| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Genetics and Molecular Biology |
2Center for Ecology and Evolutionary Biology, University of Oregon, Eugene, Oregon 97403 USA; 3Biology Department, University of Washington, Seattle, Washington 98195 USA; 4Department of Biology, Duke University, Durham, North Carolina 27708 USA
Received for publication September 13, 2004. Accepted for publication May 25, 2005.
ABSTRACT
Duplication of some floral regulatory genes has occurred repeatedly in angiosperms, whereas others are thought to be single-copy in most lineages. We selected three genes that interact in a pathway regulating floral development conserved among higher tricolpates (LFY/FLO, UFO/FIM, and AP3/DEF) and screened for copy number among families of Lamiales that are closely related to the model species Antirrhinum majus. We show that two of three genes have duplicated at least twice in the Lamiales. Phylogenetic analyses of paralogs suggest that an ancient whole genome duplication shared among many families of Lamiales occurred after the ancestor of these families diverged from the lineage leading to Veronicaceae (including the single-copy species A. majus). Duplication is consistent with previous patterns among angiosperm lineages for AP3/DEF, but this is the first report of functional duplicate copies of LFY/FLO outside of tetraploid species. We propose Lamiales taxa will be good models for understanding mechanisms of duplicate gene preservation and how floral regulatory genes may contribute to morphological diversity.
Key Words: floral regulatory genes • gene duplication • Lamiales
Morphological innovations are thought to frequently involve divergence among duplicated genes that comprise key genetic pathways regulating development (Raff, 1996
; Carroll et al., 2001
). Though most evidence comes from comparisons among animals, a similar pattern is beginning to emerge for flowering plants based primarily on studies of the MADS box family of transcriptional regulators (reviewed in Irish, 2003
). Gene duplication has played a prominent role in the evolution of several families of MADS box genes. At least four distinct gene lineages predate land plants (>400 million years ago [mya]; reviewed in Theissen et al., 2000
), and subsequent duplication has occurred within the seed plants (>200 mya; Kramer et al., 1998
; reviewed in Theissen et al., 2000
) and repeatedly within angiosperms (reviewed in Kramer et al., 2003
). Floral homeotic genes belonging to the AP3/DEF clade of MADS box genes typify this trend. The ancestor of the orthologous genes AP3 (Jack et al., 1992
) and DEF (Sommer et al., 1990
), which are necessary for development of petals and stamens in the model species Arabidopsis thaliana and Antirrhinum majus, respectively, duplicated within seed plants resulting in the ancestor of the AP3/DEF lineage as well as a second paralogous gene lineage also characterized in A. thaliana and A. majus (PI/GLO; Sundstrom et al., 1999
; Aoki et al., 2004
; Stellari et al., 2004
). A later duplication of the ancestral AP3/DEF gene occurred near the base of the tricolpate (eudicot) lineage of angiosperms resulting in the TM6 lineage (Kramer et al., 1998
). Irish (2003)
suggests duplication of several MADS box genes near the base of the tricolpates, including the ancestral AP3/DEF gene, may have contributed to key innovations such as the development of distinct petals and the fixation of whorled phyllotaxy.
In contrast to the radiation of the MADS box family, other floral regulatory genes appear to be quite conservative in copy number among angiosperms. For example, the orthologous genes LFY (Weigel et al., 1992
) and FLO (Coen et al., 1990
) are transcription factors characterized by a novel DNA-binding domain sufficient for specifying floral meristem identity in A. thaliana and A. majus and are derived from one of two paralogous genes found in gymnosperms, apparently the result of an ancient duplication within seed plants (Frohlich and Parker, 2000
). LFY/FLO is thought to be a single copy gene in all extant angiosperms previously studied (Frohlich and Parker, 2000
; reviewed in Cronk, 2001
), with the notable exception of polyploid species (e.g., Nicotiana tobaccum, Kelly, 1995
; Zea mays, Bomblies et al., 2003
). Thus all LFY/FLO-like genes found in angiosperms appear to be orthologs of a single ancestral gene. Not surprisingly, given the apparent lack of duplication in the LFY/FLO gene lineage and its key position in the floral regulatory pathway (Parcy et al., 1998
; Blazquez and Weigel, 2000
), orthologs from Z. mays and tricolpates play similar roles in both meristem identity and regulating the expression of downstream genes (Bomblies et al., 2003
).
Despite markedly different histories of duplication for LFY/ FLO and AP3/DEF, these genes interact in a pathway essentially conserved among higher tricolpates (Ingram, 1995
). Distinct from its role in floral meristem identity (Parcy et al., 1998
), LFY/FLO positively regulates the expression of AP3/ DEF in part through binding an upstream regulatory sequence of AP3 (Lamb et al., 2002
). A second gene, for which cognate homologs are well characterized in A. thaliana (UFO; Ingram et al., 1997
) and A. majus (FIM; Simon et al., 1994
), mediates LFY/FLO activity on AP3/DEF via ubiquitin-mediated degradation of transcriptional repressors of AP3 (Zhao et al., 2001
). Because of the conserved regulatory roles of LFY/FLO, UFO/FIM, and AP3/DEF during floral development, it is interesting to know whether duplication of one or more of these genes is correlated with duplication of other genes in the pathway and if so, whether duplicate genes are retained jointly. Characterizing patterns of duplication is a first step toward the ultimate goal of understanding how the genes regulating floral development have contributed toward the diversification of floral morphology.
Previous work (Fishman et al., 2001
) identified evidence of duplication for other floral regulatory genes in Mimulus species, a genus traditionally classified with the model species A. majus in the family Scrophulariaceae (Lamiales). Olmstead et al.'s (2001) molecular systematic study of taxa from Scrophulariaceae sensu lato and representatives of 15 other families of Lamiales led to a dramatic reorganization of the group. Their study found strong support for a monophyletic clade within a reclassified Veronicaceae that includes A. majus and suggested reclassification of other traditional Scrophulariaceae taxa to families such as Orobanchaceae. Significantly, the phylogenetic placement of Mimulus relative to other Lamiales families remains unclear despite additional molecular systematic studies (Beardsley and Olmstead, 2002
). Given the dramatic reorganization of Scrophulariaceae (Olmstead et al., 2001
), we undertook the present study to: (1) screen taxa historically classified with A. majus in Scrophulariaceae and taxa from related families for copy number of other floral regulatory genes including LFY/FLO, UFO/FIM, and AP3/DEF; (2) clarify when gene duplication occurred relative to the divergence among several lineages of the order Lamiales and A. majus; and (3) describe the patterns of gene duplication and retention among lineages for all three genes. By focusing on LFY/FLO, UFO/FIM, and AP3/DEF together, we examine duplication of multiple interacting members comprising a portion of a signaling pathway previously studied only individually (LFY/ FLO, AP3/DEF) or for which little is known outside of model species (UFO/FIM).
