LEAFY COTYLEDON1-LIKE Defines a Class of Regulators Essential for Embryo Development

Abstract

Arabidopsis LEAFY COTYLEDON1 (LEC1) is a critical regulator required for normal development during the early and late phases of embryogenesis that is sufficient to induce embryonic development in vegetative cells. LEC1 encodes a HAP3 subunit of the CCAAT binding transcription factor. We show that the 10 Arabidopsis HAP3 (AHAP3) subunits can be divided into two classes based on sequence identity in their central, conserved B domain. LEC1 and its most closely related subunit, LEC1-LIKE (L1L), constitute LEC1-type AHAP3 subunits, whereas the remaining AHAP3 subunits are designated non-LEC1-type. Similar to LEC1, L1L is expressed primarily during seed development. However, suppression of L1L gene expression induced defects in embryo development that differed from those of lec1 mutants, suggesting that LEC1 and L1L play unique roles in embryogenesis. We show that L1L expressed under the control of DNA sequences flanking the LEC1 gene suppressed genetically the lec1 mutation, suggesting that the LEC1-type B domains of L1L and LEC1 are critical for their function in embryogenesis. Our results also suggest that LEC1-type HAP3 subunits arose from a common origin uniquely in plants. Thus, L1L, an essential regulator of embryo development, defines a unique class of plant HAP3 subunits.

INTRODUCTION

The single-celled zygote of a flowering plant undergoes a series of controlled cell divisions and cell differentiation events that lead to the formation of a mature, multicellular embryo that is metabolically quiescent and desiccated. Early in embryogenesis, during the morphogenesis phase, the plant body is formed through the establishment of the shoot-root axis and the formation of the embryonic tissue and organ systems (West and Harada, 1993; Goldberg et al., 1994; Laux and Jurgens, 1997; Jurgens, 2001). Later, during the seed maturation phase, the embryo acquires the ability to withstand desiccation, accumulates storage macromolecules such as lipids and proteins, and becomes metabolically quiescent as a result of desiccation (reviewed by Bewley, 1997; Harada, 1997). Once environmental conditions are favorable, the seed germinates and the vegetative phase of the life cycle begins.

Genetic studies have identified regulatory genes that play critical roles in embryogenesis during either the morphogenesis or the maturation phases. For example, genes such as WUSCHEL, SHOOTMERISTEMLESS, SCARECROW, and SHORT ROOT (Dilaurenzio et al., 1996; Long et al., 1996; Mayer et al., 1998; Helariutta et al., 2000) have been shown to be essential for the formation of the shoot and root apical meristems that define the embryonic axis of developing Arabidopsis embryos. A different class of genes, including ABSCISIC ACID INSENSITIVE3 (ABI3), ABI4, and ABI5, play important roles during the maturation phase of embryogenesis (Giraudat et al., 1992; Finkelstein et al., 1998; Finkelstein and Lynch, 2000), preparing the embryo for desiccation and postgerminative growth.

Another set of genes encoding Arabidopsis LEAFY COTYLEDON (LEC) proteins, LEC1, LEC2, and FUSCA3, are unique in that they are the only known embryonic regulators required for normal development during both the morphogenesis and maturation phases (reviewed by Harada, 2001). For example, LEC1 is required to maintain suspensor cell fate, to specify cotyledon identity in the early morphogenesis phase, and to initiate and/or maintain the maturation phase and inhibit precocious germination late in embryogenesis (Meinke, 1992; Meinke et al., 1994; West et al., 1994; Parcy et al., 1997; Lotan et al., 1998; Vicient et al., 2000). Furthermore, ectopic postembryonic expression of LEC1 is sufficient to confer embryonic characteristics to seedlings and to induce somatic embryo formation from vegetative cells (Lotan et al., 1998). Because LEC1 is required for normal development both early and late during embryogenesis and is sufficient to confer embryogenic competence to vegetative cells, it is a central regulator that acts far upstream in the regulatory hierarchy that controls embryogenesis. We speculate that LEC1 establishes a cellular environment that promotes embryo development and that this environment coordinates the early and late phases of embryogenesis in flowering plants (Lotan et al., 1998). A major goal of our research is to understand, at a mechanistic level, how LEC1 establishes competence to initiate embryo development.

Given the key role of LEC1 in the control of embryogenesis, we asked if genes related to LEC1 also encode embryonic regulators. LEC1 shares significant sequence similarity with the HAP3 subunit of CCAAT binding factor (CBF, also known as NF-Y; Lotan et al., 1998). CBFs are eukaryotic transcriptional activators that serve diverse roles in dif-ferent organisms (Li et al., 1992). In yeast, CBF activates a set of genes involved in mitochondrial respiration (Guarente et al., 1984; Keng and Guarente, 1987; Trueblood et al., 1988; Schneider and Guarente, 1991), whereas mammalian CBFs are thought to act generally to enhance transcription rates, often in combination with other proteins (reviewed by Maity and De Crombrugghe, 1998; Mantovani, 1999). The transcription factor is a hetero-oligomeric complex consisting of at least three subunits, HAP2, HAP3, and HAP5, although yeast possesses a fourth subunit, HAP4 (reviewed by Maity and De Crombrugghe, 1998; Mantovani, 1999). HAP3 subunits are recognized by their central B domain, an ∼90–amino acid region of the protein that is conserved across eukaryotic organisms. For example, the LEC1 B domain has 57 and 62% sequence identity with HAP3 subunits from yeast and mammals, respectively. Thus, LEC1 appears to encode a subunit of a transcription factor that regulates the expression of genes required for embryo development.

