Introduction

Auxin has been implicated in almost every growth response in plants. Polar auxin transport (PAT) moves auxin from its sites of synthesis to its sites of action in other portions of the plant, allowing auxin to influence various plant developmental processes via long-distance signaling (Leyser 2001). Auxin is synthesized in the shoot apex and in young leaves, and directionally transported toward basal parts of the plant (Lomax et al. 1995). PAT has been linked to the initiation and maintenance of polarized growth in developing embryos (Newcomb and Wetherell 1970; Fry and Wangermann 1976; Schiavone and Cook 1987; Steinmann et al. 1999), but also plays a role in the more general establishment and maintenance of the plant axis (Scheres and Berleth 1998; Friml et al. 2003), the patterning of the vascular tissues (Sachs 1993; Galweiler et al. 1998; Reinhardt et al. 2003), the development of inflorescences (Sessions et al. 1997; Nemhauser et al. 1998), and responses of plant organs to directed stimuli (tropisms) (Philippar et al. 1999; Rashotte et al. 2000).

Uptake transporters [ AUX1and LAX (Like-Aux1) proteins] allow protonated auxin to enter the cell. The higher pH inside the cell favors the anionic form of auxin, which accumulates about 20-fold over the outside concentration. Efflux carriers at the base of the cell transport anionic auxin out of the cell down auxin’s chemical concentration gradient, with the directionality being provided by the location of the carriers (Galweiler et al. 1998). Asymmetrically localized PIN proteins (efflux carriers) are required to generate the auxin gradients, influencing many aspects of plant development, especially organ initiation (Benkova et al. 2003; Reinhardt et al. 2003). While other genes also affect auxin transport (Gil et al. 2001; Noh et al. 2001; Rashotte et al. 2001), the PIN and LAX proteins play an especially important role, and localization of transporters to the membrane appears to be a dynamic process (Estelle 2001; Geldner et al. 2001, 2003; Muday and Murphy 2002; Muday et al. 2003). Accumulating evidence suggests that members of the efflux gene family are expressed in specific cells and tissues, and are cycled between compartments within the cell and the plasma membrane in a regulated manner. Although AtAUX1 and AtPIN1 are both expressed throughout the plant and often occur in the same cell, specific pairing of PIN and LAX gene expression within cell types has not been reported. Regulation includes PIN protein relocalization and expression changes in response to gravity (Friml et al. 2002b) and the application of auxin transport inhibitors such as NPA and TIBA (Geldner et al. 2001), as well as endogenous transport inhibitors such as flavonoids, and auxin itself (Peer et al. 2004). Both flavonoids and auxin have also been implicated in nodule development.

Plant hormones are postulated to play a role in nodule development (reviewed in Hirsch and Fang 1994) because they have been shown to regulate development and growth responses, and nodule formation is both a developmental event (the formation of a new plant organ) and a growth response to stimuli (rhizobia). The localized application of auxin transport inhibitors to legume roots induces nodule-like structures in several legumes and promotes the expression of some nodulation genes (Allen et al. 1953; Hirsch et al. 1989; van de Weil et al. 1990; Scheres et al. 1992). Both direct measurements of auxin (Boot et al. 1999) and measurements of the expression of an auxin responsive GH3:GUS reporter gene fusion (Mathesius et al. 1998; Pacios Bras et al. 2003) are consistent with a change in endogenous auxin transport ability during nodulation in response to bacterial Nod factors. In addition, Mathesius et al. (1998) observed an accumulation of flavonoids, natural inhibitors of auxin transport (Jacobs and Rubery 1988), at the site of application of Nod factor. Taken together, these results suggest that changes in auxin concentration within the nodulation zone are part of the plant signal cascade that regulates early symbiotic development.

Given that auxin and flavonoids influence both PIN distribution/expression and nodule formation, we suspected that auxin transport proteins might play a role in nodule development. We also wondered if there were PIN or LAX proteins specific to nodule development or if nodule development merely co-opted developmental processes that already existed. Answering these questions requires identifying PIN and LAX genes in a legume, correlating them with Arabidopsis orthologs of known function through sequence and expression analysis, and investigating the function of PINs and LAXs that show evidence of expression in nodulating roots. We took advantage of the molecular resources now available in the model legume Medicago truncatula (Cook et al. 1997; Cook 1999; Harrison 2000; Frugoli and Harris 2001; Oldroyd and Geurts 2001) to complete the first two parts of this investigation. We identified ten MtPIN genes and five MtLAX genes. We determined the genomic sequence for each of these fifteen genes, mapped them on the publicly available map, and examined their expression patterns. With the exception of one MtPIN gene, all these genes are expressed in nodulating roots of M. truncatula.

The PIN and LAX gene families are not significantly expanded in M. truncatula relative to Arabidopsis. We identified PIN and LAX genes from other species besides Arabidopsis through an analysis of unannotated sequence data available in GenBank. Sequence comparisons of the M. truncatula PIN and LAX genes with plant PIN and LAX genes identified by cDNA sequences indicate that each gene family possesses a common overall structure with several conserved areas both within M. truncatula and across species boundaries. We used these conserved segments to construct a molecular phylogeny. By combining the phylogenetic data with information on intron position and expression, we assigned orthologous relationships between Arabidopsis genes and M. truncatula genes where possible. Our results set the stage for investigating the function of these genes in nodulation.

