/Free Full Text]

Felsenstein J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783–91.[CrossRef][Web of Science]

Goto K and Meyerowitz EM. (1994) Function and regulation of the Arabidopsis floral homeotic gene pistillata. Genes Dev 8:1548–60.[Abstract/Free Full Text]

Hartmann U, Hohmann S, Nettesheim K, Wisman E, Saedler H, Huijser P. (2000) Molecular cloning of SVP: a negative regulator of the floral transition in Arabidopsis. Plant J 21:351–60.[CrossRef][Web of Science][Medline]

Hecht V, Foucher F, Ferrandiz C, et al. (11 co-authors). (2005) Conservation of Arabidopsis flowering genes in model legumes. Plant Physiol 137:1420–34.[Abstract/Free Full Text]

Henschel K, Kofuji R, Hasebe M, Saedler H, Munster T, Theissen G. (2002) Two ancient classes of MIKC-type MADS-box genes are present in the moss Physcomitrella patens. Mol Biol Evol 19:801–14.[Abstract/Free Full Text]

Honma T and Goto K. (2001) Complexes of MADS-box proteins are sufficient to convert leaves into floral organs. Nature 409:525–9.[CrossRef][Medline]

Huelsenbeck JP and Ronquist F. (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17:754–5.[Abstract/Free Full Text]

Initiative AG. (2000) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408:796–815.[CrossRef][Medline]

Irish VF. (2003) The evolution of floral homeotic gene function. Bioessays 25:637–46.[CrossRef][Web of Science][Medline]

Irish VF and Litt A. (2005) Flower development and evolution: gene duplication, diversification and redeployment. Curr Opin Genet Dev 15:454–60.[CrossRef][Web of Science][Medline]

Jack T. (2001) Relearning our ABCs: new twists on an old model. Trends Plant Sci 6:310–6.[CrossRef][Web of Science][Medline]

Jack T. (2004) Molecular and genetic mechanisms of floral control. Plant Cell 16:S1–17.[Free Full Text]

Jack T, Brockman LL, Meyerowitz EM. (1992) The homeotic gene APETALA3 of Arabidopsis thaliana encodes a MADS box and is expressed in petals and stamens. Cell 68:683–97.[CrossRef][Web of Science][Medline]

Kaufmann K, Melzer R, Theissen G. (2005) MIKC-type MADS-domain proteins: structural modularity, protein interactions and network evolution in land plants. Gene 347:183–98.[CrossRef][Web of Science][Medline]

Kofuji R, Sumikawa N, Yamasaki M, Kondo K, Ueda K, Ito M, Hasebe M. (2003) Evolution and divergence of the MADS-box gene family based on genome-wide expression analyses. Mol Biol Evol 20:1963–77.[Abstract/Free Full Text]

Kohler C, Hennig L, Spillane C, Pien S, Gruissem W, Grossniklaus U. (2003) The polycomb-group protein MEDEA regulates seed development by controlling expression of the MADS-box gene PHERES1. Genes Dev 17:1540–53.[Abstract/Free Full Text]

Kohler C, Page DR, Gagliardini V, Grossniklaus U. (2005) The Arabidopsis thaliana MEDEA polycomb group protein controls expression of PHERES1 by parental imprinting. Nat Genet 37:28–30.[CrossRef][Web of Science][Medline]

Kramer EM, Dorit RL, Irish VF. (1998) Molecular evolution of genes controlling petal and stamen development: duplication and divergence within the APETALA3 and PISTILLATA MADS-box gene lineages. Genetics 149:765–83.[Abstract/Free Full Text]

Kramer EM and Irish VF. (1999) Evolution of genetic mechanisms controlling petal development. Nature 399:144–8.[CrossRef][Medline]

Kramer EM, Jaramillo MA, Di Stilio VS. (2004) Patterns of gene duplication and functional evolution during the diversification of the AGAMOUS subfamily of MADS box genes in angiosperms. Genetics 166:1011–23.[Abstract/Free Full Text]

Krizek BA and Meyerowitz EM. (1996) Mapping the protein regions responsible for the functional specificities of the Arabidopsis MADS domain organ-identity proteins. Proc Natl Acad Sci USA 93:4063–70.[Abstract/Free Full Text]

Krizek BA, Riechmann JL, Meyerowitz EM. (1999) Use of the APETALA1 promoter to assay the in vivo function of chimeric MADS box genes. Sex Plant Reprod 12:14–26.[CrossRef]