We used the LEC1 polypeptide sequence to identify other genes encoding Arabidopsis HAP3 (AHAP3) subunits. We show that the subunit most closely related to LEC1, designated LEC1-LIKE (L1L), is required for normal embryo development. L1L and LEC1 have distinct functions in embryogenesis, but L1L can substitute functionally for LEC1 when expressed ectopically. Comparison of the deduced amino acid sequences of L1L and LEC1 identified specific amino acid sequences that appear to be required for the function of these proteins in regulating embryo identity and development.

RESULTS

Arabidopsis HAP3 Proteins Are Encoded by a Gene Family That Can Be Divided into Two Classes

We used the amino acid sequence of LEC1 as a query to identify related Arabidopsis polypeptides. Database searches of the sequenced Arabidopsis genome showed that there are nine genes encoding proteins that share significant sequence identity and that the gene encoding L1L (At5g47670) is related most closely to LEC1. As shown in Figure 1  

Figure 1.

Analysis of Arabidopsis HAP3 Subunits.

Amino acid sequence alignment of AHAP3 proteins. Residues highlighted in black and gray represent identical and similar amino acids, respectively. B-domain residues shared between LEC1 (At1g21970) and L1L (At5g47670) but not with the other proteins are highlighted in red. The B domain is underlined.

, sequence similarity among the 10 proteins is limited primarily to the central B domain, consistent with comparisons of HAP3 subunits from other organisms (Li et al., 1992). Only two of these putative proteins, At2g37060 and At3g53340, display sequence identity with each other in the N-terminal A domain or the C-terminal C domain. Because the B domain has been shown to underlie HAP3 function in other organisms (Xing et al., 1993; Kim et al., 1996; Sinha et al., 1996) and because all of these Arabidopsis proteins possess residues that are conserved among HAP3 proteins, we conclude that there are 10 AHAP3 subunits.

Close examination of B-domain sequence alignments showed that L1L and LEC1 define a distinct class of AHAP3 subunits. The two proteins share 83% sequence identity with each other but only 52 to 71% identity with the other eight AHAP3 subunits. Furthermore, L1L and LEC1 share the amino acid residues highlighted in red in Figure 1that differ from the residues that are conserved in the other eight AHAP3 subunits. On the basis of sequence identity within the B domain, we define two classes of AHAP3 subunits: the LEC1-type and the non-LEC1-type. Thus, L1L is the AHAP3 most closely related to LEC1, opening the possibility that L1L also may be an embryonic regulator.

L1L RNA Accumulates Primarily in Developing Embryos

We analyzed L1L gene expression to obtain clues about its role during development. L1L RNA was detected in RNA gel blot hybridization experiments in developing siliques but not in vegetative organs or in flowers, as shown in Figure 2A  

Figure 2.

L1L RNA Is Detected Predominantly in Developing Siliques.

(A) Analysis of L1L and LEC1 RNA levels with RNA gel blot hybridization experiments. Each lane contained 1 μg of poly(A) RNA from siliques with zygote- to early-globular-stage seeds (S1), siliques with globular- to heart-stage seeds (S2), siliques with torpedo- to bent-cotyledon-stage seeds (S3), siliques with mature green seeds (S4), 2-day-old seedlings (Sl), mature rosette leaves (Le), 3-week-old seedling roots (Ro), stems (St), and unopened floral buds and inflorescences (Fl). Control represents the accumulation of a ribosomal protein RNA.

(B) RT-PCR amplification of L1L RNA. Abbreviations are as in (A) with the following additions: ND, no DNA; GD, wild-type genomic DNA; and lec1-1, mutant siliques with torpedo- to bent-cotyledon-stage embryos.

. This pattern of RNA accumulation closely resembled that of LEC1, which accumulates specifically in seeds and differed substantially from those of non-LEC1-type AHAP3 subunits, whose RNAs are not limited to seed development (Edwards et al., 1998; Lotan et al., 1998; Gusmaroli et al., 2001; M. Kim and J.J. Harada, unpublished results). However, differences between L1L and LEC1 RNA accumulation patterns were observed. Figure 2Ashows that L1L RNA levels peaked at a later stage of embryogenesis than did LEC1 levels. Furthermore, sensitive reverse transcriptase–mediated (RT) PCR amplification experiments indicated that L1L RNA is present in all vegetative organs, although presumably at low levels, as shown in Figure 2B. By contrast, LEC1 RNA was not detected in vegetative organs in parallel experiments. These results suggest that L1L is likely to function primarily during seed development and that it is expressed differently from LEC1.

We localized L1L RNA using in situ hybridization experiments to determine where it functions in developing siliques. As shown in Figures 3A to 3C and 3E to 3G  

Figure 3.

In Situ Detection of L1L RNA in Developing Embryos.

Wild-type seed sections were hybridized with an L1L-specific antisense probe. All sections were exposed for 10 days. (A) to (D) and (I) to (L) show bright-field micrographs, and (E) to (H) and (M) to (P) show dark-field micrographs. The sense RNA control did not bind appreciably with the sections. Bars = 50 μm.

(A) and (E) Globular-stage embryo.

(B) and (F) Heart-stage embryo.

(C) and (G) Linear cotyledon-stage embryo.

(D) and (H) Early bent-cotyledon-stage embryo.

(I) and (M) Bent-cotyledon-stage embryo.

(J) and (N) Late bent-cotyledon-stage embryo.

(K) and (O) Mature green-stage embryo.

(L) and (P) Mature yellowing-stage embryo.