MATERIALS AND METHODS
Identifying floral regulatory genes in Lamiales
We selected nine species in the order Lamiales based on classifications from recent phylogenetic studies (Olmstead et al., 2001
; Beardsley and Olmstead, 2002
; Fig. 1, Table 1) to bracket the phylogenetic placement of the model species Antirrhinum majus as well as Mimulus from which duplicates of floral regulatory genes were found previously (Fishman et al., 2001
). Selected taxa include Syringa vulgaris (Oleaceae), Chelone glabra and Antirrhinum majus (Veronicaceae), Verbena officinalis (Verbenaceae), Salvia coccinia (Lamiaceae), Mimulus guttatus and Mimulus lewisii (Phrymaceae), Paulownia tomentosa (Paulowniaceae), and Pedicularis groenlandica (Orobanchaceae; Table 1). Source material was obtained from the University of Washington medicinal herb gardens (Syringa vulgaris, Chelone glabra, Verbena officinalis, Salvia coccinia, and Paulownia tomentosa), Fishman et al. (2001
; Mimulus guttatus), Bradshaw et al. (1995
; Mimulus lewisii), John Innes Centre (Norfolk, UK; Antirrhinum majus) or was wild collected (Pedicularis groenlandica; J. Aagaard 2003–1, WTU). For each species, genomic DNA was prepared from leaf material of a single plant using the cetyltrimethylammonium bromide (CTAB) extraction method of Kelly and Willis (1998)
. Total RNA was prepared from whole flower buds of the same plants across a range of developmental stages using the RNeasy Plant Mini kit (Qiagen, Valencia, California, USA) and used to make 5'- and 3'-RACE-Ready cDNA pools (Clontech, Palo Alto, California, USA).
|
|
g genomic DNA, 50 mM KCl, 10 mM Tris pH 8.4, 2.5 mM MgCl2, 0.06% BSA, 0.2 mM dNTPs, 5 
mol of each primer, and 1 unit Promega Taq polymerase (Madison, Wisconsin, USA). Thermal cycler conditions included an initial denaturation step of 2 min at 94°C followed by 37 cycles of 20 s at 94°C, 20 s at 58°C, 90 s at 72°C, and a final extension at 72°C for 5 min. The PCR products were run on 2% Tris-borate-EDTA (TBE) gels, all visible bands excised and purified using the GENECLEAN kit (Qbiogene, Carlsbad, California, USA), and cloned using the TOPO TA cloning kit (Invitrogen, Carlsbad, California, USA). A minimum of four clones of each PCR fragment was sequenced using automated fluorescent sequencing methods (Applied Biosystems, Foster City, California, USA). (2) 3'-RACE primers were designed from sequence of all genomic clones of LFY/FLO-like sequences spanning the first intron, and 3'-RACE PCR carried out using the SMART RACE cDNA amplification kit (Clontech, Palo Alto, California, USA). 3'-RACE products were cloned directly and sequenced (four or more clones) as above for genomic PCR fragments. LFY/ FLO-like 3'-RACE sequences were compared with the coding region of genomic fragments spanning the second intron cloned earlier; for unique LFY/ FLO-like genomic sequences spanning intron 2, 5'- and 3'-RACE was used to extend coding sequence of clones as before for intron 1 genomic clones. Homologs of UFO/FIM were cloned from all nine taxa within Lamiales using a two-step approach similar to that used for LFY/FLO. (1) Two overlapping degenerate primer sets nested within conserved domains of the intronless UFO/FIM gene were designed from the consensus sequence of UFO and FIM (first 5'primer 5'-CGAGCTCGATCAGTCTGTAARMGNTGGTA-3', first 3'primer 5'-CGCTGAACGGGCTGTAGTTCATRCARTARAA-3'; second 5' primer 5'-CCGCCATTGGTTCCTCTTCTTYAARCARCA-3', second 3'primer 5'-TCTGCAAACTGATTGTACAGYTGYTGNGGCAT-3'). The PCR was performed on genomic DNA using both primer sets and an identical protocol as for LFY/FLO. The PCR fragments were purified, cloned, and at least four clones of each fragment sequenced as before. (2) Sequence of genomic clones of all UFO/FIM-like fragments from UFO/FIM primer sets 1 and 2 were compared and 3'-RACE primers designed for each unique sequence; 3'-RACE was carried out for LFY/FLO and four or more clones of each 3'-RACE product sequenced.
Lamiales homologs of AP3/DEF were cloned from all nine species directly from 3'-RACE cDNA pools. A single degenerate primer at the 5' terminus of the AP3/DEF-coding region was designed from the consensus sequence of AP3/DEF, Petunia hybrida PMADS1, and Solanum lycopersicon AP3 (Table 1; 5'-ATGGCTCGTGGGAAGATHCARAT-3'). The 3'-RACE products were cloned, and 12 or more clones of each fragment were sequenced as for LFY/ FLO and UFO/FIM.
Phylogenetic analyses of floral regulatory genes
Coding regions of all LFY/FLO- and AP3/DEF-like genes we identified were aligned with homologs from Zea mays, A. thaliana, Petunia hybrida, Solanum lycopersicon, and published sequences from A. majus (Table 1). We used only one of the paralogous LFY/FLO loci from Z. mays (ZFL1) that are believed to be the result of the tetraploid ancestry of Zea (Gaut and Doebley, 1997
). Coding regions of all UFO/FIM-like genes we cloned were aligned with homologs from A. thaliana and A. majus as well as several other UFO/FIM-like homologs. These include F-box genes from Pisum sativum (PisSaSTP, accession AF004843; Taylor et al., 2001
), Lotus japonica (LotJaPFO, accession AY156687; Zhang et al., 2003
), and Impatiens balsamina (ImpBaFIM, accession AF047392; Pouteau et al., 1998
), as well as a UFO/FIM-like expressed sequence tag (EST) from a Solanum lycopersicon floral bud library (LycEsEST, accession BI423409). The Solanum lycopersicon UFO/FIM-like EST was the highest scoring hit from BLAST searches of EST databases from Solonaceae taxa (Plant Genome Database, http://plantgdb.org) searched using A. majus FIM and all Lamiales FIM--like genes we cloned. Because published A. majus sequences for all three genes were identical to those we identified in our cloning experiments (see below) and included additional sequence 5' of ours, we excluded our A. majus sequences from further analyses. Nucleotide alignments were carried out initially based on the translated nucleotide (protein) sequence using the ClustalX algorithm implemented in BioEdit (T. Hall, North Carolina State University, Raleigh, North Carolina, USA), followed by visual alignment. Areas with ambiguous alignments were excluded, and the nucleotide alignment was analyzed using likelihood criterion implemented in PAUP* (Swofford, 2002
). Likelihood analyses employed the general time reversible model with four rate categories, estimating the gamma shape parameter and the proportion of invariable sites (GTR + 1 + 
). Heuristic search criterion included tree-bisection-reconnection (TBR) branch swapping with 10 random addition replicates. Support for nodes was estimated by 100 bootstrap replicates using the maximum likelihood estimates of substitution parameters and heuristic search criterion described previously.