Lamb RS and Irish VF. (2003) Functional divergence within the APETALA3/PISTILLATA floral homeotic gene lineages. Proc Natl Acad Sci USA 100:6558–63.[Abstract/Free Full Text]

Litt A and Irish VF. (2003) Duplication and diversification in the APETALA1/FRUITFULL floral homeotic gene lineage: implications for the evolution of floral development. Genetics 165:821–33.[Abstract/Free Full Text]

Ma H, Yanofsky MF, Meyerowitz EM. (1991) Agl1-Agl6, an Arabidopsis gene family with similarity to floral homeotic and transcription factor genes. Genes Dev 5:484–95.[Abstract/Free Full Text]

Ma L, Sun N, Liu X, Jiao Y, Zhao H, Deng XW. (2005) Organ-specific expression of Arabidopsis genome during development. Plant Physiol 138:80–91.[Abstract/Free Full Text]

Maddison WP and Maddison DR. (2000) MacClade: analysis of phylogeny and character evolution. Version 4.0. (Sinauer Associates, Sunderland, MA).

Mao L, Begum D, Chuang HW, Budiman MA, Szymkowiak EJ, Irish EE, Wing RA. (2000) JOINTLESS is a MADS-box gene controlling tomato flower abscission zone development. Nature 406:910–3.[CrossRef][Medline]

Martienssen R and Irish V. (1999) Copying out our ABCs—the role of gene redundancy in interpreting genetic hierarchies. Trends Genet 15:435–7.[CrossRef][Web of Science][Medline]

Messenguy F and Dubois E. (2003) Role of MADS box proteins and their cofactors in combinatorial control of gene expression and cell development. Gene 316:1–21.[CrossRef][Web of Science][Medline]

Michaels SD and Amasino RM. (1999) FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 11:949–56.[Abstract/Free Full Text]

Michaels SD, Ditta G, Gustafson-Brown C, Pelaz S, Yanofsky MF, Amasino RM. (2003) AGL24 acts as a promoter of flowering in Arabidopsis and is positively regulated by vernalization. Plant J 33:867–74.[CrossRef][Web of Science][Medline]

Mo Y, Ho W, Johnston K, Marmorstein R. (2001) Crystal structure of a ternary SAP-1/SRF/c-fos SIRE DNA complex. J Mol Biol 314:495–506.[CrossRef][Web of Science][Medline]

Moon J, Suh SS, Lee H, Choi KR, Hong CB, Paek NC, Kim SG, Lee I. (2003) The SOC1 MADS-box gene integrates vernalization and gibberellin signals for flowering in Arabidopsis. Plant J 35:613–23.[CrossRef][Web of Science][Medline]

Moore RC and Purugganan MD. (2005) The evolutionary dynamics of plant duplicate genes. Curr Opin Plant Biol 8:122–8.[CrossRef][Web of Science][Medline]

Nam J, Kim J, Lee S, An GH, Ma H, Nei MS. (2004) Type I MADS-box genes have experienced faster birth-and-death evolution than type II MADS-box genes in angiosperms. Proc Natl Acad Sci USA 101:1910–5.[Abstract/Free Full Text]

Nawy T, Lee JY, Colinas J, Wang JY, Thongrod SC, Malamy JE, Birnbaum K, Benfey PN. (2005) Transcriptional profile of the Arabidopsis root quiescent center. Plant Cell 17:1908–25.[Abstract/Free Full Text]

Nesi N, Debeaujon I, Jond C, Stewart AJ, Jenkins GI, Caboche M, Lepiniec L. (2002) The TRANSPARENT TESTA16 locus encodes the Arabidopsis B sister MADS domain protein and is required for proper development and pigmentation of the seed coat. Plant Cell 14:2463–79.[Abstract/Free Full Text]

Parenicova L, de Folter S, Kieffer M, et al. (12 co-authors). (2003) Molecular and phylogenetic analyses of the complete MADS-box transcription factor family in Arabidopsis: new openings to the MADS world. Plant Cell 15:1538–51.[Abstract/Free Full Text]

Pelaz S, Ditta GS, Baumann E, Wisman E, Yanofsky MF. (2000) B and C floral organ identity functions require SEPALLATA MADS-box genes. Nature 405:200–3.[CrossRef][Medline]

Pelaz S, Gustafson-Brown C, Kohalmi SE, Crosby WL, Yanofsky MF. (2001) APETALA1 and SEPALLATA3 interact to promote flower development. Plant J 26:385–94.[CrossRef][Web of Science][Medline]