Materials and methods

Isolation of MtPIN genes

An alignment of three A. thaliana PIN proteins and a rice PIN protein (Luschnig et al. 1998) was used to develop a degenerate primer pair for use in PCR amplification of PIN gene fragments from M. truncatula genomic DNA: AEC1 (5′ CKTTCCAYTTCATCTCCWCCA 3′) and AEC2 (5′ TGSTGCTAGACGTGGTAMTCC 3′). Amplification products were cloned into pBS-KS(+) (Stratagene, La Jolla, Calif.) and sequenced with primers from the polylinker. The insert from one clone with homology to an auxin transporter as determined by BLAST (Altschul et al. 1997) (Genbank Accession No. AQ629001) was used to screen a M. truncatula BAC library with 5× genome coverage [MT_ABa; 30,720 clones; Nam et al. 1999), available to order at http://www.genome.clemson.edu]. The insert was labeled with [32P]dCTP by random priming, and hybridization was performed according to protocols from the Clemson University Genomics Institute Web site (http://www.genome.clemson.edu/protocols/). Hybridizing clones were grouped into five contigs based on the sizes and/or sequences of the PCR products generated with primers AEC1 and AEC2 and fingerprinting of HindIII digested BAC DNA. The following prototypes for each contig were selected: 28J21 ( MtPIN1), 25F22 (MtPIN2), 71F7 (MtPIN3), 10J7 (MtPIN4), and 30I16 (MtPIN5). These BACs were analyzed by Southern hybridization with a probe from the 5′ end of MtPIN3 (positions 676–856 of Genbank Accession No. AY115838), and single hybridizing bands were detected for the MtPIN1, MtPIN3 and MtPIN4 BACs. The MtPIN5 BAC 30I16 failed to yield a signal. The MtPIN2 BAC yielded two strongly hybridizing bands, one of approximately 4 kb and the other >10 kb. Because the MtPIN2 sequence (determined as described below) was predicted to have no internal HindIII sites, we suspected that the BAC contained another PIN gene. The 4-kb hybridizing fragment was cloned into pBS-KS(+), sequenced, and indeed found to encode an additional PIN gene, designated MtPIN7.

In an effort to identify more PIN genes, a screen was performed on a portion (36,864 clones) of a second publicly available M. truncatula BAC library with 20× genome coverage (MtH2; available to order at http://www.genome.clemson.edu) using the MtPIN3 5′ end probe. Twenty-five hybridizing BACs were identified, and Southern analysis of HindIII-digested DNA from these BACs yielded two groups of BACs with a hybridizing HindIII fragment profile distinct from the MtPIN1 -4 bands previously identified. A PIN gene was sequenced from representatives of each of these groups—BAC 1A14 (MtPIN6) and BAC 71A5 (MtPIN10).

Two additional putative PIN genes were found by BLAST analysis of M. truncatula sequences from the ongoing M. truncatula Genome Sequencing Project. MtH2 BAC 23F15 (AC137994) carries a complete putative PIN gene (named MtPIN8). MtH2 BAC 30B20 carries the 5′ end of a putative PIN gene; overlapping adjacent BACs were selected from data provided by the M. truncatula physical mapping project (http://www.medicago.org/genome/) and MtH2 BAC 64A5 was used to sequence the rest of this PIN gene, which was designated MtPIN9.

Isolation of MtLAX genes

The complete sequences of MtLAX1, MtLAX2 and MtLAX3 cDNAs were published previously (de Billy et al. 2001). Full-length cDNA clones for these genes, and partial cDNA clones of MtLAX4 and MtLAX5 from the same library, were obtained (de Billy et al. 2001). The partial sequences of MtLAX4 and MtLAX5 were determined from the cDNA library clones. A portion of genomic library MtH2 (36,864 clones) was screened with a probe mixture consisting of labeled MtLAX1, 2 , 3 , and 5 cDNA clone inserts, and subsequently with a probe representing the last 555 bp of the coding region plus 260 bp of the 3′ untranslated region of the MtLAX4 cDNA. Thirty-two BACs were identified and grouped based on amplification of PCR products with pairs of MtLAX gene-specific primers. All but two of these were found to carry one of the five known MtLAX genes. Analysis of the remaining two BACs by PCR using degenerate primers based on a highly conserved region of the LAX genes corresponding to codons 49–55 and 141–148 of MtLAX1 (5′-GAYGCNTGGTTYWSITGYGC-3′ and 5′-CCRAAIARIARRAAIGTRCARTT-3′) failed to yield products. Because of the location of the primers with respect to known HindIII sites in the genes, if these two BACs contained partial sequences of MtLAX1, 2 or 3 they would not be expected to yield PCR products.

Sequencing

The sequences of the PCR fragments were used as the starting point for sequencing the putative transporter genes found in the initial five PIN contigs. Sequencing was performed directly from BAC DNA or from PCR products of BAC DNA generated by (1) pairing contig-specific forward primers with primers designed to recognize the 3′ untranslated regions of ESTs identified as putative auxin transporter transcripts, (2) pairing contig-specific reverse primers with a degenerate primer (5′-TGYTCCGGIATIAAYCGITTYGT-3′) designed to recognize a conserved amino acid motif (CSGINRFV) found near the N-terminus of auxin transporters reported from other species, and (3) using other contig-specific primers. The LAX genes were sequenced from PCR products generated from BAC DNA using cDNA-based gene-specific primer pairs and gene-specific primers paired with a degenerate primer designed to recognize a conserved amino acid motif (DAWFSCA) near the N-terminus of auxin permeases. For MtLAX4 and 5, the 5′ ends of the genes were sequenced directly from BAC DNA. All sequencing was performed on an ABI 370 Sequencer using Big Dye Terminator 2 or 3 according to the manufacturer’s standard protocols.

Mapping

Chromosomal locations of PIN and LAX genes from BACs whose location had not already been determined as part of the M. truncatula Genome Project were determined by scoring CAPS (Cleaved Amplified Polymorphic Sequence) and dCAPS (derived Cleaved Amplified Polymorphic Sequence) markers developed from BAC end sequences or from intron sequences in the genes [e.g. in Choi et al. (2004)] (Table 1). The CAPS and dCAPS markers were used to determine genomic locations by scoring them on DNA from an established mapping population consisting of individuals obtained from a cross between the M. truncatula lines A17 and A20 (Kulikova et al. 2001). The PIN and LAX markers were placed with respect to markers on the public map (Choi et al. 2004) via Colormapping. Colormapping is a non-mathematical method which uses a color matrix to display the genotypes of a core set of markers in members of a mapping population, allowing new markers to be placed in relation to known markers by visual analysis (Kiss et al. 1998; Kalo et al. 2000).

Table 1 Primers used for mapping genes
Full size table

PIN and LAX gene expression analysis

Intron-spanning, gene-specific PCR primer pairs were chosen to differentiate between possible products amplified from genomic DNA contaminants and products amplified from cDNA (Table 2). The primers were tested for gene specificity using PIN gene BACs or LAX gene cDNA clones as template. PCRs (10 μl) consisted of PCR Mastermix (Eppendorf AG, Hamburg, Germany), each primer at 0.5 μM, and 2 ng of template DNA. Thermocycling conditions were 95°C for 4 min, followed by 30 cycles of 95°C for 20 s, 61°C (PIN genes) or 63°C (LAX genes) for 20 s, and 72°C for 45 s. Products were analyzed on 1% or 2% (w/v) agarose gels in 0.5× Tris-Borate-EDTA buffer and visualized by staining with ethidium bromide. In each case, only the intended target yielded PCR products.

Table 2 Primers used for RT-PCR
Full size table

RNA was purified from plant tissues using the RNeasy Plant Mini Kit (Qiagen, Valencia, Calif.) according to the manufacturer’s instructions. Roots (minus 1 mm of root tip) and shoots of 12-day old nodulating plants were separated for extraction. RNA (20 μg) was treated with 10 U of RQ1 RNase-free DNase in (Promega, Madison, Wis.) in a 100-μl reaction at 37°C for 30 min. Following phenol and chloroform extractions and ethanol precipitation, the RNA was resuspended in 20 μl of distilled water and quantified by measuring the absorbance of a 1:50 dilution at 260 nm. cDNA was synthesized in 20-μl reactions from 1-μg aliquots of RNA, using random hexanucleotide primers (Invitrogen, La Jolla, Calif.) and Superscript Reverse Transcriptase II (Invitrogen) following the manufacturer’s recommendations. PCRs were performed and analyzed as above using 0.5 to 4.0 μl of cDNA as template. The quantity used depended on the primer set, but was always the same for shoots and roots.

Phylogenetic analysis

Additional PIN and LAX sequences from other species where obtained by searching the Genbank Non-Redundant Database (nr-July 2004) with TBLASTN. PIN and LAX sequences for which a genomic sequence and/or a full length cDNA was found in Genbank or the TIGR Gene Index of Tentative Consensus sequences (TCs) were included in the analysis. The presence of a TC alone was not considered sufficient to include a gene in the analysis because the compilation program separates multiple TCs when only a few ESTs or short ESTs from different parts of the sequence are present, confounding the phylogeny with partial length sequences.

Protein sequence alignments were generated using the CLUSTAL method in MegAlign (DNASTAR, Madison, Wis.) with minor adjustments being made manually. Positions in gaps in the alignments were eliminated from the analyses. Unrooted most parsimonious trees were produced from the results of 100 bootstrap replicates using the PROTPARS and CONSENSE algorithms of Phylip 3.6 with default parameters. The corresponding nucleotide alignments were used to generate unrooted maximum likelihood trees from 100 bootstrap replicates using DNAML and CONSENSE (Phylip 3.6) with default parameters. Trees were viewed and edited with TreeView (Page 1996) and show only branches supported by more than 50% of the replicates.