Tests of alternate topologies
Likelihood-based tests of alternate topologies (Goldman, 1993
; Goldman et al., 2000
) were used to test different hypotheses regarding the timing of duplication of multiple LFY/FLO-like and AP3/DEF-like copies found in the Lamiales. Likelihood scores (–lnL) of topologies and maximum likelihood (ML) estimates of substitution parameters (GTR + 1 + ) were calculated under (1) one constraint topology for LFY/ FLO and (2) three constraint topologies for AP3/DEF using PAUP* (Swofford, 2002
). (1) The constraint topology for LFY/FLO (LFY/FLO H01) collapsed the nodes within paralogous gene clades with bootstrap support less than 95% and placed Antirrhinum majus FLO (AntMaFLO) along with Chelone glabra FLO (CheGlFLO) and Syringa vulgaris FLO (SyrVuFLO) at an unresolved node sister below two paralogous gene clades. We excluded Z. mays ZFL1 from the analyses because of problems with long branch attraction (see below). (2) Constraint 1 (AP3/DEF H01) collapsed nodes within the paralogous gene clades of the maximum likelihood tree having bootstrap support less than 95% and moved Syringa vulgaris DEF (SyrVuDEF) to the arbitrarily defined A clade. Constraint 2 (AP3/DEF H02) is identical to constraint 1 except SyrVuDEF is placed within the B clade. Constraint 3 (AP3/DEF H03) collapsed nodes in a similar manner and placed Chelone glabra (CheGlDEF) and Antirrhinum majus DEF (AntMaDEF) along with SyrVuDEF at an unresolved node sister below the paralogous gene clades (Fig. 3). The test statistic 
= (–lnL–ML)–(–lnLMLHo) was calculated for each constraint, where (–lnL–ML) is the likelihood score maximized over all topologies, and (–lnLMLHo) is the likelihood score given the topological constraint (Goldman et al., 2000
).
|
). Simulated data used the ML estimates of substitution parameters and topology (including branch lengths) under H0 (GTR + 1 + ), and were identical in sequence length to the original alignments used in phylogenetic analyses (LFY/FLO = 783 nucleotides, AP3/DEF = 666 nucleotides). Simulated data were analyzed according to Goldman et al. (2000)
using an approximation under HA (posPpud) in which substitution parameters (GTR + 1 + ) are provided when maximizing the likelihood over all topologies, and the test statistic = (–lnL–ML)–(–lnLMLHo) is calculated for each simulated data set. RESULTS
Copy number of floral regulatory genes among Lamiales
A single copy of a LFY/FLO-like gene was found for both Syringa vulgaris and Antirrhinum majus (Table 1). Overlapping sequence from 3'-RACE clones corresponding to genomic sequence amplified using first and second intron primer sets were identical. Similarly, a single copy of an AP3/DEF-like gene was found for both species based on our copy number assay using 3'-RACE directly. For both genes, the A. majus clones we identified were identical to previously published FLO (Coen et al., 1990
) and DEF (Sommer et al., 1990
) sequences. In contrast, two copies of both LFY/FLO-like and AP3/DEF-like genes were found for seven taxa representing five families within Lamiales: Chelone glabra (Veronicaceae), Verbena officinalis (Verbenaceae), Salvia coccinia (Lamiaceae), Paulownia tomentosa (Paulowniaceae), Pedicularis groenlandica (Orobanchaceae), and Mimulus lewisii and M. guttatus (Phrymaceae; Table 1). For LFY/FLO-like genes, the first intron primer set amplified both copies from most species except P. groenlandica, for which we found a single copy using this primer combination. The second intron primer set amplified two copies from V. officinalis and P. tomentosa but a single copy from all other species. Comparison of sequence from 3'-RACE experiments showed clones corresponding to genomic sequence from first and second intron primer sets were identical for at least one copy except for P. groenlandica. Thus first and second intron primer sets together identified two distinct LFY/FLO-like genes from P. groenlandica.
The single copy of LFY/FLO-like and AP3/DEF-like genes we found for S. vulgaris as well as duplicate copies found for V. officinalis, P. tomentosa, P. groenlandica, M. lewisii, and M. guttatus appear functional based on correctly spliced cDNAs. Duplicates are well diverged at synonymous codon positions, on average having 44% (LFY/FLO-like) and 38% (AP3/DEF-like) of silent sites with substitutions. Duplicate copies of LFY/FLO-like and AP3/DEF-like genes found for S. coccinia also appear functional based on cDNAs, but divergence at silent sites is approximately half that seen for other duplicates (19% and 17% for S. coccinia LFY/FLO-like and AP3/DEF-like genes, respectively). In contrast with other LFY/ FLO-like duplicates, one copy of both LFY/FLO-like and AP3/ DEF-like genes in C. glabra may be nonfunctional as evidenced by no detectable expression in cDNA pools (LFY/FLO-like) or premature stop codons found in cDNAs (AP3/DEF-like). Duplicates from C. glabra that appear nonfunctional (and were sister to the functional copies; see below) were excluded from further analyses.
A single copy of an UFO/FIM-like gene was found within all taxa we studied (Table 1). Sequence from genomic clones of both primer sets were identical within the region they overlapped (approximately 500 nucleotides) for all species, and the A. majus sequence we identified was identical to previously published FIM (Simon et al., 1994
). Therefore 3'-RACE using primers designed only from genomic clones of the 5'primer sets were used to obtain cDNA sequences for all taxa, with the exception of P. groenlandica. We found no evidence of expression in our floral bud cDNA pools from P. groenlandica, and additional 3'- as well as 5'-RACE experiments using multiple nested primer sets showed no evidence of expression in any of three cDNA pools constructed from individual P. groenlandica plants collected from separate populations. This suggests the UFO/FIM-like genomic sequence from P. groenlandica either (1) may not be expressed at detectable levels in the floral tissues we assayed or (2) may correspond to a nonfunctional copy, although there is no evidence of missense or stop codons (see below).