Pellegrini L, Song T, Richmond TJ. (1995) Structure of serum response factor core bound to DNA. Nature 376:490–8.[CrossRef][Medline]

Pnueli L, Abuabeid M, Zamir D, Nacken W, Schwarzsommer Z, Lifschitz E. (1991) The Mads box gene family in tomato—temporal expression during floral development, conserved secondary structures and homology with homeotic genes from antirrhinum and Arabidopsis. Plant J 1:255–66.[Web of Science][Medline]

Pnueli L, Hareven D, Broday L, Hurwitz C, Lifschitz E. (1994) The Tm5 Mads box gene mediates organ differentiation in the 3 inner whorls of tomato flowers. Plant Cell 6:175–86.[Abstract]

Pnueli L, Hareven D, Rounsley SD, Yanofsky MF, Lifschitz E. (1994) Isolation of the tomato AGAMOUS gene TAG1 and analysis of its homeotic role in transgenic plants. Plant Cell 6:163–73.[Abstract]

Prasad K, Sriram P, Kumar CS, Kushalappa K, Vijayraghavan U. (2001) Ectopic expression of rice OsMADS1 reveals a role in specifying the lemma and palea, grass floral organs analogous to sepals. Dev Genes Evol 211:281–90.[CrossRef][Web of Science][Medline]

Project IRGS. (2005) The map-based sequence of the rice genome. Nature 436:793–800.[CrossRef][Medline]

Riechmann JL, Krizek BA, Meyerowitz EM. (1996) Dimerization specificity of Arabidopsis MADS domain homeotic proteins APETALA1, APETALA3, PISTILLATA, and AGAMOUS. Proc Natl Acad Sci USA 93:4793–8.[Abstract/Free Full Text]

Riechmann JL and Meyerowitz EM. (1997) Determination of floral organ identity by Arabidopsis MADS domain homeotic proteins AP1, AP3, PI, and AG is independent of their DNA-binding specificity. Mol Biol Cell 8:1243–59.[Abstract]

Ronquist F and Huelsenbeck JP. (2003) MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–4.[Abstract/Free Full Text]

Rounsley SD, Ditta GS, Yanofsky MF. (1995) Diverse roles for Mads box genes in Arabidopsis development. Plant Cell 7:1259–69.[Abstract]

Santelli E and Richmond TJ. (2000) Crystal structure of MEF2A core bound to DNA at 1.5 angstrom resolution. J Mol Biol 297:437–49.[CrossRef][Web of Science][Medline]

Shore P and Sharrocks D. (1995) The Mads-box family of transcription factors. Eur J Biochem 229:1–13.[Web of Science][Medline]

Soltis DE, Soltis PS, Albert VA, Oppenheimer DG, dePamphilis CW, Ma H, Frohlich MW, Theissen G. (2002) Missing links: the genetic architecture of flower and floral diversification. Trends Plant Sci 7:22–31.[CrossRef][Web of Science][Medline]

Sonnhammer ELL and Koonin EV. (2002) Orthology, paralogy and proposed classification for paralog subtypes. Trends Genet 18:619–20.[CrossRef][Web of Science][Medline]

Stellari GM, Jaramillo MA, Kramer EM. (2004) Evolution of the APETALA3 and PISTILLATA lineages of MADS-box-containing genes in the basal angiosperms. Mol Biol Evol 21:506–19.[Abstract/Free Full Text]

Sung S and Amasino RM. (2005) Remembering winter: toward a molecular understanding of vernalization. Annu Rev Plant Biol 56:491–508.[CrossRef][Medline]

Svensson ME, Johannesson H, Engstrom P. (2000) The LAMB1 gene from the clubmoss, Lycopodium annotinum, is a divergent MADS-box gene, expressed specifically in sporogenic structures. Gene 253:31–43.[CrossRef][Web of Science][Medline]

Swofford DL. (2002) PAUP* 4.0: phylogenetic analysis using parsimony (*and other methods)(Sinauer Associates, Sunderland, MA).