, L1L RNA was detected at low but statistically significant levels in the developing embryo proper, suspensor, and endosperm at early stages, including zygotes (data not shown). During the torpedo stage (Figures 3C and 3G) and the linear cotyledon stage (Figures 3D and 3H), L1L RNA became prevalent primarily in the outer cell layers of the embryo, similar to the distribution of LEC1 RNA (Lotan et al., 1998). L1L RNA became evenly distributed throughout the embryo at a high level by the bent-cotyledon stage (Figures 3I, 3J, 3M, and 3N) and was present at low levels in mature-stage embryos (Figures 3K, 3L, 3O, and 3P). This temporal pattern of RNA accumulation corresponds with the results from RNA gel blot analyses. Sense RNA did not bind appreciably with the sections, showing the specificity of the hybridization reactions (data not shown). Together, the RNA accumulation patterns suggest a role for L1L in embryogenesis.

L1L Is Required for Embryogenesis

Given that L1L is expressed primarily during embryogenesis, we used RNA interference (RNAi) experiments to determine if the suppression of L1L RNA levels affected embryo development. Because L1L shares substantial identity with other AHAP3 subunits in the central B domain, L1L-specific nucleotide sequences encoding the C domain were used for targeted suppression (see Methods). Wild-type plants were transformed with the L1L RNAi construct under the control of the 35S promoter, and transgenic plants were recovered. Thirteen of 172 T1 transgenic lines produced defective T2 seeds. More specifically, T1 plants from three independently derived lines segregated 30.1% (n = 1372), 31.9% (n = 1156), and 21.7% (n = 1327) defective T2 seeds. Although the RNAi construct was incompletely penetrant, these results suggest that L1L is required for embryo development. We also found that 4 of 15 lines containing a 35S:L1L transgene segregated defective seeds. Together, these results suggest that cosuppression of L1L gene expression induces defects in embryogenesis (see below).

As shown in Figures 4B to 4D  

Figure 4.

RNAi Suppression of L1L Gene Expression Induces Embryo Defects.

(A) Seed with a wild-type embryo at the bent-cotyledon stage. The seed was cleared and viewed with Nomarski optics.

(B) to (D) Cleared seeds containing defective embryos from lines containing the L1L RNAi constructs. Progeny segregating with a wild-type phenotype in the same silique were at the bent-cotyledon stage.

(E) to (H)  L1L RNA accumulation in defective embryos. Sections were hybridized with an antisense L1L probe and exposed for 10 days.

(E) and (F) Bright- and dark-field micrographs of a defective embryo from a line containing the L1L RNAi construct.

(G) and (H) Bright- and dark-field micrographs of a wild-type embryo at the mature green stage.

Bars = 50 μm in (A), (D), (E), and (H) and 25 μm in (B) and (C).

, RNAi mutants arrested at a number of different embryonic stages with a range of morphological phenotypes. Some embryos arrested at the globular stage (Figures 4B and 4C) but had extra cells in the suspensor. Other mutants arrested at later embryonic stages and had reduced cotyledons (Figure 4D). Seeds containing these defective embryos did not germinate, nor did immature seeds collected before desiccation germinate in culture. However, the effects appeared to be limited to seed development, because no defects in the vegetative development of the RNAi transgenic lines were detected. Suppression of L1L gene expression induced embryonic defects that differed from those caused by the lec1 mutation (Lotan et al., 1998).

To confirm that defects in embryo development resulted from the silencing of the L1L gene, we analyzed L1L RNA levels in transgenic lines using in situ hybridization experiments with L1L-specific probes that excluded sequences encoding the C domain. L1L RNA was not detected at significant levels in 60 of 62 embryos with a mutant phenotype (Figures 4E and 4F) from four independent transgenic lines and was present only at a low level compared with the wild type in the 2 other mutant embryos. Of embryos that segregated with a wild-type phenotype, 31% (n = 352) possessed high levels of L1L RNA, similar to embryos with a wild-type genotype (Figures 4G and 4H), whereas the remainder had only intermediate levels. We interpret these results to indicate that very low levels of L1L RNA do not support embryo development but intermediate levels are sufficient for normal embryogenesis. The incomplete penetrance and variable expressivity of RNAi suppression of L1L gene expression probably allowed us to recover viable progeny containing the transgene. We also demonstrated the specificity of gene silencing by showing that LEC1, oleosin, and cruciferin storage protein RNAs were detected in RNAi embryos exhibiting a mutant phenotype as they were in wild-type embryos (data not shown). Together, these data suggest strongly that L1L is essential for embryo development.

Ectopically Expressed L1L Can Function in Place of LEC1

To examine the functional relationship between L1L and LEC1, we asked if L1L could suppress the lec1 mutation when expressed ectopically. L1L RNA was detected in lec1-1 null mutants (Figure 2B), indicating that the endogenous L1L gene cannot substitute completely for the LEC1 gene. Although the spatial distribution of L1L RNA was similar to that of LEC1, there were temporal differences in accumulation during embryogenesis (Figures 2and 3) (Lotan et al., 1998). Therefore, we fused the L1L coding region with 1997 and 774 bp of sequence 5′ and 3′, respectively, of the LEC1 coding region and transferred the chimeric gene into lec1-1 null mutants. The lec1 mutation causes embryos to become intolerant of desiccation, and no mutant seeds germinate (Meinke, 1992; West et al., 1994; Lotan et al., 1998). As shown in Table 1  

Table 1.

Transgene Suppression of the lec1-1 Mutation

Genea	Total Seeds Screened	Desiccation- Tolerant Seedsb	Percentage of Viable Seedsc
LEC1	13,600	82	0.60
L1L	10,885	71	0.65
At4g14540	12,800	0	0
At3g53340	12,800	7	0.05

^a

Genes were fused with the LEC1 promoter and terminator and transferred into lec1-1 mutants.

^b

Number of seeds that germinated after drying for 2 weeks at 28°C.

^c

Percentage of viable seeds reflects both transformation efficiency and the ability of the construct to suppress the lec1 mutation.

and Figure 5A  

Figure 5.