Phylogenetic analyses of floral regulatory genes
LFY/ FLO
After excluding ambiguous regions, our alignment of LFY/FLO homologs from 13 species includes approximately 783 nucleotides from each of 19 homologs (Table 1). Based on this alignment, likelihood criterion found a maximum likelihood tree (–lnL = 6520.4862) placing Zea mays ZFL1 sister to one of two S. coccinia LFY/FLO homologs. However, this topology appears to be the effect of the long branch leading to ZFL1 because an alternate tree of marginally lower score (–lnL = 6523.76183; Fig. 2) was found that has a topology identical to that of the maximum likelihood tree when ZFL1 is excluded from the analysis. This alternate likelihood tree (Fig. 2) places all LFY/FLO homologs from taxa within Lamiales in a well-supported clade (94% bootstrap support) sister to LFY/FLO homologs from the Solanales. Within Lamiales, the S. vulgaris homolog branches first, sister to a clade comprising the remaining sequences (71%), followed by the single expressed (see above) C. glabra homolog. When both C. glabra LFY/FLO-like genes are included, they form a well-supported clade (100% bootstrap support; data not shown). Two gene clades belong sister, each including one copy of the LFY/ FLO homologs from V. officinalis, P. tomentosa, P. groenlandica, M. lewisii, and M. guttatus. Antirrhinum majus FLO is sister to one of these gene clades (FLOB), though its placement is weakly supported (49% bootstrap support). The two LFY/FLO homologs from S. coccinia cluster together as a well-supported sister clade (99%) within the FLOA clade.
|
; Fig. 3). The alternate topology places A. majus FLO sister to the two paralogy groups in Lamiales (LFY/FLO H01). The maximum likelihood tree given this constraint is only slightly lower (–lnL = 5903.04329), although the test statistic comparing the constrained tree with that placing A. majus FLO within the FLO B clade ( = 1.8) is near the significance threshold (P = 0.03; Fig. 4). This tight distribution apparently reflects the very short branch leading to the node placing A. majus FLO sister to the FLOB clade (Fig. 2) and the high proportion of invariant sites (36%) in the model used to simulate data.
|
We investigated three alternate topologies for AP3/DEF homologs using the likelihood based approach of Goldman et al. (2000
; Fig. 3). The first places S. vulgaris AP3/DEF within the DEFA clade (AP3/DEF H01) along with A. majus DEF and C. glabra AP3/DEF. The maximum likelihood score given this constraint (–lnL = 6989.50125) is significantly lower than the unconstrained maximum likelihood tree, and the test statistic ( = 12.5) lies well beyond 95% of test statistics from data simulated under the constraint (Fig. 4). Similarly, an alternate topology placing S. vulgaris within the DEFB clade (AP3/DEF H02) is rejected ( = 11.7; Fig. 4). However, the third topology placing Veronicaceae homologs (A. majus DEF and C. glabra AP3/DEF-like) within a clade sister to paralogy groups DEFA and DEFB cannot be rejected, because the test statistic (= 1.7) lies near the 50th percentile of data simulated under the constraint topology of AP3/DEF H03 (Fig. 4).
UFO/FIM
Phylogenetic analyses of our alignment (approximately 1113 nucleotides, excluding all ambiguous regions) of UFO, FIM, UFO/FIM-like homologs or ESTs, and the eight UFO/FIM-like genes we identified among Lamiales taxa show generally weakly supported topologies based on likelihood criterion. Likelihood analyses found a maximum likelihood tree placing FIM and all UFO/FIM-like genes we cloned from Lamiales taxa in a weakly supported monophyletic clade sister to other Asterid UFO/FIM-like genes, including the UFO/FIM-like EST from S. esculentum (LycEsEST; Fig. 2). This is strong evidence the UFO/FIM-like genes we cloned from Lamiales taxa are orthologous to A. majus FIM. Within the Lamiales clade, the S. vulgaris UFO/FIM-like gene branches first. UFO/FIM-like homologs from the Veronicaceae including FIM form a weakly supported clade sister to the remaining Lamiales UFO/FIM-like homologs. Several of the nodes within this clade have moderate bootstrap support including Mimulus UFO/FIM-like genes (78%) and a clade with the strongest bootstrap support across the tree (79%) consisting of the S. coccinia UFO/FIM-like homolog and the gene from P. groenlandica, for which we found no evidence of expression within multiple floral bud cDNAs.
DISCUSSION
Multiple duplications of LFY/FLO and AP3/DEF in Lamiales
We have identified at least two independent duplications of LFY/FLO and AP3/DEF homologs within the Lamiales. The first occurred within the Veronicaceae (sensu Olmstead et al., 2001
), after the split between the lineage leading to Antirrhinum majus and Chelone glabra. This duplication is likely the result of a whole genome duplication (polyploidization) as evidenced by the approximate doubling of chromosomes in C. glabra (N = 14) relative to A. majus (N = 8; Goldblatt and Krukoff, 1984
). In this case, one duplicate of each gene appears to have been silenced based on no detectable expression in RACE cDNA pools (LFY/FLO) or premature stop codons (AP3/DEF). We also found two copies of both LFY/FLO-like and AP3/DEF-like genes in representative taxa from Verbenaceae (Verbena officinalis), Phrymaceae (sensu Beardsley and Olmstead, 2002
; Mimulus guttatus and M. lewisii), Paulowniaceae (Paulownia tomentosa), and Orobanchaceae (Pedicularis groenlandica; Fig. 1). Duplicate copies in these taxa are highly diverged at silent sites (44% and 38% for LFY/FLO-like and AP3/DEF-like, respectively), and phylogenetic analyses consistently place one copy from each species in separate gene clades (Fig. 2). This evidence is consistent with a second ancient duplication(s) in the ancestral lineage of these families, resulting in two paralogs that appear functional based on full-length cDNAs. Finally, two putative copies of LFY/FLO-like and AP3/DEF-like genes were found in Salvia coccinia (Lamiaceae) that cluster together with strong support within the FLOA and DEFB paralogy clades, respectively (Fig. 2). This may represent a third duplication event specific to the Lamiaceae, although there is no evidence of recent polyploidy as for Chelone glabra from chromosome numbers. However, we cannot rule out the possibility that duplicate copies in Lamiaceae could be of a hybrid origin (i.e., alleles of a single locus) because divergence at silent sites (19% and 17% for LFY/FLO-like and AP3/DEF-like, respectively) is near the range found for alleles segregating between interfertile Mimulus species (10% and 13%, respectively; J. E. Aagaard, unpublished data; University of Washington).
We examined several alternate topologies for gene trees (Fig. 2) that suggest different points for the ancient duplication(s) of LFY/FLO-like and AP3/DEF-like genes relative to divergence among families of Lamiales. First, we tested alternate topologies using a likelihood-based approach (Goldman et al., 2000
), suggesting duplication(s) occurred at the base of the Lamiales (AP3/DEF H01, H02; Fig. 3). In all cases, alternate topologies supporting this hypothesis were rejected (Fig. 4). Second, because the LFY/FLO tree places Veronicaceae homologs before (C. glabra) and after (A. majus) the duplication, we tested alternate topologies constituting a null hypothesis, which implies duplication after the split leading to the Veronicaceae (LFY/FLO H01; AP3/DEF H03; Fig. 3). In this case, we clearly cannot reject the null hypothesis for AP3/ DEF, though the test result is somewhat ambiguous for LFY/ FLO (Fig. 4). However, because there is no evidence of duplicate copies from past work in the well-studied model species A. majus and Antirrhinum and Chelone represent a well-supported clade (Olmstead et al., 2001
), this argues against rejecting the null hypothesis for LFY/FLO as well. In addition, Southern blots probed with the highly conserved C-terminus portion of FLO showed no evidence of more than a single copy in taxa from Veronicaceae (data not shown). In sum, our phylogenetic analyses and tests of alternate topologies support ancient duplication(s) of LFY/FLO-like and AP3/DEF-like genes after the divergence between Veronicaceae and the lineage leading to many of the other families of Lamiales including Verbenaceae, Lamiaceae, Phrymaceae, Paulowniaceae, and Orobanchaceae. We consider this scenario more likely because it is much more parsimonious from the standpoint of number of inferred gene duplication and loss events. Accordingly, we suggest a gene nomenclature for Lamiales LFY/FLO and AP3/DEF homologs that reflects the precedence of FLO (Coen et al., 1990
) relative to LFY (Weigel et al., 1992
), DEF (Sommer et al., 1990
) relative to AP3 (Jack et al., 1992
), and co-orthology with FLO and DEF. Below we refer to Lamiales homologs as FLO or DEF, with co-orthologs further delineated as A or B (see Table 1).
Duplication of both genes at the same relative position in the Lamiales phylogeny suggests an ancient whole genome duplication event in Lamiales. FLOA and DEFB lie on separate linkage groups in Mimulus, as do putative paralogs of another floral regulatory gene (CYC; Fishman et al., 2001
) that have a similar level of divergence at silent sites (45%) as FLOA/B and DEFA/B (J. E. Aagaard, unpublished data; University of Washington). Linkage relationships for FLO and DEF and additional unlinked paralogs are again consistent with whole genome duplication. However, taxa for which we found a single copy of FLO and DEF varied in their chromosome number (S. vulgaris, n = 22, A. majus, n = 8; Goldblatt and Krukoff, 1984
) similar to taxa from which two copies of FLO and DEF were found (V. officinalis, n = 7; S. coccinia, n = 11; M. lewisii, n = 8; M. guttatus, n = 14; P. tomentosa, n = 20; P. groenlandica, n = 8; Goldblatt and Krukoff, 1984
). Thus extant chromosome numbers appear to provide little evidence supporting whole genome duplication in the Lamiales phylogeny. As additional gene-based markers are mapped in Lamiales taxa sharing the hypothesized whole-genome duplication event (e.g., Mimulus; Bradshaw and Schemske, 1995
; Fishman et al., 2001
), syntenic groups can be compared with linkage relationships for A. majus (Schwarz-Sommer et al., 2003
) in order to clarify the source of the ancient FLO and DEF paralogs we found. Regardless, the clear presence of paralogs for multiple genes among many families of Lamiales suggests future studies in this clade carefully consider orthology/paralogy relationships among homologs.
Patterns of duplication in the AP3/DEF and LFY/FLO gene lineages
The DEF duplicates within Lamiales we identified complement a pattern of repeated duplications throughout the AP3/DEF lineage reported previously. Duplications have occurred in seed plants (Theissen et al., 2000
), within angiosperms at the base of the tricolpates (Kramer et al., 1998
), and more recently, among lineages of basal angiosperms (Stellari et al., 2004
) and other basal tricolpate families such as Ranunculaceae (Kramer et al., 2003
). Interestingly, Irish (2003)
suggests the duplication at the base of the tricolpates corresponds with duplication of several other MADS box genes, consistent with an ancient whole genome duplication at the base of the tricolpate lineage. This is similar to what we propose for Lamiales based on the concordant patterns for multiple floral regulatory genes including FLO and DEF. Duplication of MADS box genes at the base of the tricolpates including the ancestral AP3/DEF gene are thought to have contributed to major morphological innovations, including the development of distinct petals and the fixation of whorled phyllotaxy (Kramer and Irish, 1999
; Irish, 2003
). However, because of the time scale and phylogenetic position of most model species relative to this duplication, it is difficult to reconstruct ancestral gene function. Thus the contribution of earlier duplications to morphological evolution may prove difficult to study. In contrast, we suggest DEF duplicates in Lamiales may prove a good system in which to study the mechanisms responsible for duplicate gene preservation and diversification. Specifically, DEFA and DEFB are present within taxa from multiple Lamiales families separated for more than 30 million years (Wikstrom et al., 2001
), providing several independently evolving lineages for comparison with the (single copy outgroup) model species A. majus. In addition, because a whole genome duplication is the likely cause of duplication in Lamiales, we believe paralogs of other floral regulatory genes will likely be found, such as those that are known to dimerize with AP3/DEF, including PI/GLO (Trobner et al., 1992
; Goto and Meyerowitz, 1994
), or whose expression is regulated by LFY/FLO, including AG/PLENA (Yanofsky et al., 1990
; Bradley et al., 1993
).
Significantly, ours is the first report of duplicate functional copies of LFY/FLO homologs outside of tetraploids (e.g., Kelly et al., 1995
; Bomblies et al., 2003
). Cronk (2001)
suggested the conspicuous absence of LFY/FLO duplication across angiosperms reported previously (e.g., Frohlich and Parker, 2000
) was in sharp contrast to patterns for other loci such as MADS box genes and could be the result of selection against duplicate LFY/FLO copies. However, our results suggest duplicate copies of LFY/FLO might have simply gone undetected in previous studies due to insufficient sampling of species and genomes. We used two independent primer sets nested within conserved domains of exons based on the A. majus FLO sequence (an ingroup) to amplify over introns with known length polymorphism, followed by exhaustive cloning and sequencing of all detectable PCR fragments regardless of signal intensity. Despite our experimental design, in at least one case (Pedicularis groenlandica), individual primer sets consistently amplified only one or the other FLO paralog. It would be interesting to apply a similar screening approach to assay LFY/ FLO copy number in other angiosperm lineages, particularly where duplications of other genes regulated by LFY/FLO are known (e.g., AP3/DEF-like genes in Ranunculaceae; Kramer et al., 2003
). Such work will clarify issues regarding the frequency of duplication for LFY/FLO relative to other floral regulatory genes and may contribute to an understanding of the evolution of regulatory pathways following duplication (see below).
Copy number of UFO/FIM-like genes is inconclusive
Unlike LFY/FLO and AP3/DEF, our cloning experiments and phylogenetic analyses of UFO, FIM, and UFO/FIM-like homologs, including the eight UFO/FIM-like genes we identified among Lamiales taxa, provide little convincing evidence regarding copy number in Lamiales. Although our cloning experiments only identified a single UFO/FIM-like gene among all Lamiales taxa from which we cloned, our phylogenetic analyses are suggestive of a pattern of duplication for this gene. For example, it is surprising that support for nodes of the UFO/FIM phylogeny are weaker than the LFY/FLO tree since our UFO/FIM alignment is more than 40% longer than for LFY/FLO (1113 and 783 nucleotides, respectively), and the UFO/FIM alignment has fewer invariable sites than for LFY/FLO (20% and 31%, respectively). In addition, the relatively strong support (79%) for the clade containing S. coccinia and P. groenlandica UFO/FIM-like genes to the exclusion of other homologs is surprising, given we do not see this relationship among any other genes we have studied, and this clustering does not reflect the phylogenetic relationships among these taxa in Olmstead et al.'s (2001)
study (Fig. 1). Possible explanations include (1) duplicate copies of UFO/ FIM in the Lamiales, which we have incompletely sampled in our PCR-based screen; or (2) duplication of an ancestral UFO/ FIM homolog, followed by repeated loss along the lineages from which we sampled (lineage sorting). These explanations are intriguing, because we can find no evidence of expression for the P. groenlandica UFO/FIM-like gene we cloned from genomic DNA in any of four cDNA pools constructed from separate individuals (collected from different populations), despite isolating cDNAs for both paralogs of FLO and DEF from the same RACE pools, and FIM should be expressed coincidently (Ingram et al., 1995
). This suggests we might have missed a second copy expressed in the P. groanlandica floral bud cDNAs in our initial PCR screens from genomic DNA. Further studies, both within Lamiales and among appropriate outgroup taxa (e.g., Solanales), are needed to resolve the issue of copy number and possible duplication of UFO/FIM within the Lamiales.
Preservation of FLO and DEF paralogs in Lamiales
Paralogs of FLO and DEF in Lamiales appear to have avoided silencing, the typical fate of most gene duplicates. Lynch and Conery (2000)
estimate the average half-life of fully redundant gene duplicates is around 4 million years, with a 95% probability of loss after approximately 17 million years. The FLO and DEF paralogs have been maintained within at least four independently evolving lineages of Lamiales (Verbenaceae, Phrymaceae, Paulowniaceae, and Orobanchaceae) more than twice as long (estimated divergence times for families range between 32 and 53 million years; Wikstrom et al., 2001
). Thus selection must be acting on paralogs in order to maintain them within families over these time scales. Preservation of duplicate genes including floral MADS box genes (e.g., Mena et al., 1996
; Lowman and Purugganan, 1999
) and orthologs of LFY/FLO (Bomblies et al., 2003
) have been reported previously and is typically associated with restricted expression of one or both paralogs relative to expression in single-copy lineages. Such parsing of ancestral gene function or subfunctionalization is thought to be the predominant mode of duplicate gene preservation (Force et al., 1999
). Consistent with a subfunctionalization hypothesis for both, our studies of the molecular evolution of FLO and DEF paralogs found no evidence of adaptive divergence acting on duplicate copies but relaxed purifying selection in one (FLOB) or both (DEFA and DEFB) copies following the duplication event (J. E. Aagaard, unpublished data; University of Washington). Future rigorous tests of a subfunctionalization hypothesis for the preservation of FLO and DEF paralogs will require expression data during floral development as well as information on regulatory elements, which have only been well characterized for AP3 (Hill et al., 1998
).
Our gene trees show an interesting pattern suggesting that preservation of FLO and DEF paralogs in Lamiales may not have occurred independently. Specifically, all families of Lamiales for which we found two functional co-orthologs of FLO also retain two functional co-orthologs of DEF (Fig. 2). Such a pattern might be the result of rapid fixation of both FLO and DEF paralogs independently following duplication through a mechanism such as subfunctionalization (sensu Force et al., 1999
), followed by phylogenetic divergence of the Lamiales. Alternately, joint preservation might have occurred non-independently as the result of the cognate relationships between genes. LFY/FLO regulates the expression of multiple MADS box genes including AP3/DEF (reviewed in Zhao et al., 2001
), in part through binding upstream regulatory sequence (Lamb et al., 2002
). Fixation of one set of duplicates (e.g., DEFA and DEFB) via a mechanism such as subfunctionalization might increase the probability of preserving an interacting gene (FLOA and FLOB) via a distinctly different mechanism. One possibility is dosage compensation, though this seems unlikely for transcription factors such as LFY/FLO and AP3/DEF, which can activate downstream targets when expressed at low levels (LFY/FLO, Coen et al., 1990
; Weigel et al., 1992
) or are positively self-regulating (AP3/DEF, Sommer et al., 1990
; Jack et al., 1992
). Alternately, a mechanism analogous to subfunctionalization acting at the molecular level could contribute to joint preservation of FLOA/B and DEFA/B. Joint preservation of duplicate pairs of interacting genes via divergent molecular interactions and/or expression domains is referred to as concerted divergence of gene expression and has been proposed as a mechanism responsible for the preservation of parallel networks of duplicated genes in Arabidopsis resulting from ancient polyploidy (Blanc and Wolfe, 2004
). In our case, sequence divergence between regulatory elements of subfunctionalized DEF paralogs might require FLO paralogs with divergent DNA-binding domains. This hypothesis is testable because it predicts increased regulatory activity between pairs of duplicate genes. Interestingly, nonsynonymous substitutions are significantly elevated within the putative DNA binding domain of one of the Lamiales FLO paralogs (FLOB; J. E. Aagaard, unpublished data; University of Washington). Regardless of the mechanism, it will be interesting in future studies to examine the possibility of a similar pattern of correlated preservation of gene duplicates for PLE (Bradley et al., 1993
), which is also known to be regulated by FLO (McSteen et al., 1998
) as well as Lamiales homologs of other genes recently identified as regulated by LFY (William et al., 2004
).
Conclusions
The DEFA and DEFB paralogs we cloned in Lamiales represent another example of repeated duplications within the AP3/DEF gene lineage (Kramer et al., 1998
, 2003
). While previous reports have suggested LFY/FLO may be constrained as a single copy gene (Cronk, 2001
), this is clearly not the case in the Lamiales based on preservation of FLOA and FLOB paralogs within multiple families. Because our work suggests paralogs of FLO and DEF may be the result of an ancient whole-genome duplication event in the Lamiales, screens for homologs of other floral regulatory genes may reveal additional paralogous loci. Because duplications appear to have occurred after the split leading to the Veronicaceae, the model species A. majus presents a convenient outgroup for phylogenetic and developmental studies of paralogous floral regulatory genes. We believe Lamiales taxa will prove valuable for understanding the mechanisms of duplicate gene preservation and how key regulatory pathways may contribute to morphological diversity in plants.
FOOTNOTES
1 This work was supported by an NSF IGERT predoctoral fellowship and NSF dissertation improvement grant (DEB-0105176) to J. E. A., an NSF grant (DEB-0090313) to R. G. O., an NSF grant (DEB-0075704) to J. H. W., and an NIH grant to P. C. P. (GM54185). 
5 Author for correspondence (e-mail jaagaard{at}gs.washington.edu
) present address: Department of Genome Sciences, University of Washington, Box 357730, Seattle, Washington 98195-7730 USA; phone 206-221-5124, fax 206)-685-7301
LITERATURE CITED
Ambrose B. D. Lerner P. Ciceri C. Padilla M. Yanofsky R. Schmidt 2000 Molecular and genetic analyses of the Silky1 gene reveals conservation in floral organ specification between eudicots and monocots. Molecular Cell 5: 569-579[CrossRef][ISI][Medline]
Aoki S. K. Uehara M. Imafuku M. Hasebe M. Ito 2004 Phylogeny and divergence of basal angiosperms inferred from APETALA3- and PISTILLATA-like MADS-box genes. Journal of Plant Research 117: 229-244[ISI][Medline]
Beardsley P. R. Olmstead 2002 Redefining Phrymaceae: the placement of Mimulus, tribe Mimuleae, and Phryma. American Journal of Botany 89: 1093-1102
Blanc G. K. Wolfe 2004 Divergence of duplicated genes formed by polyploidy during Arabidopsis evolution. Plant Cell 16: 1679-1691
Blazquez M. D. Weigel 2000 Integration of floral inductive signals in Arabidopsis. Nature 404: 889-892[CrossRef][Medline]
Bomblies K. R. Wang B. Ambrose R. Schmidt R. Meeley J. Doebley 2003 Duplicate FLORICAULA/LEAFY homologs zfl1 and zfl2 control inflorescence architecture and flower patterning in maize. Development 130: 2385-2395
Bradley D. R. Carpenter H. Sommer N. Hartley E. Coen 1993 Complementary floral homeotic phenotypes result from opposite orientations of a transposon at the plena locus of Antirrhinum. Cell 72: 85-95[CrossRef][ISI][Medline]
Bradshaw H. S. Wilbert K. Otto D. Schemske 1995 Genetic mapping of floral traits associated with reproductive isolation in monkeyflowers (Mimulus). Nature 376: 762-765[CrossRef]
Carroll S. J. Grenier S. Weatherbee 2001 From DNA to diversity. Blackwell Science, Malden, Massachusetts, USA
Coen E. J. Romero S. Doyle R. Elliott G. Murphy R. Carpenter 1990 floricaula: a homeotic gene required for flower development in Antirrhinum majus. Cell 63: 1311-1322[CrossRef][ISI][Medline]
Cronk Q. 2001 Plant evolution and development in a post-genomic context. Nature Reviews Genetics 2: 607-619[CrossRef][ISI][Medline]
Fishman L. A. Kelly E. Morgan J. Willis 2001 A genetic map in the Mimulus guttatus species complex reveals transmission ratio distortion due to heterospecific interactions. Genetics 159: 1701-1716
Force A. M. Lynch F. Pickett A. Amores Y. Yan J. Postlethwait 1999 Preservation of duplicate genes by complementary, degenerative mutations. Genetics 151: 1531-1545
Frohlich M. D. Parker 2000 The mostly male theory of flower evolutionary origins: from genes to fossils. Systematic Botany 25: 155-170
Gaut B. J. Doebley 1997 DNA evidence for the segmental allotetraploid origin of maize. Proceedings of the National Academy of Sciences, USA 94: 6809-6814
Goldblatt P. B. Krukoff 1984 Index to plant chromosome numbers 1979–1981. Monographs in Systematic Botany 4: 1-427
Goldman N. 1993 Statistical tests of models of DNA substitution. Journal of Molecular Evolution 36: 182-198[CrossRef][ISI][Medline]
Goldman N. J. Anderson A. Rodrigo 2000 Likelihood-based tests of topologies in phylogenetics. Systematic Biology 49: 652-670[CrossRef][ISI][Medline]
Goto K. E. Meyerowitz 1994 Function and regulation of the Arabidopsis floral homeotic gene Pistillata. Genes & Development 8: 1548-1560
Hill T. C. Day S. Zondio A. Thackeray V. Irish 1998 Discrete spatial and temporal cis-acting elements regulate transcription of the Arabidopsis floral homeotic gene APETALA3. Development 125: 1711-1721[Abstract]
Hughes A. 1999 Adaptive evolution of genes and genomes. Oxford University Press, Oxford, UK
Ingram G. S. Doyle R. Carpenter E. Schultz R. Simon E. Coen 1997 Dual role for fimbriata in regulating floral homeotic genes and cell division in Antirrhinum. European Molecular Biology Organization Journal 16: 6521-6534
Ingram G. J. Goodrich M. Wilkinson R. Simon G. Haughn E. Coen 1995 Parallels between UNUSUAL FLORAL ORGANS and FIMBRIATA, genes controlling flower development in Arabidopsis and Antirrhinum. Plant Cell 7: 1501-1510[Abstract]
Irish V. 2003 The evolution of floral homeotic gene function. Bioessays 25: 637-646[CrossRef][ISI][Medline]
Jack T. L. Brockman E. Meyerowitz 1992 The homeotic gene APETALA3 of Arabidopsis thaliana encodes a MADS box and is expressed in petals and stamens. Cell 68: 683-697[CrossRef][ISI][Medline]
Kelly A. M. Bonnlander D. Meeks-Wagner 1995 NFL, the tobacco homolog of FLORICAULA and LEAFY, is transcriptionally expressed in both vegetative and floral meristems. Plant Cell 7: 225-234[Abstract]