Results

Identification and mapping of members of a PIN gene family in M. truncatula

Ten members of the PIN gene family of auxin efflux transporters (MtPIN1–MtPIN10) were identified by a combination of genomic library screening and analysis of publicly available sequences of BAC library clones (see Materials and methods). The ten predicted proteins range in size from 357 to 660 amino acids and are 28 to 79% identical to each other. PIN proteins comprise a conserved N-terminal region of transmembrane segments, a variable middle region thought to be a cytoplasmic domain, and a conserved C-terminal region of transmembrane segments. The differences between the proteins arise mainly from differences in the length and sequence of the central variable region. The genes can be broadly grouped into two categories based on the length of this middle region, with one group of genes having a much shorter sequence than the other. Use of only conserved regions of the proteins in phylogeny analysis resulted in the proteins with similar variable regions clustering together, even though this portion of the sequence was not included in the analysis.

The genes all have a similar overall intron/exon structure, with most having a large first exon corresponding to the N-terminal transmembrane segments and the majority of the variable segment, followed by a second exon corresponding to the rest of the variable segment and the beginning of the C-terminal transmembrane region. Four small exons of 86, 158, 77, and 64 bp, respectively, encode the remaining transmembrane segments. This same overall gene structure is observed in the PIN genes from Arabidopsis (Fig. 1C).

Fig. 1A–C
figure 1

Phylogenetic analysis and intron exon structure of PIN genes. A Parsimony analysis based on an alignment of the conserved N- and C-terminal regions of the PIN proteins (representing 302 amino acids-see Materials and methods). Only branches with greater than 50% support are displayed. B Maximium likelihood tree of dicot nucleotide sequences encoding conserved N- and C-terminal regions of PIN gene products. C Deduced intron/exon structure of PIN genes for which genomic sequences were available. The dotted lines indicate gaps in the nucleotide sequence. The numbers at the bottom indicate the number of amino acids and nucleotides aligned from each section. Sequences used in these alignments were the following: MtPIN1 (AY115836), MtPIN2 (AY115837), MtPIN3 (AY115838), MtPIN4 (AY115839), MtPIN5 (AY115840), MtPIN6 (AY553209), MtPIN7 (AY553210), MtPIN8 (BK005119), MtPIN9 (AY553211) and MtPIN10 (AY553212) from Medicago truncatula); AtPIN1 (AF089085), AtPIN2 (AF086907), AtPIN3 (AF087818), AtPIN4 (AF087016), AtPIN5 (AB005242), AtPIN6 (AF087819), AtPIN7 (AF087820) and AtPIN8 (AL391146) from Arabidopsis thaliana; CsPIN1 (AB08589) from Cucumis sativus; BjPIN1 (AJ132363), BjPIN2 (AJ249297) and BjPIN3 (AJ249298) from Brassica juncea; LjPIN1 (BK005120) and LjPIN2 (BK005121) from Lotus japonicus; MCm-AEC1 (AF246995), MCm-AEC2 (AF247004) and MCm-AEC3 (AF247005) from Momordica charantia; OsREH1 (AF056027), AK059229, AK103181, AK063976, AK101191, AK066552, AK100297, AK103208 and AP003408c from Oryza sativa (the coding sequence of this gene has been reanalyzed; the start site has been redefined to include an additional 198 bp of upstream sequence); PsPIN1 (AY222857) and PsPIN2 (AB112364) from Pisum sativum; PttPIN1 (AF190881), PttPIN2 (AF515435) and PttPIN3 (AF515434) from Populus tremula x P. tremuloides; and TaPIN1 (AY496058) and TaPIN2 (BK005137) from Triticum aestivum

Full size image

Comparison of the M. truncatula PINs reveals that Mt PIN2 and Mt PIN7 are highly similar to each other, with 81% amino acid sequence identity, 87% nucleotide sequence identity throughout the coding regions, and 73% identity when intron sequences (excluding a region missing from the variable segment of MtPIN7) are included in the comparison. The two genes are located within 20 kb of each other. The first intron of MtPIN7, which occurs earlier in the sequence than that of MtPIN2, contains several small insertions, but 45 of the remaining 86 bp in the MtPIN7 intron are identical to the coding sequence of MtPIN2.

MtPIN1 and MtPIN3 show 79% amino acid sequence identity and 74% nucleotide sequence identity within their coding regions; MtPIN4 and MtPIN5 have 73% amino acid identity and 69% nucleotide sequence identity within the coding regions (excluding short stretches of the MtPIN4 variable region which are missing from MtPIN5). MtPIN6 possesses an intron near the 5′ end which is not present in other M. truncatula PIN genes but is found at an identical position in AtPIN6. MtPIN8 and MtPIN9 are predicted to encode shorter proteins than the other PIN genes, lacking most of the variable segment; such a structure is also observed in AtPIN5 and AtPIN8.

The genomic locations of MtPIN1 – MtPIN7 were determined by developing BAC-specific markers that were polymorphic between the related M. truncatula ecotypes A17 and A20, and testing these markers on a mapping population on which several dozen markers have already been mapped (Table 3). MtPIN8, MtPIN9 and MtPIN10 are contained in contigs that have been anchored on the M. truncatula physical map during the course of the genome sequencing project. Four of the genes (MtPIN2, MtPIN3, MtPIN6, and MtPIN7) map to an 8-cM region on the short arm of chromosome 1, two to the long arm (MtPIN9 and MtPIN10) and one to the short arm (MtPIN8) of chromosome 7, and one each to chromosomes 4 (MtPIN1; long arm), 6 (MtPIN4; near centromere), and 8 (MtPIN5; long arm).

Table 3 Genomic locations of Medicago truncatula PIN and LAX genes
Full size table

PIN genes from other plant species

BLAST analysis of the complete Arabidopsis genome sequence indicates that there are eight PIN genes in this species. Expression of each of these has been demonstrated (Galweiler et al. 1998; Muller et al. 1998; Zhong and Ye 2001; Friml et al. 2002a, 2002b; Benkova et al. 2003; Reinhardt et al. 2003). Expressed PIN genes have also been described in other dicot species, including hybrid aspen ( Populus tremula × Populus tremuloides; Schrader et al. 2003), pea (Pisum sativum), bitter gourd (Momordica charantia) and Brassica juncea. We have identified by BLAST analysis two putative PIN genes in Lotus japonicus from completed TAC clone sequences, and named these LjPIN1 (Chr2; clone LjT06K23) and LjPIN2 (Chr4; clone LjT05D12).

Among the monocots, one PIN gene each had previously been identified in rice (OsREH1; Luschnig et al. 1998) and wheat (TaPIN1). BLAST analysis of rice BAC sequences and the full-length cDNA database (http://cdna01.dna.affrc.go.jp/cDNA/) yielded sequences of seven additional PIN genes (left unnamed pending annotation by the Knowledge-based Oryza Molecular Biological Encyclopedia). An additional wheat PIN gene was found among cDNA sequences in GenBank and has been named TaPIN2.

Relationships between PIN genes from M. truncatula and other plant species

PIN protein sequences and coding sequences were used to resolve evolutionary relationships within this gene family and aid in determining orthology. The aligned N- and C-terminal conserved regions of the sequences (representing 302 amino acids) were used in parsimony analysis (PROTPARS; Phylip 3.6) to generate a phylogenetic tree (Fig. 1A). An attempt was made to further refine the tree by using the corresponding nucleotide sequences for maximum likelihood tree generation. Examination of the alignment, however, showed codon bias in the monocot sequences, with high G or C preference in the third codon position. Such codon bias is common in rice genes (Yu et al. 2002) and was also noted in the catalase gene family in maize (Guan and Scandalios 1996). This bias would skew the phylogenetic analysis, and therefore only dicot species were used in the nucleotide sequence analysis (Fig. 1B) with DNAML (Phylip 3.6).

The sequences fell into two major groups. One group contained all the shorter proteins plus AtPIN6 and MtPIN6. The other group contained the remaining longer proteins. When nucleotide sequence identities are compared within species, a similar range is observed. The Arabidopsis PINs share 36.5–84.5% identity, while the Medicago genes range from 38.9 to 81.4% identity. When nucleotide sequence identities are compared within groupings however, identities within clades are higher than identities within species.

For instance, PsPIN2 and MtPIN3—two legume genes—are 92.8% identical, while the closest M. truncatula gene to MtPIN3 (MtPIN1) shares only 74.2% identity with it. Likewise, two other legume genes (LjPIN1and MtPIN8) are 75.6% identical to each other, but only ~56% identical to AtPIN8. Thus within-species identity is often lower than identity within classes of PIN sequences.

Identification and mapping of LAX genes in M. truncatula

Five MtLAX genes were identified from cDNA clones and BAC libraries (see Materials and methods). Four of the genes were localized on the M. truncatula genetic map using intron-based markers (Table 3). MtLAX1 mapped to a 5.5-cM region on the long arm of chromosome 5, MtLAX2 to a 22-cM region in the middle of chromosome 7 and MtLAX3 to a 3-cM region on the long arm of chromosome 3, while MtLAX5 is tightly linked to the markers DNABP and NUM1 on the long arm of chromosome 4. MtLAX4 was mapped by analysis of available M. truncatula genome fingerprinting and sequence data. Sequences from the ends of two BACs carrying MtLAX4 (4L2 and 62N11) are found near the end of the completely sequenced BAC MtH2 35F15 (AC138016), which carries the marker TE016. TE016 has been mapped to the telomeric region of the short arm of chromosome 4 (Choi et al. 2004). Various pairs of gene-specific primers for MtLAX4 failed to yield PCR products from the A20 and R108 ecotypes (data not shown), suggesting that MtLAX4 may be absent in these ecotypes.

All of the M. truncatula LAX genes are highly similar at the protein level (72.8–85.6% identity; PAM250 residue weight table), although parsimony analysis splits them into two major groupings, with MtLAX1, 2 and 4 forming one branch and MtLAX 3 and 5 another (Fig. 2A). Examination of sequence pairs shows that MtLAX2 and MtLAX4 show the highest level of sequence identity (85.6%), but MtLAX2 displays even greater similarity to a pea sequence, PsAux1 (93.8%)—so again, within-species identity is often lower than identity within classes. Examination of their nucleotide sequences shows that all of the M. truncatula LAX genes have seven introns within the coding region.

Fig. 2
figure 2

A, B Phylogenetic analysis of LAX proteins. A Parsimony analysis of LAX proteins. Only branches with greater than 50% support are displayed. B Maximum likelihood tree based on nucleotide sequences of LAX genes. The sequences used in these alignments were the following: MtLAX1 (AY115841), MtLAX2 (AY115843), MtLAX3 (AY115842), MtLAX4 (AY115844) and MtLAX5 (AY115845) from Medicago truncatula, AtAUX1 (X98772), AtLAX1 (AJ249442), AtLAX2 (AJ243221) and AtLAX3 (AY127575) from Arabidopsis thaliana, AK102729, AK103524, AK100090, and AK111849 from Oryza sativa, PttLAX1 (AF115543), PttLAX2 (AF190880), and PttLAX3 (AF263100) from Populus tremulus, PsAUX1 (AB107919) from Pisum sativum, LjLAX1 (BK005138) from Lotus japonicus, ZmAUX1 (ZmAUX1) from Zea mays, and CsAUX1 (AB085896) from Cucumis sativus

Full size image

LAX genes in other plant species

Four members of the auxin influx carrier family are present in the Arabidopsis genome, and these vary in intron number (having 5–7 introns within the coding region) with the variation occurring at the C-terminal end of the genes (Parry et al. 2001). Expressed auxin influx carrier genes have been described in hybrid aspen, pea, maize ( Zea mays), and cucumber ( Cucumis sativus). BLAST analysis of the full-length rice cDNA database yielded sequences of four auxin influx carriers. One additional putative auxin influx carrier gene was identified among Lotus japonica TAC clone sequences and was named LjLAX1. The TIGR Gene Index for L. japonica (version 3.0) lists four partial LAX TCs, which represent three separate genes. Each of these sequences is different from the LjLAX1, suggesting the number of LAX sequences in L. japonica is at least four, but without full-length sequence the Lotus sequences could not be included in the phylogenic analysis. Partial sequence comparison suggests that the excluded L. japonica sequences are most similar to MtLAX1, 2 and 4.

Relationships between LAX genes of M. truncatula and other plant species

Because of their extreme similarity, nearly the entire sequence of the LAX genes could be used for analysis. Two groups of LAX proteins emerge from the parsimony analysis (Fig. 2A). There is insufficient information based on protein parsimony to elucidate relationships between many of the sequences in the top clade, but maximum likelihood analysis of the nucleotide sequences (Fig. 2B) allows some comparisons. For instance, the two Arabidopsis sequences are more similar to each other than to any other gene in this clade, an observation that does not hold for the multiple sequences from poplar or M. truncatula.

Expression of M. truncatula PIN and LAX genes

With the exception of MtPIN5, all M. truncatula PIN genes could be amplified from cDNA obtained from nodulating plants (Fig. 3). MtPIN2 could only be amplified from roots. The bands seen in MtPIN5 reactions were also observed in no-template controls (not shown) and are assumed to be primer dimers, since they do not match the expected size of the product from the predicted cDNA. The major bands observed in other reactions migrate at sizes expected based on predicted intron/exon structures (Table 2).

Fig. 3
figure 3

Confirmation of expression of MtPIN1 – MtPIN10 in nodulating roots and shoots of the wild-type M. truncatula ecotype A17. Total RNA was isolated from 12-day old nodulating plants, and used to test for the presence or absence of gene expression by RT-PCR. Primers specific for each gene (Table 2) were used to generate the products in the indicated lanes. Bands seen in the MtPIN5 reactions were also observed in the no-template control (not shown) and are assumed to be primer dimers. The major bands observed in other reactions migrate at sizes expected based on the predicted intron/exon structures listed in Table 2

Full size image

Further evidence that MtPIN5 is not expressed comes from its absence from the public collection of almost 190,000 ESTs (http://www.tigr.org/tigr-scripts/tgi/T_index.cgi?species=medicago). There is no obvious defect in the sequence of the coding region of MtPIN5, including the intron/exon boundaries, that would prevent the gene from being expressed, but the intron/exon boundaries have shifted (Fig. 1C).

The availability of the TIGR Medicago Gene Index made it possible to assemble “electronic Northerns” for PIN and LAX genes represented in the EST database. MtPIN5, 6 , 7 , 8 , 9 and 10 do not appear as ESTs in release 7.0 of the database. MtPIN2, 3 and 9 appear as single ESTs from root tips (TC91788), phosphate-starved leaves (TC83905), and seedling roots 24 h after inoculation (TC81304), respectively. Only MtPIN1 is represented by more than one EST: three independent MtPIN1 ESTs come from a developing flower library (TC80729).

The original report identifying the initial members of the MtLAX gene family reported expression of MtLAX1, 2 and 3 in nodulating roots by in situ hybridization (de Billy et al. 2001). RT-PCR of the same cDNAs used for the PIN analysis with primers from Table 2 demonstrated amplification of transcripts of all five LAX genes from both shoots and roots of nodulating plants (data not shown). “Electronic Northerns” based on release 7.0 of the TIGR Medicago Gene Index were more informative. MtLAX4 does not appear among the ESTs, and MtLAX5 is represented by a single EST from an early seed development library (TC84343). MtLAX1 is represented twice, once from developing roots and once from a library made from leaves subjected to insect herbivory (TC89992), confirming the RT-PCR results. MtLAX2 is represented five times—twice from the early seed development library, and once each from phosphate starved leaves, developing leaves and nematode infected roots (TC88482). MtLAX3 is by far the most widely/abundantly expressed of the LAX genes with 12 independent ESTs: two from phosphate starved leaves, one from developing flowers, two from roots colonized with Glomus versiforme (mycorrhizae), and one each from stems, developing stems, developing seeds, Colletotrichum trifolii  infected leaves, seedling roots, root tips, and a mixed tissues library (TC86757).

Discussion

The combination of protein and DNA sequence comparison, mapping, analysis of gene structure, and expression data support some direct orthologies between Arabidopsis and M. truncatula auxin transport genes. For six expressed Medicago truncatula auxin transport genes we were able to find a direct ortholog in Arabidopsis, while for seven auxin transport genes the relationships were more complex (Table 4).

Table 4 Summary of Arabidopsis/M. truncatula PIN and LAX gene relationships
Full size table

PIN genes

Of the ten M. truncatula PIN genes isolated, nine are expressed (Fig. 3). Two pairs of MtPIN genes appear to be the result of gene duplications. MtPIN4 and MtPIN5 are closely related in sequence, but the change in intron structure and the lack of expression of MtPIN5 suggest that it may be silenced—one possible fate of duplicated genes (Lynch and Conery 2000).

A second gene duplication event is evident in the MtPIN2/MtPIN7 pair of efflux transporter genes. MtPIN7 is located on the same BAC as MtPIN2 (Table 3), is highly homologous to MtPIN2 at both the DNA and protein sequence levels, contains intron sequences with homology to the coding sequence of MtPIN2 (which could result from the intron boundary shifting in MtPIN7 when it duplicated and the coding sequence degenerated), and is part of a clade in which all other species examined contribute only one member (Fig. 1), suggesting that MtPIN7 is the result of a duplication of MtPIN2 that occurred recently in the Medicago lineage. Since MtPIN2 is expressed in the roots and not in the shoots while MtPIN7 is expressed in both the shoots and roots (Fig. 3), the functions of the two genes may have diverged since their duplication. Under this scenario, MtPIN2 and MtPIN7 could be considered as a pair of paralogs, sometimes termed a derived and an ancestral ortholog (Koonin 2001). Although the half-life of a duplicated gene has been estimated at only 3–7 Myr (Lynch and Conery 2000), genes are more likely to survive duplication if they undergo subfunctionalization (including changes in levels of expression) or partitioning of functions (Wolfe 2001). The MtPIN2/MtPIN7 pair provides an excellent example of such a case, and detailed expression and mutational analysis are ongoing in our laboratory.

Several direct orthologies are suggested by the DNA and protein comparisons and the intron/exon analysis in Fig. 1 and summarized in Table 4. AtPIN6 appears to be the ortholog of MtPIN6 based on their sequence identity and the conserved structure of their intron/exon boundaries. Little has been published about the expression pattern of AtPIN6, except that transcripts are present in lateral root primordia (Benkova et al. 2003), which is consistent with MtPIN6 expression in roots. Likewise AtPIN8/MtPIN8 and AtPIN5/MtPIN9 appear to be orthologs. These three pairs occur in a separate clade from the rest of the PIN genes and, except for the PIN6 grouping, contain members from both monocots and dicots, suggesting that they are ancestral groupings. Both M. truncatula genes are expressed in roots and shoots, but the expression pattern of neither Arabidopsis gene has been reported. If the expression data for MtPIN7 is taken into account, AtPIN2 and MtPIN2 in the upper clade could also be orthologs, because AtPIN2 is expressed only in roots, and MtPIN2 shares the same expression pattern (Fig. 3), despite the unusual intron structure of AtPIN2 (Fig. 1C).

Other relationships are not as clear. In one case, three Arabidopsis genes (AtPIN4, AtPIN7 and AtPIN3) are in the same clade with two M. truncatula genes (MtPIN1and MtPIN3). In another case, three M. truncatula genes (MtPIN4, MtPIN5 and MtPIN10) are in a clade to which Arabidopsis contributes only one member (AtPIN1). In the first case, the Arabidopsis genes appear to be more closely related to each other at both the DNA and protein level than to either of the M. truncatula genes, and all of the intron structures are similar. The three Arabidopsis genes have all been shown to have roles in root growth and responses to gravity (Friml et al. 2002a, 2002b; Benkova et al. 2003), but their expression is not specific to roots (Benkova et al. 2003) and neither is that of the M. truncatula genes (Fig. 3). In the second case, if MtPIN5 is assumed to be silent because of its structure and lack of detectable expression, the relationships within this clade at the DNA level (Fig. 1B) suggest that MtPIN4 and MtPIN10 are closer to each other in sequence than to AtPIN1. Arabidopsis stands out as contributing only one member to the clade, while M. truncatula, poplar, pea, wheat and rice all contribute multiple members. Since the Arabidopsis genome has been completely sequenced, the presence of multiple genes from both monocots and dicots in this clade implies that Arabidopsis lost a PIN gene after its evolutionary divergence from the organisms represented in the tree rather than that a PIN gene was duplicated in M. truncatula. In addition, there is no evidence for subfunctionalization or partitioning of functions in the M. truncatula genes: AtPIN1 is expressed in seedlings, flowers, roots, leaves and stems (Galweiler et al. 1998) and MtPIN4 and MtPIN10 are expressed in both roots and shoots.