Suppression of the lec1 Mutation by L1L.

(A) A representative lec1-1 seedling containing the LEC1:L1L:LEC1 transgene that has survived seed desiccation. The transgene confers desiccation tolerance to lec1 mutant embryos.

(B) A lec1-1 seedling transformed with 35S:L1L. The transgene allows lec1-1 mutant embryos to withstand desiccation and confers embryonic characteristics to the seedling, including a lack of cotyledon expansion, failure of hypocotyls and roots to extend, and production of cotyledon-like organs at the positions of leaves.

(C) and (D) Hybridization of cruciferin storage protein and oleosin probes, respectively, with embryo-like 35S:L1L seedlings. Sections were exposed for 2 days.

Bars = 300 μm in (A) and (B) and 100 μm in (C) and (D).

, transgenic lec1-1 mutant seeds that were dried extensively (see Methods) produced viable seedlings, indicating that L1L expressed under the control of LEC1 flanking DNA sequences was able to rescue the desiccation intolerance of lec1 mutants. Moreover, no embryonic or postembryonic abnormalities were detected in lec1 mutant plants containing the L1L transgene, and T1 plants segregated progeny with wild-type and mutant phenotypes at ratios indicating the presence of one, two, or multiple transgenes (data not shown). By contrast, the expression of two genes encoding non-LEC1-type AHAP3 subunits, At4g14540 and At3g53340, under the control of LEC1 5′ and 3′ flanking sequences did not rescue the desiccation intolerance of the lec1 mutants significantly (Table 1). Together, these results suggest that L1L but not non-LEC1-type AHAP3 genes can function in place of LEC1 when expressed ectopically.

We also showed that L1L reproduced effects caused by LEC1 when both were expressed postembryonically. We fused the L1L coding region with the 35S promoter from Cauliflower mosaic virus (Odell et al., 1985) and transferred the construct into lec1-1 mutants. Plants containing 35S:L1L produced viable seedlings, indicating that the transgene could rescue the desiccation intolerance of lec1 mutants. However, the seedlings did not resemble wild-type. Rather, as shown in Figure 5B, transgenic seedlings developed thick and fleshy cotyledons and hypocotyls. Between 25 and 59% of the transgenic seedlings developed multiple pairs of fleshy, cotyledon-like structures at positions normally occupied by leaves, whereas the other seedlings remained arrested in their development, with only two cotyledons. Seedlings with a wild-type genotype transformed with the 35S:L1L construct displayed similar phenotypes (data not shown). These morphological defects reproduced those observed in transgenic seedlings overexpressing LEC1 (Lotan et al., 1998), although we did not detect somatic embryos on 35S:L1L seedlings.

We used in situ hybridization experiments to determine if these fleshy seedlings express embryonic programs of development. As shown in Figures 5C and 5D, RNAs encoding cruciferin storage protein and oleosin lipid body protein, which normally accumulate specifically during embryo development, were detected in these transgenic seedlings. Thus, ectopic expression of L1L reproduces the effects of LEC1 overexpression by creating an environment sufficient to induce embryonic characteristics in vegetative organs. By contrast, expression of a non-LEC1-type AHAP3, At4g14540, under the control of the 35S promoter did not induce detectable developmental abnormalities (data not shown). Together, these results demonstrate that L1L can confer embryonic characteristics to seedlings and support the idea that the LEC1-type B domain may underlie L1L and LEC1 function in embryogenesis.

L1L Genes Are Present in Other Plants

Because L1L and LEC1 both are required for embryo development, we determined the extent to which LEC1-type AHAP3 subunits are present in other organisms. As part of a study to obtain ESTs from the embryo proper of globular-stage scarlet runner bean embryos at 6 days after pollination (A.Q. Bui, K. Weterings, and R.B. Goldberg, unpublished results), cDNA clones encoding a HAP3 subunit with a LEC1-type B domain were identified. Because sequence analysis revealed that the predicted protein shared 94% sequence identity with the Arabidopsis L1L B domain and only 85% identity with that from LEC1, we named the cDNA PcL1L.

To obtain clues about PcL1L function, we examined its temporal and spatial patterns of RNA accumulation. Figure 6A  

Figure 6.

PcL1L RNA Is Present Primarily in Developing Seeds.

(A) Gel blot analysis of PcL1L RNA accumulation. Twenty-five micrograms of total RNA was analyzed from leaves (Le), stems (St), 2-week-old seedling leaves (Sl), 2-week-old seedling roots (SlRo), 2-week-old seedling stems (SlSt), inflorescences (In), ovules (Ov), 2-DAP seeds (Se I), 4- to 5-DAP seeds (Se II), 6-DAP seeds (Se III), 12- to 14-DAP embryos (Em I), and 19- to 21-DAP embryos (Em II).

(B) RT-PCR analysis of PcL1L RNA accumulation. Each lane corresponds to the RNA gel blot sample in (A).

(C) to (F) Distribution of PcL1L RNA. Sections were hybridized with a PcL1L antisense probe (C) to (E) or a sense RNA control (F). (C) and (D) were exposed for 4 days, whereas (E) was exposed for 47 days.

(C) Preglobular-stage seed. PcL1L RNA is high in the embryo proper and suspensor.

(D) Globular-stage seed. PcL1L RNA is at its highest levels in outer tissue layers of the embryo.

(E) Unfertilized ovule. PcL1L RNA is present at low levels throughout the ovule.

(F) Unfertilized ovule that does not bind sense RNA probe.

Bars = 100 μm.

shows that PcL1L RNA was detected only in developing embryos by RNA gel blot analysis. However, RT-PCR analysis showed that PcL1L RNA was present, presumably at low levels, at all developmental stages tested, including in vegetative organs, inflorescences, and unfertilized ovules (Figure 6B). Thus, the pattern of PcL1L RNA accumulation is more similar to that of L1L than that of LEC1 in Arabidopsis (Figure 2). In situ hybridization analysis showed that PcL1L RNA accumulated at high levels in the embryo proper and suspensor of preglobular-stage (5 days after pollination [DAP]) and globular-stage (7 DAP) embryos but at very low levels in the endosperm of the seeds and integuments of the unfertilized ovules (Figures 6C to 6E). PcL1L RNA was present at the highest levels in the epidermal and subepidermal layers of the embryo proper at the globular stage (Figure 6D). This pattern is similar to the distribution of Arabidopsis L1L RNA observed at the torpedo and bent-cotyledon stages (Figures 3C and 3D). The finding that scarlet runner bean appears to possess a protein with a LEC1-type B domain and that RNA encoding this protein accumulates with a pattern similar to that of Arabidopsis L1L suggests that the two L1L genes are orthologous.

Through database searches, we extended this analysis by identifying 12 other HAP3 subunits from other plants with conserved amino acid residues characteristic of the LEC1-type B domain. Most were identified from embryo or seed-derived cDNA libraries, although sequences from pine pollen cone and lotus root nodule cDNA libraries were identified. As shown in Figure 7A  

Figure 7.

Identification of L1L Proteins from Other Plants.

(A) Amino acid sequence alignment of the B domains of plant L1L proteins. Conserved amino acid residues are highlighted in gray, and residues unique to L1L proteins are highlighted in black. Accession numbers are given at the end of Methods.

(B) Phylogenetic relationships between L1L and non-LEC1-type- HAP3 subunits. The cladogram illustrates the most parsimonious consensus pattern of relationships obtained using maximum parsimony analysis. Bootstrap values generated with 1000 replicates are indicated before the nodes. Nodes with bootstrap scores of <50% are not shown. The high bootstrap values provide strong support for the monophyletic L1L clade.

, alignment of the B domains of these 15 plant HAP3 subunits revealed 17 amino acid residues that are shared between LEC1-type B domains but that differ from residues conserved in non-LEC1-type B domains. By contrast, no conserved amino acid residues were detected in the A and C domains of these 15 proteins, although some similarities were observed in these regions between L1L, PcL1L, and some of the other plant L1L proteins (data not shown). The phylogenetic tree obtained by maximum parsimony analysis (PAUP 4.0) shown in Figure 7Bindicates that the L1L proteins constitute a well-supported, monophyletic clade. Based on their sequence identity within the B domain, their origin in seed RNA populations, and the ability of L1L to suppress the lec1 mutation when expressed ectopically, we speculate that some of these L1L genes also may play roles in embryogenesis.

DISCUSSION

L1L and LEC1 Constitute a Subclass of HAP3 Subunits

L1L and LEC1 display substantial sequence identity with HAP3 subunits of CBFs. Although a CBF from plants has not been isolated, several lines of evidence suggest that L1L and LEC1 function as part of a CBF that regulates embryogenesis. First, in addition to HAP3, paralogs of the other two CBF subunits required for DNA binding activity, HAP2 and HAP5, have been identified in plants (Li et al., 1992; Albani and Robert, 1995; Edwards et al., 1998; Kusnetsov et al., 1999; Gusmaroli et al., 2001). Unlike yeast and mammals, which possess single genes for each subunit, plants possess families of subunits. For example, there are 6, 10, and 8 genes encoding the AHAP2, AHAP3, and AHAP5 subunits, respectively, in Arabidopsis, opening the possibility that different combinations of subunits may regulate diverse processes (Edwards et al., 1998; Gusmaroli et al., 2001; M. Kim, H. Lee, R.W. Kwong, and J.J. Harada, unpublished results). Second, an AHAP2 gene has been shown to complement a yeast HAP2 mutation, indicating that the Arabidopsis protein can function in a CBF (Edwards et al., 1998). Third, an AHAP5 subunit has been shown to interact with other nuclear proteins, presumably AHAP2 and AHAP3, to form a complex that binds a double-stranded oligonucleotide containing a CAAT box (Kusnetsov et al., 1999). Fourth, loss-of-function mutations of L1L and LEC1, two HAP3 paralogs, have severe consequences on plant development, suggesting that these subunits play essential roles.

Our results show that there are at least two distinct classes of AHAP3 subunits that differ in several respects. LEC1-type and non-LEC1-type HAP3 subunits differ by 16 amino acid residues that serve as signatures of their B domains (Figure 1) (Gusmaroli et al., 2001). Phylogenetic analysis suggests that HAP3 subunits possessing these signature residues have a common evolutionary origin (Figure 7). Residues at corresponding positions of yeast and mammalian HAP3 subunits are more similar to non-LEC1-type than to LEC1-type AHAP3 subunits. This finding opens the possibility that L1L and LEC1 represent novel HAP3 subunits of CBF. Next, sequence diversity between the two types of subunits appears to underlie the functional differences, because L1L but not two other genes that encode non-LEC1-type AHAP3 subunits suppressed the lec1 mutation when expressed under the control of LEC1 flanking sequences (Table 1). Similarly, a non-LEC1-type AHAP3 did not induce embryonic characteristics in seedlings when fused with the 35S promoter, as did L1L and LEC1 (Table 1, Figure 5). Finally, genes encoding L1L and LEC1 are expressed predominately or exclusively during seed development (Figure 2) (Lotan et al., 1998), whereas the non-LEC1-type AHAP3 genes generally are expressed at high levels in nonembryonic tissues (Edwards et al., 1998; Gusmaroli et al., 2001; M. Kim and J.J. Harada, unpublished results). In this regard, LEC1, L1L, and PcL1L exhibit similar spatial patterns of RNA accumulation in developing embryos (Figures 3and 6) (Lotan et al., 1998).

The B Domain Underlies the Function of LEC1-Type HAP3 Subunits in Embryogenesis

We present strong evidence that the B domain of LEC1-type HAP3 subunits underlies their function in embryogenesis. Sequence similarity between L1L and LEC1 was observed exclusively in the B domain and not in the A and C domains (Figure 1). Suppression of the lec1 mutation by L1L (Figure 5, Table 1) suggests that some or all of the 16 residues unique to LEC1-type B domains account for the ability of L1L to substitute functionally for LEC1 when expressed ectopically. Most HAP3 subunits from other plants that possess LEC1-type B domains (Figure 7) are present in embryos or seeds, consistent with the expression patterns of L1L and LEC1. This finding opens the possibility that other L1Ls play important roles in seed development. However, two L1Ls are present in pollen cones and root nodules, suggesting that the LEC1-type HAP3 subunit may function at other developmental stages. This class of HAP3 subunit is present in gymnosperms and monocotyledonous and dicotyledonous angiosperms, but it has not been detected in nonplant organisms, suggesting that the LEC1-type B domain evolved uniquely in plants.

The central B domain of HAP3 subunits serves critical roles in CBF function. Studies of yeast and mammalian HAP3s show that the B domain contains amino acid residues that account for its ability to interact with HAP2 and HAP5 subunits and for the CBF to bind DNA (Xing et al., 1993; Kim et al., 1996; Sinha et al., 1996). One explanation for the differences in the activities of L1L and LEC1 versus non-LEC1-type AHAP3 subunits is that LEC1-type B domains may mediate interactions with specific AHAP2 and AHAP5 subunits to form a CBF that activates the genes required for embryo development. Because AHAP2 and AHAP5 are encoded by six and eight genes, respectively, and most are expressed in nonseed tissues (Gusmaroli et al., 2001; H. Lee, M. Kim, and J.J. Harada, unpublished results), the possibility exists that defined combinations of AHAP subunits confer specific functions to the transcription complex. However, this interpretation requires that the specific AHAP2 and AHAP5 subunits present in embryos also are present in vegetative tissues, because ectopic expression of L1L and LEC1 confers embryonic characteristics to vegetative tissues (Figure 5) (Lotan et al., 1998).

A second possibility is that the LEC1-type B domain may recruit other transcription factors to the CBF that confer unique specificity to the complex. CBFs have been shown to interact with other transcription factors to activate specific sets of genes. For example, activation of the MHC class II gene promoter requires the binding of CBF and an X-box binding factor, and activation of the 3-hydroxy-3-methylglutaryl-CoA synthase gene requires the binding of both CBF and sterol regulatory element binding proteins (Wright et al., 1994; Linhoff et al., 1997; Dooley et al., 1998). In addition, no Arabidopsis paralog of the HAP4 subunit that provides a transcriptional activation domain to the yeast CBF has been identified, and transcriptional activation domains are not apparent in the HAP2 and HAP5 subunits as they are in their mammalian counterparts (Forsburg and Guarente, 1989; Coustry et al., 1996). Thus, a protein with transcriptional activation function may be recruited to the complex by L1L and LEC1.

A final alternative is that B domain residues unique to LEC1-type HAP3 subunits may confer a novel DNA binding specificity to the CBF that differs from that afforded by non-LEC1-type AHAP3 subunits. Thus, CBFs containing L1L and LEC1 would bind and modulate the transcription of genes required for embryo development, whereas non-LEC1-type AHAP3s would not. This is the simplest alternative, because there is no need to invoke novel interactions with other proteins. However, to our knowledge, no HAP3 subunit has been identified that alters the binding specificity of CBFs in other organisms.

L1L and LEC1 Have Distinct Functions during Embryogenesis

Although our studies have shown that L1L can substitute for LEC1 if expressed ectopically (Table 1, Figure 5), other evidence suggests that L1L and LEC1 normally have distinct functions during embryogenesis. The first and most compelling argument is that monogenic, loss-of-function mutations in either L1L or LEC1 cause defects in embryo development. These results show that the endogenous genes cannot substitute for one another, although we cannot exclude the possibility that the genes have partial overlaps in function. Consistent with this interpretation is the finding that the suppression of L1L and LEC1 gene expression induces different embryonic phenotypes. lec1 mutants arrest at a late stage of embryo development, with complete though misshapen cotyledons and embryonic axes, and mutant embryos can be rescued before desiccation to produce viable seedlings (reviewed by Harada, 2001). By contrast, RNAi suppression of L1L caused embryos to arrest in their development as early as the globular stage, and mutant embryos cannot be rescued to produce postembryonic plants (Figure 4). We conclude that although L1L clearly is required for embryo development, it appears to play a fundamentally different role in embryogenesis than LEC1.

With regard to the L1L mutant phenotype, RNAi suppression of L1L gene expression is characterized by incomplete penetrance and variable expressivity. Not all embryos containing the construct exhibit a mutant phenotype, and those that do arrest at a number of different embryonic stages with a variety of defects (Figure 4). However, transgenic embryos displaying a mutant phenotype possessed low to undetectable levels of L1L RNA, whereas those with a wild-type phenotype had intermediate to high RNA levels (Figure 4). Therefore, defects in embryo development appear to result from the suppression of L1L expression. Although we have not yet identified an insertional mutation of L1L, we note the possibility that the RNAi suppression of L1L may not produce a mutant phenotype as severe as that of a genetic null mutation. The RNAi construct is controlled by the 35S promoter, and we have shown that this promoter does not become active detectably during embryogenesis until the globular stage (J. Pelletier and J.J. Harada, unpublished results). Thus, L1L RNA may accumulate early during embryogenesis in RNAi lines, albeit at a very low level, and decline only after the globular stage. Despite these qualifications, it is unlikely that a null l1l mutant would share similar characteristics with the lec class of mutants, because l1l mutant embryos arrest earlier in embryogenesis than do lec1 mutants.

Three other lines of evidence support the conclusion that L1L and LEC1 have distinct endogenous functions. First, L1L RNA accumulates later in embryogenesis than does LEC1 RNA (Figure 2). Second, L1L RNA is present in developing seeds and at low levels in vegetative organs, whereas LEC1 RNA is detected only in developing seeds (Lotan et al., 1998) (Figure 2). Third, lec1 mutants display an abnormal phenotype even though L1L RNA is detected in the mutant seeds, indicating that the endogenous L1L gene is not sufficient to completely prevent defects induced by the lec1 mutation (Figure 2).

There are several potential explanations to reconcile the findings that the endogenous L1L and LEC1 genes do not act redundantly, yet L1L can be made to substitute functionally for LEC1. One hypothesis is that the specific pattern of LEC1 gene expression is critical for its function. Although the distribution of LEC1 and L1L RNAs in embryos is similar, there are differences in the timing of their accumulation (Figures 2and 3) (Lotan et al., 1998). Similar situations have been described for two Arabidopsis MYB genes, WEREWOLF and GLABROUS1. Genes encoding these functionally equivalent proteins play different roles in plant development because they are transcribed in distinct cell types (Lee and Schiefelbein, 2001). Alternatively, increased dosage of the L1L gene and, by inference, increased L1L RNA levels in transgenic lec1 mutants containing the LEC1:L1L:LEC1 transgene may account for the suppression of the mutation. Dosage suppression has been described in microorganisms (Puziss et al., 1994). A third possibility is that because the accumulation of LEC1 and L1L RNA does not differ substantially, RNA sequences in the LEC1 5′ and/or 3′ untranslated regions, which are included in the LEC1:L1L:LEC1 gene (see Methods), may regulate LEC1 function at the translational level. Additional information is needed to distinguish between these possibilities.

In conclusion, we have shown that L1L, the AHAP3 subunit most closely related to LEC1, is a regulator of embryo development. L1L is expressed predominately during embryo development, and it is required for the completion of embryogenesis. The ability of L1L but not non-LEC1-type AHAP3 subunits to function in place of LEC1 when expressed ectopically implicates the B domain as the region of L1L and LEC1 that is critical for their function. Mutagenesis studies are needed to define which of the 16 amino acid residues of LEC1-type AHAP3 subunits differentiate their functions from non-LEC1-type subunits. Although L1L can function redundantly with LEC1 when expressed ectopically, the two subunits have distinct functions during embryogenesis. Thus, L1L is a novel regulatory protein that plays an essential role during embryogenesis.

METHODS

Plant Materials and Manipulations

lec1-1 mutants and wild-type plants of Arabidopsis thaliana (ecotype Wassilewskija) were grown as described previously (West et al., 1994). Seeds of the day-neutral scarlet runner bean (Phaseolus coccineus cv Hammond's Dwarf Red Flower) were grown in the greenhouse as described by Weterings et al. (2001). Seeds were germinated in vermiculite to obtain seedlings. Flowers were pollinated and collected at specific days after pollination (DAP) as described previously (Weterings et al., 2001).

Approximately 500 unfertilized runner bean ovules were collected from young, open flowers. Approximately 90, 66, and 50 seeds were collected from 2-DAP, 4- to 5-DAP, and 6-DAP pods, respectively. Approximately 100 cotyledon-stage embryos were isolated from seeds of 12- to 14-DAP and 19- to 21-DAP pods. Seed and embryo stages were according to Weterings et al. (2001). Small young leaves, stems, and inflorescences were collected from lateral branches of flowering plants. True leaves, roots, and stems were collected from 2-week-old seedlings. Upon collection, tissues were frozen immediately in liquid nitrogen and stored at −80°C until use.

Agrobacterium tumefaciens strain GV3101 containing transformation constructs was infiltrated into lec1-1 and wild-type plants (Bechtold et al., 1993). Seeds from T0 plants were germinated on medium containing 60 μg/mL glufosinate ammonium to select for transgenic plants (Finale; AgrEvo Environmental Health, Montvale, NJ). Plant genotypes were verified in PCR amplification experiments.

Isolation and Preparation of cDNA and Genomic Clones

PCR was used to amplify the genomic fragments containing AHAP3 genes. Primers that flanked the putative L1L open reading frame with the addition of BamHI and XbaI sites for subcloning purposes were used (BAMMNJ7-5, 5′-AGGATCCATGGAACGTGGAGGCTTCCAT-3′; and 3-MNJ7XBA, 5′-ATCTAGATCAGTACTTATGTTGTTGAGTCG-3′). The AHAP3 genes At4g14540 (3-224) and At3g53340 (3-180) were amplified using primer combinations 3-224-F/3-224-R (5′-CCTATC-TCGAGATGGCGGATTCGGACAACGATTC-3′/5′-CCCGGTCTAGAT-TAAGAAAAATGATGGGAAAATTGATGTCC-3′) and AH3-180-F/AH3- 180-R (5′-CCCGGGGAGATCTATGGCGGATACGCCTTCGAGCCC-AGC-3′/5′-GGGCCCCTAGGCTTTTACCAGCTCGGCATTTCTTCA-CC-3′), respectively. Nucleotide sequences of the genomic clones were verified.