Tan S and Richmond TJ. (1998) Crystal structure of the yeast MAT alpha 2/MCM1/DNA ternary complex. Nature 391:660–6.[CrossRef][Medline]

Uimari A, Kotilainen M, Elomaa P, Yu D, Albert VA, Teeri TH. (2004) Integration of reproductive meristem fates by a SEPALLATA-like MADS-box gene. Proc Natl Acad Sci USA 101:15817–22.[Abstract/Free Full Text]

Vandenbussche M, Theissen G, Van de Peer Y, Gerats T. (2003) Structural diversification and neo-functionalization during floral MADS-box gene evolution by C-terminal frameshift mutations. Nucleic Acids Res 31:4401–9.[Abstract/Free Full Text]

Vrebalov J, Ruezinsky D, Padmanabhan V, White R, Medrano D, Drake R, Schuch W, Giovannoni J. (2002) A MADS-box gene necessary for fruit ripening at the tomato ripening-inhibitor (Rin) locus. Science 296:343–6.[Abstract/Free Full Text]

Yalovsky S, Rodriguez-Concepcion M, Bracha K, Toledo-Ortiz G, Gruissem W. (2000) Prenylation of the floral transcription factor APETALA1 modulates its function. Plant Cell 12:1257–66.[Abstract/Free Full Text]

Yu H, Xu Y, Tan EL, Kumar PP. (2002) AGAMOUS-LIKE 24, a dosage-dependent mediator of the flowering signals. Proc Natl Acad Sci USA 99:16336–41.[Abstract/Free Full Text]

Yu JJ, Wang W, Lin S, et al. (117 co-authors). (2005) The genomes of Oryza sativa: a history of duplications. PLoS Biol 3:e38.[CrossRef][Medline]

Zachgo S, Saedler H, SchwarzSommer Z. (1997) Pollen-specific expression of DEFH125, a MADS-box transcription factor in Antirrhinum with unusual features. Plant J 11:1043–50.[CrossRef][Web of Science][Medline]

Zahn LM, King HZ, Leebens-Mack JH, Kim S, Soltis PS, Landherr LL, Soltis DE, dePamphilis CW, Ma H. (2005) The evolution of the SEPALLATA subfamily of MADS-box genes: a preangiosperm origin with multiple duplications throughout angiosperm history. Genetics 169:2209–23.[Abstract/Free Full Text]

Zhang H and Forde BG. (1998) An Arabidopsis MADS box gene that controls nutrient-induced changes in root architecture. Science 279:407–9.[Abstract/Free Full Text]

Accepted for publication August 16, 2006.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				V. F. Irish Evolution of petal identity J. Exp. Bot., July 1, 2009; 60(9): 2517 - 2527. [Abstract] [Full Text] [PDF]

				Z. Li, G. L. Reighard, A. G. Abbott, and D. G. Bielenberg Dormancy-associated MADS genes from the EVG locus of peach [Prunus persica (L.) Batsch] have distinct seasonal and photoperiodic expression patterns J. Exp. Bot., June 24, 2009; (2009) erp195v1. [Abstract] [Full Text] [PDF]

				A. Tadiello, A. Pavanello, D. Zanin, E. Caporali, L. Colombo, G. L. Rotino, L. Trainotti, and G. Casadoro A PLENA-like gene of peach is involved in carpel formation and subsequent transformation into a fleshy fruit J. Exp. Bot., February 1, 2009; 60(2): 651 - 661. [Abstract] [Full Text] [PDF]

				C. H. Leseberg, C. L. Eissler, X. Wang, M. A. Johns, M. R. Duvall, and L. Mao Interaction study of MADS-domain proteins in tomato J. Exp. Bot., May 17, 2008; (2008) ern094v1. [Abstract] [Full Text] [PDF]

				D. E. Soltis, A. S. Chanderbali, S. Kim, M. Buzgo, and P. S. Soltis The ABC Model and its Applicability to Basal Angiosperms Ann. Bot., August 1, 2007; 100(2): 155 - 163. [Abstract] [Full Text] [PDF]

				C. He, H. Sommer, B. Grosardt, P. Huijser, and H. Saedler PFMAGO, a MAGO NASHI-Like Factor, Interacts with the MADS-Domain Protein MPF2 from Physalis floridana Mol. Biol. Evol., May 1, 2007; 24(5): 1229 - 1241. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 23/11/2245    most recent msl095v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Hileman, L. C. Articles by Irish, V. F. Search for Related Content PubMed PubMed Citation Articles by Hileman, L. C. Articles by Irish, V. F. Social Bookmarking          What's this?

Online ISSN 1537-1719 - Print ISSN 0737-4038

Copyright © 2009 Society for Molecular Biology and Evolution

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics