Control of dissected leaf morphology by a Cys(2)His(2) zinc finger transcription factor in the model legume Medicago truncatula

Abstract

Plant leaves are diverse in their morphology, reflecting to a large degree the plant diversity in the natural environment. How different leaf morphology is determined is not yet understood. The leguminous plant Medicago truncatula exhibits dissected leaves with three leaflets at the tip. We show that development of the trifoliate leaves is determined by the Cys(2)His(2) zinc finger transcription factor PALM1. Loss-of-function mutants of PALM1 develop dissected leaves with five leaflets clustered at the tip. We demonstrate that PALM1 binds a specific promoter sequence and down-regulates the expression of the M. truncatula LEAFY/UNIFOLIATA orthologue SINGLE LEAFLET1 (SGL1), encoding an indeterminacy factor necessary for leaflet initiation. Our data indicate that SGL1 is required for leaflet proliferation in the palm1 mutant. Interestingly, ectopic expression of PALM1 effectively suppresses the lobed leaf phenotype from overexpression of a class 1 KNOTTED1-like homeobox protein in Arabidopsis plants. Taken together, our results show that PALM1 acts as a determinacy factor, regulates the spatial-temporal expression of SGL1 during leaf morphogenesis and together with the LEAFY/UNIFOLIATA orthologue plays an important role in orchestrating the compound leaf morphology in M. truncatula.

Plant leaves are lateral organs initiated as a peg-like structure from the flank of the shoot apical meristem (SAM), a pluripotent structure that is capable of self-renewal. They can be simple, consisting of a single flattened blade subtended by a petiole, or compound (or dissected), consisting of multiple blade units known as leaflets. The class 1 Knotted1-like homeobox proteins (KNOXIs) are required to maintain indeterminacy of the SAM (1). During early development, KNOXI genes are down- regulated at the incipient leaf primordium at the periphery of the SAM (2). This down-regulation marks the site of primordia initiation and is permanent in developing primordia that lead to simple leaves. However, the KNOXI genes are transiently reactivated in leaf primordia in most eudicot species that have compound leaves, indicating a requirement for a transient phase of indeterminacy in the initiation of leaflet primordia at leaf margins during compound leaf development (3). However, the transient indeterminacy is not sufficient for compound leaf development, as it can also lead to simple leaves as a result of secondary morphogenesis in some plants (3). In some leguminous plants (Fabaceae) that belong to the inverted repeat lacking clade (IRLC), including garden pea (Pisum sativum) and alfalfa (Medicago sativa), the role of KNOXIs in maintaining indeterminacy is replaced by the FLORICAULA (FLO)/LEAFY(LFY) transcription factor UNIFOLIATA (UNI)/SINGLE LEAFLET1 (SGL1) (4–6), because KNOXI proteins are not detected in leaves in the IRLC legumes (6). However, conflicting evidence exists to support the expression of KNOXI transcripts in leaves of IRLC legumes (7, 8). Recently, common molecular frameworks using NAM/CUC transcription factors and auxin gradients have been shown to play a role in compound leaf development in diverse species (9–12). Leguminous plants belong to the third largest family of flowering plants with significant economic value (13). Leaves of many leguminous species are pinnate, palmate, or higher-ordered compound. However, how different leaf morphology is determined in legumes is not yet understood.

Results

Isolation and Characterization of palmate-like pentafoliata1 Mutants.

To identify additional regulators of leaf morphogenesis in legumes, we screened a mutant collection in the model legume Medicago truncatula (cv. Jemalong A17) derived from fast neutron bombardment deletion mutagenesis and isolated two leaf mutants, M469 and M534. Mature leaves developed in these two mutants are palmate-like pentafoliate in contrast to the trifoliate WT leaves (Fig. 1 A–E). We named these mutants palmate-like pentafoliata1-1 (palm1-1) and palm1-2. Compared with WT compound leaves, which have a terminal and two lateral leaflets (LLs), mature leaves in the palm1 mutants have a terminal and two pairs of LLs clustered at the tip of the petiole. In addition, the two distally oriented LLs (LLd) are subtended by rachis structures similarly as the terminal leaflet (TL; Fig. 1 B–E). SEM and histochemical analysis show the presence of elongated epidermal cells at the surface and three vascular bundles, with two on the adaxial side of the rachis structures, indicating changes of LLd to the TL morphology in the palm1 mutant (Fig. 1 H, I, M, and N and Fig. S1).