L1L, At4g14540, and At3g53340 genomic clones were inserted between the LEC1 promoter and terminator within the plant transformation vector BJ49 (Gleave, 1992). The LEC1 promoter/terminator cassette consists of 1992 bp of DNA 5′ of the LEC1 translation start codon plus 770 bp 3′ of the LEC1 stop codon (H. Lee and J.J. Harada, unpublished results). The L1L gene was fused with the 35S promoter from Cauliflower mosaic virus and the octopine synthase terminator of the plasmid pART7 and transferred into the binary transformation vector pMLBART (Gleave, 1992).

cDNA clone pPCEP112 was identified from a scarlet runner bean cDNA library by EST sequencing analysis. This cDNA library was constructed with total RNA isolated from the embryo proper of 6-DAP seeds using the SMART PCR cDNA Library Construction Kit (Clontech, Palo Alto, CA) (A.Q. Bui, K. Weterings, and R.B. Goldberg, unpublished results).

Protein Sequence Analysis

Amino acid sequences were aligned with the PileUp program (Seqweb version 2.0.2; Accelrys, Burlington, MA), and alignments were prepared with BOXSHADE (http://www.ch.EMBnet.org). Database searches were performed with the LEC1 protein sequence as a query (http://www.ncbi.nlm.nih.gov/blast/ and http://www.arabidopsis.org/Blast/ [as of August 5, 2002]). The analysis identified the following AHAP3 genes: At2g47810, At1g09030, At2g37060, At3g53340, At2g38880, At5g47640, At4g14540, At2g13570, and At5g47670 [the last of which we renamed LEC1-LIKE].

Parsimony trees of the B domains of HAP3 subunits were generated with CLUSTAL X (version 1.8; Thompson et al., 1997) and the heuristic search algorithm of the PAUP program (version 4.0 beta; Swofford et al., 1996). One hundred replicates were used for weighted analysis in generated consensus parsimony trees. For maximum parsimony analysis, 1000 iterations were used to create bootstrap percentages.

RNA Analyses

Arabidopsis RNA was isolated as described previously (Stone et al., 2001). Total RNA from scarlet runner bean was isolated using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA) and treated with RNase-free DNase I (Boehringer Mannheim, Indianapolis, IN) according to the protocol of Ausubel et al. (1995). Approximately 25 μg of RNA was loaded on a formaldehyde gel, and RNA gel blot analysis was performed as described previously (Harada et al., 1988). In situ hybridization experiments with Arabidopsis and scarlet runner bean tissues were performed as described previously (Dietrich et al., 1989; Weterings et al., 2001).

The presence of L1L RNA in an organ system was assessed using nonquantitative reverse transcriptase–mediated PCR analysis. For Arabidopsis, first-strand cDNA was generated from 5 μg of each RNA in a 20-μL reaction volume using the Thermal Script reverse transcriptase system (Invitrogen, Carlsbad, CA). One microliter of each reaction was amplified in a 20-μL reaction volume according to the manufacturer's specifications using primers for L1L (see above), LEC1 (LP/UP, 5′-GACATACAACACTTTTCCTTAAAG-3′/5′-CAGCAA-CAACCCACCCCCAATG-3′), and a ribosomal protein gene (TIN1/TIN2, 5′-TTTGGTGGATGCCCCTGATA-3′/5′-TAATTTCCGAATCCA-AAATC-3′) (T. Lotan and J.J. Harada, unpublished results). Amplification products were fractionated by agarose gel electrophoresis.

For scarlet runner bean RNA, first-strand cDNA was generated from 2 μg of each total RNA in a 20-μL reaction using Superscript II Reverse Transcriptase according to the manufacturer's specifications (Gibco BRL, Rockville, MD). PCR amplification was performed using the primers PcL1L-F (5′-AGATTCTTCCTCCACATGCCAAGAT-3′) and PcL1L-R (5′-CCTTAATCCCATCCATCCCCTTAAT-3′) with 2 μL of each reverse transcriptase reaction in a 50-μL reaction volume.

RNA Interference Suppression of L1L

The primer combination 3LEFTXX (5′-TCTAGACTCGAGCTTAGCTGCAGTGCTGGG-3′) and 3RIGHTBAM (5′-GGATCCTTGAACCAAGACGCATTACG-3′) was used to amplify a 500-bp fragment unique to the C domain of L1L. The fragment was placed in both orientations into the RNA interference vector pRNA69, which contains the 35S promoter (J.F. Emery and J.L. Bowman, unpublished results). This construct then was placed into the pMLBART binary vector.

Upon request, all novel materials described in this article will be made available in a timely manner for noncommercial research purposes.

Accession Numbers

Accession numbers for the CCAAT binding factor HAP3 subunits shown in Figure 7are as follows: LEC1, AF036684; L1L, AY138461; PcL1L, AF533650; barley (A), AL506199 and AL509098; wheat, AY058921; pine A, AW754604; pine B, AW981729; Argemone, AY058920; rice, AU088581; maize, AF410176; barley B, BE603222; Vernonia, AY058919; soybean A, AY058917; soybean B, AY058918; and lotus, AW719547 and AW720671.

ACKNOWLEDGMENTS

We thank John Emery and John Bowman for the gift of pRNA69, Keith Lowe and Bill Gordon-Kamm for sharing unpublished data, Kook-Hyun Chung and Neelima Sinha for help with the phylogenetic analysis, the ABRC for seeds, and Abeba Kiros, Diana Lee, Julie Pelletier, and Kelly Matsudaira Yee for technical assistance. This work was supported by grants to R.B.G. and J.J.H. from the U.S. Department of Energy and Ceres, Inc.

Author notes

¹

These two authors contributed equally to this work.

2

To whom correspondence should be addressed. E-mail ; fax [phone]

Article, publication date, and citation information can be found at www.plantcell.org/cgi/doi/10.1105/tpc.006973.