Kelly A. J. Willis 1998 Polymorphic microsatellite loci in Mimulus guttatus and related species. Molecular Ecology 7: 769-774[CrossRef]
Kramer E. V. Di Stilio P. Schluter 2003 Complex patterns of gene duplication in the AP3 and PI lineages of the Ranunculaceae. International Journal of Plant Sciences 164: 1-11[CrossRef]
Kramer E. R. Dorit V. Irish 1998 Molecular evolution of genes controlling petal and stamen development: duplication and divergence within the APETALA3 and PISTILLATA MADS-box gene lineages. Genetics 149: 765-783
Kramer E. V. Irish 1999 Evolution of genetic mechanisms controlling petal development. Nature 399: 144-148[CrossRef][Medline]
Kuroda H. N. Takahashi H. Shimada M. Seki K. Shinozaki M. Matsui 2002 Classification and expression analysis of Arabidopsis F-box-containing protein genes. Plant Cell Physiology 43: 1073-1085
Kush A. A. Brunelle D. Shevell N. Chua 1993 The cDNA sequence of two MADS box proteins in petunia. Plant Physiology 102: 1051-1052[CrossRef][ISI][Medline]
Lamb R. T. Hill Q. Tan V. Irish 2002 Regulation of APETALA3 floral homeotic gene expression by meristem identity genes. Development 129: 2079-2086
Lowman A. M. Purugganan 1999 Duplication of the Brassica oleracea APETALA1 floral homeotic gene and the evolution of domesticated cauliflower. Journal of Heredity 90: 514-520
Lynch M. J. Conery 2000 The evolutionary fate and consequences of duplicate genes. Science 290: 1151-1155
McSteen P. C. Vincent S. Doyle R. Carpenter E. Coen 1998 Control of floral homeotic gene expression and organ morphogenesis in Antirrhinum. Development 125: 2359-2369[Abstract]
Mena M. B. Ambrose R. Meeley S. Briggs M. Yanofsky R. Schmidt 1996 Diversification of C-function activity in maize flower development. Science 274: 1537-1540
Molinero-Rosales N. M. Jamilena S. Zurita P. Gomez J. Capel R. Lozano 1999 FALSIFLORA, the tomato orthologue of FLORICAULA and LEAFY, controls flowering time and floral meristem identity. Plant Journal 20: 685-693[CrossRef][ISI][Medline]
Olmstead R. C. dePamphilis A. Wolfe N. Young W. Elisons P. Reeves 2001 Disintegration of the Scrophulariaceae. American Journal of Botany 88: 348-361
Parcy F. O. Nilsson M. Busch I. Lee D. Weigel 1998 A genetic framework for floral patterning. Nature 395: 561-566[CrossRef][Medline]
Pnueli L. M. Abu-Abeid D. Zamir W. Nacken Z. Schwarz-Sommer E. Lifschitz 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 Journal 1: 255-266[ISI][Medline]
Pouteau S. D. Nicholls F. Tooke E. Coen N. Battey 1998 Transcription pattern of a FIM homologue in Impatiens during floral development and reversion. Plant Journal 14: 235-246[CrossRef][ISI][Medline]
Raff R. 1996 The shape of life. The University of Chicago Press, Chicago, Illinois, USA
Rambaut A. N. Grassley 1997 Seq-Gen: an application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Computer Applications in the Biosciences 13: 235-238
Schwarz-Sommer Z. E. Silva R. Berndtgen W. Lonnig A. Muller I. Nindl K. Stuber J. Wunder H. Saedler T. Gubitz A. Borking J. Golz E. Ritter A. Hudson 2003 A linkage map of an F2 hybrid population of Antirrhinum majus and A. molle. Genetics 163: 699-710
Simon R. R. Carpenter S. Doyle E. Coen 1994 Fimbriata controls flower development by mediating between meristem and organ-identity genes. Cell 78: 99-107[CrossRef][ISI][Medline]
Sommer H. J. Beltran P. Huijser H. Pape W. Lonnig H. Saedler Z. Schwarz-Sommer 1990 Deficiens, a homeotic gene involved in the control of flower morphogenesis in Antirrhinum majus: the protein shows homology to transcription factors. European Molecular Biology Organization Journal 9: 605-613
Souer E. A. van der Krol D. Kloos C. Spelt M. Bliek J. Mol R. Koes 1998 Genetic control of branching pattern and floral identity during Petunia inflorescence development. Development 125: 733-742[Abstract]
Stellari G. M. Jaramillo E. Kramer 2004 Evolution of the APETALA3 and PISTILLATA lineages of MADS-box-containing genes in the basal angiosperms. Molecular Biology and Evolution 21: 506-519
Sundstrom J. A. Carlsbecker M. Svenson M. E. Svensson P. Engstrom 1999 MADS-box genes active in developing pollen cones of Norway spruce are homologous to the B-class floral homeotic genes in angiosperms. Developmental Genetics 25: 253-266[CrossRef][ISI][Medline]
Swofford D. 2002 PAUP*: phylogenetic analysis using parsimony (* and other methods), version 4.0 beta for Macintosh. Sinauer, Sunderland, Massachusetts, USA
Taylor S. J. Hofer I. Murfet 2001 Stamina pistilloida, the pea ortholog of FIM and UFO, is required for normal development of flowers, inflorescences, and leaves. Plant Cell 13: 31-46
Theissen G. A. Becker A. Di Rosa A. Kanno J. Kim T. Munster K. Winter H. Saedler 2000 A short history of MADS-box genes in plants. Plant Molecular Biology 42: 115-149[CrossRef][ISI][Medline]
Trobner W. L. Ramirez P. Motte I. Hue P. Huijser W. Lonnig H. Saedler H. Sommer Z. Schwarz-Sommer 1992 GLOBOSA: a homeotic gene which interacts with DEFICIENS in the control of Antirrhinum floral organogenesis. European Molecular Biology Organization Journal 11: 4693-4704
Weigel D. J. Alvarez D. Smyth M. Yanofsky E. Meyerowitz 1992 LEAFY controls floral meristem identity in Arabidopsis. Cell 69: 843-859[CrossRef][ISI][Medline]
Wikstrom N. V. Savolainen M. Chase 2001 Evolution of the angiosperms: calibrating the family tree. Proceedings of the Royal Society of London, B, Biological Sciences 268: 2211-2220
William D. Y. Su M. Smith M. Lu D. Baldwin D. Wagner 2004 Genomic identification of direct target genes of LEAFY. Proceedings of the National Academy of Sciences, USA 101: 1775-1780
Yanofsky M. H. Ma J. Bowman G. Drews K. Feldman E. Meyerowitz 1990 The protein encoded by the Arabidopsis homeotic gene agamous resembles transcription factors. Nature 346: 35-39[CrossRef][Medline]
50% support. Because of problems with the long branch leading to ZeaMaZFL1, bootstrap support for nodes in the LFY/FLO tree was calculated from the maximum likelihood tree that excluded ZFL1