Thus five MtPIN genes can be assigned one-to-one with four Arabidopsis PIN genes, with one gene pair (MtPIN2/MtPIN7) assigned as a metalog pair of AtPIN2. The remaining four expressed MtPINs cannot with confidence be paired on a one-to-one basis with the remaining four AtPINs based on the present data.

AUX / LAX genes

AUX / LAX genes comprise a smaller family than PIN genes in all organisms for which LAX gene sequences have been deposited in Genbank, although the incomplete nature of most plant genome sequences must of course be taken into account. Unlike the PIN gene comparisons, molecular phylogeny of the LAX genes suggests an evolutionary divergence in this gene family between monocots and dicots (Fig. 2). All monocot sequences analyzed thus far fall into two clades, and each of these clades is an individual branch on one of the two large branches that divide the AUX/LAX gene family into two groups.

As with PIN genes, Arabidopsis has fewer gene family members than M. truncatula A17 (four as opposed to five). However, because of the ecotype specificity of MtLAX4, the two organisms may share the same number of gene family members at the species level. MtLAX5 and AtLAX2 appear to be orthologs, as do MtLAX3 and AtLAX3. However, the relationships between AtAUX1 and AtLAX1 and M. truncatula LAX genes are more complex. AtAUX1 is expressed in leaves and roots and often occurs in the same cells as AtPIN1 (Swarup et al. 2001). For many Arabidopsis PIN gene family members, physiological and genetic data are available to support their roles in auxin transport. However, among the AtAUX/LAX genes, only for AtAUX1 is supporting data available to demonstrate its involvement in auxin transport (Marchant et al. 2002). The rest of the Arabidopsis genes have been identified and grouped in this class by sequence homology or based on unpublished observations (Parry et al. 2001). The MtLAX genes are expressed in both roots and shoots. The expression patterns of the AtLAX genes have not yet been published, but like the AtLAX genes, the presumed involvement of MtLAXs in auxin transport is based on sequence homology.

While AtAUX1and AtLAX1 share close sequence identity to each other, their relationships to MtLAX4, MtLAX2 and MtLAX1 are unclear, even when DNA sequence comparisons are made (Fig. 2B). Based on a smaller number of sequences Parry et al. (2001) concluded that AtLAX2 was the founding member of the AUX/LAX family in Arabidopsis, and that other family members arose by gene/genome duplication prior to angiosperm evolution. The data reported here support this conclusion, with MtLAX5 being the ortholog of AtLAX2 and MtLAX3 the ortholog of AtLAX3. While the DNA comparison suggests that MtLAX2 and MtLAX4 are more closely related to AtAUX1 and AtLAX1 than MtLAX1, we were unable to discern any further relationships within this group due to the similarities in the expression patterns of the genes. In Arabidopsis, AtAUX1 is critical for many developmental processes, and the ubiquitous expression of MtLAX2 could imply orthology to AtAUX1. MtLAX4 may be a duplication of MtLAX2 in the A17 ecotype because, although we were able to detect MtLAX4 sequence in the A17 ecotype of M. truncatula by PCR and it is represented on at least four BACs in the MtH2 BAC library from this ecotype, we were unable to amplify MtLAX4 from the A20 ecotype or the more closely related R108 ecotype.

Thus two MtLAX genes (MtLAX3 and 5) form orthologous pairs with AtLAX3 and AtLAX2, respectively. Of the remaining three MtLAX genes, none correspond directly with the AtLAX1 and AtAUX1, but MtLAX2 is the best candidate for an ortholog of AtAUX1 based on expression pattern.

PIN or LAX genes specific to M. truncatula

A simplistic approach to identifying auxin transporters involved in nodulation would be to identify a PIN or LAX gene sequence specific to M. truncatula (i.e., lacking an orthologous counterpart in the non-nodulating Arabidopsis), which we were unable to do. The fact that no auxin transporters are exclusive to nodule development certainly does not preclude their involvement in the process. The MtPIN2/MtPIN7 pair is the most recent duplication for which both genes are expressed in M. truncatula. Only one gene from another legume is present in the same clade, and since a comprehensive assessment of gene family members in other legume species besides M. truncatula has not been done, it is difficult to determine if the duplication is legume specific, M. truncatula specific, or if Arabidopsis has lost a PIN sequence. However, the extremely close sequence identity between the intron region of MtPIN7 corresponding to a region coding for protein in MtPIN2, leads us to suggest a recent event. While neither gene family is dramatically expanded in M. truncatula relative to Arabidopsis, M. truncatula has more expressed PIN and LAX genes than Arabidopsis (8 PINs and 4 AUX/LAXs in Arabidopsis versus 9 PINs and 5 LAXs in M. truncatula), so it is possible that some of these sequences may represent genes used exclusively in nodule development. Indeed, we suffer from an abundance of candidates: all of the expressed genes identified in this study are expressed in nodulating roots. If nodulation simply co-opts normal plant developmental processes, we would expect several of these genes to be involved in nodulation. We are currently pursuing this investigation using antibodies and RNA interference expressed against various PIN sequences, and preliminary analysis suggests that at least four of the MtPIN genes have some role in nodule development (Huo, Schnabel and Frugoli, unpublished data).