Fig. 1.

M. truncatula palm1-1 mutant exhibits altered leaf form: (A) 3-wk-old WT (Jemalong A17; Left) and palm1-1 mutant (Right); (B) juvenile (Left) and compound (Right) leaves of WT plant; (C) high-magnification view of B; (D) juvenile (Left) and compound (Right) leaves of the palm1-1 mutant; and (E) high-magnification view of D. (F) Measurements of the petiole length of compound leaves on the fifth node of 6-wk-old WT and palm1-1 mutant plants. (G) Measurements of the rachis length of compound leaves. Shown are means ± SE (n = 10; F and G). (H–Q) SEM images of compound leaves (H and M), rachis (I, N; arrowheads in H and M), and leaf primordia at different developmental stages (J–L; O–Q) of WT (H–L) and the palm1-1 mutant (M–Q). Asterisks in H and M indicating petiolules. Arrowhead in K indicates the boundary between St and LL. Arrows in P and Q indicate LLp developed in the P₃ primordia. P, Plastochron; Rac, rachis; Pet, petiole. (Scale bars, 1 cm in B and D.)

Accompanying these changes, the palm1 mutants also exhibit alterations in the proximal-distal axis of compound leaves. Compared with WT leaves, the petiole length of mature leaves in 6-wk-old palm1-1 mutant was increased by approximately 20% (Fig. 1F). Conversely, the length of the central rachis was reduced by approximately 19%, although rachis structures were developed on LLd in the mutant (Fig. 1G).

We attempted to identify the earliest morphological alterations during leaf development in the palm1 mutant using SEM. In WT plants, leaf primordia after initiation (P₀ for Plastochron 0) from the periphery of the SAM developed a pair of stipule (St) primordia at P₁, a pair of LL primordia, boundaries between St and LL, and LL and TL at P₂, and the differentiation of TL and St as indicated by trichomes developed on their abaxial surface at P₃ (5) (Fig. 1 J–L). Leaf development progressed normally in the palm1-1 mutant until the P₃ stage, when a pair of extra leaflet primordia, the proximally oriented LL (LLp), developed at the base of LLd, which were initiated at the P₂ stage (Fig. 1 O–Q; arrows). The earliest morphological alteration in the palm1 mutant, the development of extra leaflet primordia in a basal position, suggest that PALM1 plays a key role in the suppression of the morphogenetic activity in the proximal region of the compound leaf primordium, which is required to maintain the trifoliate morphology of compound leaves and the morphology of LL without the rachis structure in WT plants.

PALM1 encodes a Cys(2)His(2) zinc finger transcription factor.

By using a map-based approach, we mapped the PALM1 locus to a 45-kb interval on chromosome 5 (Fig. 2 A–C; Table S1 and Figs. S2 and S3). There are eight annotated ORFs in this genomic interval that are deleted in both palm1-1 and palm1-2 mutants (Fig. 2D). Three show syntenic relationships with Arabidopsis thaliana homologues on chromosome 4 (Table S2). To test which ORF is the candidate gene, we first screened two other mutant collections and isolated four mutants with the same phenotype as the original palm1-1 and palm1-2 mutants. We named these additional mutants palm1-3, -4, -5, and -6 (Table S3). Sequence analysis indicates that palm1-3 carries a 26-bp deletion between positions 243 and 269, and palm1-4, palm1-5, and palm1-6 carry tobacco Tnt1 retrotransposons at positions 114, 302, and 583, respectively, within the coding region of ORF3 (Table S3). Second, introducing the WT ORF3 locus, including 2.718-kb 5′-flanking sequence, 0.756-kb ORF, and 1.028-kb 3′-downstream sequence, into the palm1-1 mutant rescued the mutant phenotype (Fig. 2E). Based on these results, we conclude that PALM1 corresponds to ORF3, an intron-less gene that encodes a small protein of 251 aa.

Fig. 2.

Map-based cloning and characterization of PALM1. (A) palm1 was mapped to contig 77 of chromosome 5 closely linked to the CR932963-SSR1 marker. Top shows markers that cosegregate with palm1; bottom shows number of recombinants. (B) Bacterial artificial chromosome clones in the region. (C) Deletion borders identified in palm1-1 and palm1-2 mutants using chromosomal walking. (D) Eight ORFs annotated within the deleted region in palm1-1 and palm1-2 alleles. Solid boxes/vertical lines denoting exons and horizontal lines denoting introns. (E) Introducing a plant transformation vector pCAMBIA3300-PALM1 that contains 2.718-kb 5′-flanking sequence, 0.756-kb ORF, and 1.028-kb 3′-downstream sequence of PALM1 into the palm1-1 mutant rescued the mutant phenotype. (F) Sequence alignments of conserved C₂H₂ zinc finger domain and EAR transcription repressor domain of PALM1 with its homologues from alfalfa (M. sativa), L. japonicus, soybean (G. max), and A. thaliana. (G-J) RNA in situ hybridization. (G and H) Representative longitudinal sections of compound leaf primordia in 3-wk-old vegetative shoot apices used to detect PALM1 transcripts in LL primordia (arrows; G and H) at the P₂ stage. (I) A longitudinal section of a vegetative shoot apex showing PALM1 transcripts in LL. (J) A sense probe used as a negative control (arrows indicate LL). (K–M) Nuclear localization of 35S::GFP-PALM1 fusion protein detected in onion epidermal cells. (N–P) Nuclei stained by DAPI. (Q–S) Subcellular localization of 35S::GFP as a control. (K, N, and Q) Laser confocal images of GFP; (L, O, and R) DIC images of onion epidermal cells; and (M, O, and S) merged images. (Scale bars, 1 cm in F; 100 μm in G–S.)

Sequence comparison indicates that PALM1 and its close homologues from other plant species share syntenic chromosomal locations and are highly conserved in the EPF-type Cys(2)His(2) zinc finger DNA-binding domain at their N-termini (14) and in the EAR repressor domain identified in the class II ERF transcriptional repressors at their C-termini (15) (Fig. 2F and Fig. S4). Furthermore, PALM1 and its homologues from closely related legume species such as alfalfa (M. sativa), Lotus japonicus, and soybean (Glycine max) share a higher degree of sequence similarity than those from more distantly related species such as A. thaliana (Fig. 2F and Fig. S4). However, the role of these homologous genes is still not yet known.

Expression Pattern of PALM1 and Subcellular Localization of the Encoded Protein.

The tissue-specific expression of PALM1 was analyzed by ways of in silico expression (16) and RNA in situ hybridization. The microarray-based expression analysis indicates that PALM1 transcripts are expressed in vegetative shoot buds, leaves, and developing seeds, but remain low or hardly detectable in other tissues including roots, petioles, stems, flowers, pods, and the seed coat (Fig. S5). RNA in situ hybridization using a series of longitudinal sections of vegetative shoot apices shows that PALM1 transcripts were detected in the LL primordia as early as the P₂ stage (Fig. 2 G–I; arrows). PALM1 transcripts remained low or were barely detected in other tissues including SAM, TL, and stipules (Fig. 2 G–I). A sense probe, serving as a negative control, did not give any hybridization signals (Fig. 2J). Subcellular localization prediction, using Plant-PLoc (http://www.csbio.sjtu.edu.cn/cgi-bin/PlantPLoc.cgi), suggests that PALM1 is likely localized to nuclei. To verify this, we transiently expressed a GFP-PALM1 fusion protein driven by the constitutive Cauliflower Mosaic Virus 35S promoter in onion epidermal cells. Fig. 2 K–S shows that the fusion protein was specifically localized to nuclei, consistent with its predicted role as a transcription factor.

PALM1 Negatively Regulates SGL1 Expression.

Previously, we had shown that loss-of-function mutations in the M. truncatula FLO/LFY/UNI orthologue SGL1 completely abolished the initiation of LL primordia at the P₂ stage, resulting in simple leaves (5). SGL1 is expressed in both SAM and entire leaf primordia (5), the latter of which is partially overlapping with PALM1 (Fig. 2 G–I). However, SGL1 expression is greatly reduced in expanding leaflets (5). We hypothesized that SGL1 may be required for the proliferation of LL in the palm1 mutant. Quantitative RT-PCR data reveal that the SGL1 transcript level was increased by 2.7-fold in vegetative shoot apices in the palm1-1 mutant compared with WT (Fig. 3A). To further test whether the increase in the SGL1 expression is simply caused by an increase in the number of leaflet primordia or an alteration in the expression pattern in the palm1 mutant, we compared the expression of the SGL1pro::uidA (GUS) reporter gene in WT and in palm1-1 mutant. Fig. 3B shows that in palm1-1 mutant plants the SGL1pro::uidA reporter gene was expressed in all five leaflets, and its expression remained at a high level in expanding leaflets. In contrast, as previously reported, the same reporter gene was expressed in only the SAM and young leaflets, and its expression was greatly reduced in expanding leaflets in WT (5) (Fig. 3C). These results indicate that the loss-of-function mutation in PALM1 up-regulated and expanded the spatial-temporal expression of SGL1, a positive regulator of leaflet initiation in M. truncatula (5).

Fig. 3.

PALM1 and SGL1 interaction. (A) Quantitative RT-PCR analysis of the SGL1 expression relative to the expression of MtActin2. The level of SGL1 transcripts was increased by 2.7-fold in palm1-1 mutant compared with WT plants 3 wk after germination. Shown are means ± SE (n = 3). (B) GUS activity staining of 3-wk-old M. truncatula palm1-1 mutant harboring the SGL1pro::uidA (GUS) reporter gene. GUS was expressed in the SAM and young and expanding leaflets. (C) The same reporter gene was expressed in the SAM and young leaflets, but not in expanding leaflets (arrow) in WT. (D) Compound leaf phenotype of palm1-3 sgl1-1 double mutants. Shown from left to right are mature compound leaves of WT, palm1-3, sgl1-1, and palm1-3 sgl1-1 mutants (all in the R108 ecotype). Leaves of the double mutants were simple, similar to those of the sgl1 mutant, indicating an epistatic interaction between sgl1 and palm1. (E) The SGL1pro::uidA reporter gene expressed in the SAM and rosette leaves of 3-wk-old WT A. thaliana plants. (F) The expression of the reporter gene was greatly suppressed in 35S::PALM1 plants. (G) Quantitative RT-PCR analysis of the reporter gene expression. The level of GUS transcripts was greatly reduced in 35S::PALM1 SGL1pro::GUS plants compared with SGL1pro::GUS plants. Shown are means ± SE (n = 3). (H) Schematic drawing of the SGL1 promoter including F1, F2, and F3 sequences upstream from the translation initiation codon and a series of deletion fragments of the F2 sequence. (I) EMSA of the F2 deletion sequences labeled with biotin in the presence and absence of purified His-tagged PALM1. (J) EMSA of the F2-3 sequence. Tested were the unlabeled F2-3 sequence in 10-, 20-, and 50-fold excess relative to the biotin-labeled sequence as specific competitors (indicated by +, ++, and +++), unlabeled F3-1 sequence in 50-fold excess as a nonspecific (NS) competitor, and His-tagged tobacco etch virus protease (TEV), an unrelated protein as a negative control. Arrows indicate shifted bands. (Scale bars, 1 cm in B–D.)

To genetically test the involvement of SGL1 in the proliferation of LL primordia in the palm1 mutant, we generated palm1-3 sgl1-1 double mutants (single mutant alleles all from the R108 ecotype). All leaves that developed in the double mutants were simple, similar to those in the sgl1 single mutant (Fig. 3D), indicating an epistatic interaction between sgl1 and palm1. The genetic interaction data support the requirement of SGL1 in the proliferation of LL primordia in the loss-of-function palm1 mutant.

To further elucidate potential mechanisms that underlie PALM1 regulation of SGL1 expression, we took several approaches. We first ectopically expressed the M. truncatula PALM1 gene under control of the constitutive 35S promoter in A. thaliana (Col-0) plants and then introduced the transgene (35S::PALM1) into the plant, through genetic crosses, that carries the SGL1pro::GUS reporter gene (5). Ectopic expression of PALM1 did not affect the simple leaf morphology and flower development of the transgenic plants (Fig. 3 E and F), but it almost completely abolished the SGL1pro::GUS gene expression in leaves as indicated by quantitative RT-PCR and GUS staining data (Fig. 3 E–G), indicating that ectopic expression of PALM1 suppresses the SGL1 promoter activity in A. thaliana leaves.

Next, we used an electrophoretic mobility shift assay (EMSA) to determine the ability of PALM1 to bind the SGL1 5′-flanking sequence. The results indicate that PALM1 bound only sequences within the nucleotide region between −354 and −747 upstream of the translation initiation codon (Fig. S6). In addition, the interaction was abolished in the presence of high molar ratios of unlabeled specific competitor or the ion chelator EDTA (Fig. S6). Furthermore, by using a series of deletions, we narrowed the region in the SGL1 5′-flanking sequence that interacts with PALM1 to a 152-bp sequence between nucleotides −596 and −747 (Fig. 3 H and I). We further show that this interaction is specific, because (i) it was out-competed by unlabeled specific sequence, but not by a nonspecific sequence from a different region of the promoter; (ii) this interaction was not due to the His tag present in the fusion protein; and (iii) a smaller 80-bp deletion sequence lost the binding activity (Fig. 3J). Collectively, these results indicate that PALM1 may negatively regulate SGL1 expression by directly binding its promoter sequence.

PALM1 Antagonizes the KNOXI Protein, KNAT1, in A. thaliana.

In tomato plants, KNOXI genes are initially down-regulated at the incipient sites of leaf primordia (i.e., P₀) at the periphery of the SAM and subsequently reactivated in the developing leaf primordia to promote indeterminacy for compound leaf development (17–20). Depending on the developmental context, ectopic expression of TKNs, tomato KNOXI genes, has different effects on leaf shape, supporting a role for TKNs in stage-specific suppression of leaf maturation in tomato (20). The KNOXI protein, kn1, has been postulated to play a role in the establishment of the proximal-distal polarity in maize (Zea mays) leaves (21). In A. thaliana, a plant with simple leaves, ectopic expression of an A. thaliana KNOXI gene, KNAT1/BP, leads to excessive lobing of leaf margins and uneven growth of laminae (20, 22–24) (Fig. 4A). To test the ability of PALM1 to suppress the effects of overexpression of KNAT1, we made double transgenic lines, through genetic crossing, that ectopically express both PALM1and KNAT1 (35S::PALM1 35S::KNAT1). We show that both leaf lobing and lamina outgrowth were completely abolished in the double transgenic lines (Fig. 4 B and C). Quantitative RT-PCR data show that the KNAT1 transcript level was only slightly reduced in the double transgenic lines compared with the 35S::KNAT1 lines, in line with the transgene being driven by the constitutive 35S promoter. However, these results suggest that PALM1 may suppress the effects of overexpression of KNAT1 by regulating its downstream targets, instead of its transcription, in A. thaliana (Fig. 4D). Although KNOXI proteins are not detected in compound leaves in the IRLC legumes (6), these results are reminiscent of the previous observation that compound leaf development in IRLC legumes can still respond to ectopic expression of KNOXI genes (6) and suggest that PALM1 is capable of regulating leaf morphogenetic processes that are sensitive to the KNOXI regulation.

Fig. 4.

PALM1 and KNAT1 interaction. (A) Rosette leaves of 35S::KNAT1Arabidopsis thaliana plants. The fifth and older leaves were deeply lobed and outgrown. (B) Rosette leaves of 35S::KNAT1 35S::PALM1 A. thaliana plants. Ectopic expression of PALM1 completely suppressed KNAT1-induced lobing of leaf margins and the uneven outgrowth of the laminae. (C) Cauline leaves of 35S::KNAT1 plants (two on the left) and 35S::KNAT1 35S::PALM1 plants (two on the right). Similarly, KNAT1-induced leaf lobing and outgrowth of cauline leaves were suppressed by ectopic expression of PALM1. (D) Quantitative RT-PCR analysis of KNAT1 expression relative to the level of AtEF1a expression. The level of KNAT1 transcripts was only slightly reduced by the ectopic expression of PALM1. (Scale bars, 1 cm.)

Discussion

Mature leaves in M. truncatula, an IRLC legume, are dissected with three leaflets at the tip. Previous studies have shown that the initiation of two LL primordia is controlled by the M. truncatula LFY/UNI orthologue SGL1 (5) (Fig. 3D). In this study, we show that the M. truncatula PALM1 gene encodes a Cys(2)His(2) zinc finger transcription factor and is required to maintain the trifoliate morphology of mature leaves. Several striking phenotypic changes in loss-of-function palm1 mutants, development of two extra leaflets in a basal position, development of the rachis structure on two distally oriented LLs and alteration of the petiole and rachis length, suggest that PALM1 suppresses the morphogenetic activity in developing leaf primordia and serves as a determinacy factor for leaf morphogenesis in M. truncatula (Fig. 1).

Cys(2)His(2) zinc finger transcription factors are one of the largest families of transcription factors in plants (25, 26). Homologous sequences exist in lower land plants, suggesting an ancient origin and divergent function of these transcription factors. A similar situation also exists in mammals, in which a large increase in the diversity of Cys(2)His(2) zinc finger transcription factors is attributed to recent gene duplication and retraction (27). Recently, the tomato LYRATE gene, the orthologue of the Arabidopsis JAGGED (At1G68480) with five exons, has been shown to encode a Cys(2)His(2) zinc finger protein with an EAR domain at the N terminus, a C₂H₂ domain in the middle, and a Pro-rich motif at the C terminus (28). LYRATE appears to play a role in the outgrowth of lateral organs through modulation of the KNOX and auxin transcriptional networks in tomato (28). Outside of the Cys(2)His(2) domain, LYRATE and PALM1 do not share sequence similarities. In addition, LYRATE and PALM1 do not share the same exon-intron structure as PALM1 does not contain any introns. Thus, PALM1 and LYRATE represent two distinct classes of Cys(2)His(2) zinc finger transcription factors. However, they both play a role in compound leaf development, albeit in different species.

Our results indicate that PALM1 binds a specific sequence in the promoter and negatively regulates the transcription of SGL1. Whereas SGL1 is expressed in the SAM and the entire young leaf primordia (5), the expression of PALM1 in LL primordia partly overlaps with that of SGL1. Consistently, the role of PALM1 in the regulation of SGL1 expression and leaf morphogenesis is more pronounced at late stages of leaf development as indicated by the up-regulation and expansion of SGL1pro::GUS reporter gene expression in expanding leaflets in the palm1 mutant compared with WT (Fig. 3 B and C) and along the proximal-distal axis of leaves as indicated by the altered petiole and rachis length and the ectopic formation of rachis on LLs in the palm1 mutants (Fig. 1). Our results support a model in which the negative regulator, PALM1, through its own spatial-temporal expression, defines the spatial-temporal expression of SGL1 and the associated morphogenetic activity in leaf primordia, and through this regulation determines the trifoliate morphology of mature leaves (Fig. 5 A–D). In loss-of-function palm1 mutants, the lack of the negative regulation resulting from loss of PALM1 results in the up-regulation and expansion of SGL1 expression and an increase in the morphogenetic activity, which leads to the development of extra leaflets at a basal position of leaves, ectopic formation of the rachis structure on the distally oriented LLs, and altered development of the proximal-distal axis of leaves (Fig. 5 E and F).

Fig. 5.

Models of compound leaf development in M. truncatula. (A and B) Leaf development in the single leaflet1 (sgl1) mutant. (A) SEM image of leaf primordia [adapted from Wang et al. (5)]. (B) A diagram of interactions between SGL1 and PALMATE-LIKE PENTAFOLIATA1 (PALM1). Loss of function of SGL1 results in lack of formation of LL primordia and simple leaves. (C and D) Leaf development in WT plants. (C) SEM image of leaf primordia development. (D) A diagram of interactions between SGL1 and PALM1. Dashed arrow indicates an epistatic genetic interaction between SGL1 and PALM1. Distinct and partially overlapping spatial-temporal expression of SGL1 and PALM1 in leaf primordia and transcriptional repression of SGL1 by PALM1 define the morphogenetic activity at the leaf margin and the development of LL primordia and the trifoliate leaves. (E and F) Leaf development in the palm1 mutant. (E) SEM image of leaf primordia development. (F) A diagram of interactions between SGL1 and PALM1. Mutations in the PALM1 gene up-regulate and expand the spatial-temporal expression of SGL1 and result in an increased morphogenetic activity at the leaf margin and the development of LLd and LLp and pentafoliate leaves.

Taken together, our studies identify PALM1 as a key regulator of dissected leaf morphogenesis in M. truncatula, an IRLC legume. Our analysis further shows that PALM1 homologues exist in non-IRLC legumes including soybean and L. japonicus (Fig. 2F and Fig. S4), in which KNOXI proteins are expressed in leaves and likely associated with compound leaf development in these plants (6). Although functions of the PALM1 homologues in these non-IRLC legumes are not yet known, it is tempting to speculate that they may be involved in compound leaf development in these plants through an antagonistic interaction with the KNOXI-mediated morphogenetic processes as suggested by the antagonistic interaction between PALM1 and KNAT1 in a heterologous system (Fig. 4). Alternately, PALM1 homologues may have coevolved with the FLO/LFY orthologues in the IRLC legumes in which they regulate the FLO/LFY-type transcription factors during compound leaf development (6).

Methods

Plant Materials.

Seeds of M. truncatula cv. Jemalong A17 (i.e., WT) were exposed to fast neutron radiation at a dosage level of 40 Gy and germinated in a greenhouse with a controlled environment. Approximately 30,000 M₂ plants derived from 5,000 M₁ lines were screened, resulting in the isolation of palm1-1 (M469) and palm1-2 (M534). Additional alleles in the M. truncatula R108 ecotype, palm1-3 (GKB483), and palm1-4 (NF1271), palm1-5 (NF227), and palm1-6 (NF5022) were isolated from a collection of 1,400 T₂ T-DNA mutants (29) and a collection of 6,000 R₂ tobacco Tnt1 retrotransposon mutants (30), respectively. M. truncatula sgl1-1 mutant and SGL1::uidA transgenic lines (5) and A. thaliana SGL1::uidA (5) and 35S::KNAT1 (23) transgenic lines were as previously described. The palm1-1 and palm1-3 lines were backcrossed to their respective parental lines for three generations. BC₃ lines were used to characterize growth phenotypes.

Genetic Mapping.

We generated F₂ mapping populations derived from crosses between palm1-1 and M. truncatula cv. Jemalong A20 (SI Methods). The PALM1 locus was identified using bulked segregant analysis, fine genetic mapping, and chromosomal walking (Table S1 and Figs. S2 and S3).

SEM.

Shoot apices of 2- to 4-wk-old seedlings were subjected to vacuum infiltration in a fixative solution (5% formaldehyde, 5% acetic acid, 50% ethanol) for 30 min and then kept at room temperature overnight. SEM was carried out as described previously (5).

Subcellular Localization.

Subcellular localization of PALM1 was determined using the coding sequence of GFP fused in-frame to the 5′-end of the coding sequence of PALM1 and transcribed from the Cauliflower mosaic virus 35S promoter. The resulting plasmid was bombarded into onion epidermal cells using a helium Biolistic device (PDS-1000; Bio-Rad). The GFP-PALM1 fusion protein was examined using a confocal laser scanning microscope (TCS SP2 AOBS; Leica). Primer sequences are listed in Table S4.

RNA in Situ Hybridization.

RNA in situ hybridization was performed as previously described (31) with minor modifications. The PALM1 probes correspond to a 300-bp sequence in the 3′-region of the PALM1 mRNA. Ten-micrometer sections from shoot apices of 2- to 4-wk-old seedlings were processed and hybridized with digoxigenin-labeled sense and antisense probes.

Complementation and Ectopic Expression.

For complementation studies, a genomic fragment, including 2.718-kb 5′-flanking sequence, 0.756-kb ORF, and 1.028-kb 3′-downstream sequence of PALM1 was amplified by PCR and cloned into pGEM-T Easy vector (Promega). After sequence verification, the insert was digested with EcoRI and XbaI and subcloned into pCAMBIA3300. The resulting plasmid was introduced into Agrobacterium tumefaciens EHA105 strain by electroporation. For ectopic expression, a 756-bp genomic fragment containing the entire coding region of PALM1 was amplified by PCR, digested with NcoI and BstEII and cloned into pCAMBIA3301. The resulting plasmid was introduced into A. tumefaciens EHA105 and GV3101 strains, and used to transform M. truncatula and A. thaliana, respectively. Primer sequences are listed in Table S4.

Stable Plant Transformation.

M. truncatula and A. thaliana (Col-0) were transformed as previously described (5).

Quantitative RT-PCR.

Total RNA samples were isolated from tissues using an RNeasy Plant Mini Kit (Qiagen). The quality of the RNA samples was determined by a Nanodrop Analyzer (BioMedical Solutions). Reverse transcription and cDNA synthesis were carried out with 2 μg of total RNA, using an Omniscript RT Kit (Qiagen) and oligo(dT)₁₅ columns. Real-time RT-PCR analysis was carried out as previously described (32). M. truncatula Actin2 and A. thaliana EF1a were used as internal controls. Primer sequences are listed in Table S4.

Escherichia coli Expression and EMSA.

His(6)-tagged PALM1 was expressed in E. coli BL21 strain using pET32a vector and purified with the QIAexpressionist kit, following the manufacturer’s instructions (Qiagen). EMSA was carried out with a Light Shift Chemiluminescent EMSA kit, following the manufacturer’s instruction (Pierce). Briefly, we used 200 ng purified recombinant protein and 20 fmol biotin-labeled DNA fragment in a 20 μL reaction mix containing 10 mM Tris (pH 7.5), 50 mM KCl, 1 mM DTT, 2.5% glycerol, 0.05% Nonidet P-40, 5 mM MgCl₂, 0.5 mM EDTA, 5 ng/mL poly(dI•dC), and unlabeled DNA fragment at various molar ratios as competitors. Primer sequences are listed in Table S4.