Recessive mutations that suppress the late-flowering phenotype conferred by FRIGIDA (FRI) and FLOWERING LOCUS C (FLC) and which also result in serrated leaf morphology were identified in T-DNA and fast-neutron mutant populations. Molecular analysis showed that the mutations are caused by lesions in the gene encoding the large subunit of the nuclear mRNA cap-binding protein, ABH1 (ABA hypersensitive1). The suppression of late flowering is caused by the inability of FRI to increase FLC mRNA levels in the abh1 mutant background. The serrated leaf morphology of abh1 is similar to the serrate (se) mutant and, like abh1, se is also a suppressor of FRI-mediated late flowering although it is a weaker suppressor than abh1. Unlike se, in abh1 the rate of leaf production and the number of juvenile leaves are not altered. The abh1 lesion affects several developmental processes, perhaps because the processing of certain mRNAs in these pathways is more sensitive to loss of cap-binding activity than the majority of cellular mRNAs.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
In many plant species, flowering is promoted by the sensing of environmental cues that accompany seasonal changes, such as changes in day-length or temperature. Within a given species, there can be variation in the response to environmental cues. In Arabidopsis thaliana, this variation has been well characterized. In rapid-flowering (summer-annual) accessions of Arabidopsis the major environmental cue for flowering is day-length; in these accessions the promotion of flowering occurs through a photoperiod pathway that perceives long days as inductive (Koornneef et al., 1998a,b; Mouradov et al., 2002; Simpson and Dean, 2002). At a molecular level, this pathway operates through a coincidence mechanism in which the presence of CONSTANS during the light period leads to the activation of genes involved in the initiation of flowering (Suarez-Lopez et al., 2001; Valverde et al., 2004; Yanovsky and Kay, 2002).
There are also winter-annual accessions of Arabidopsis. In these accessions flowering is promoted by prolonged exposure to cold temperature (vernalization) in addition to exposure to long days (for review see Michaels and Amasino, 2000; Simpson and Dean, 2002). Winter annuals typically begin growing in the fall and, after exposure to the vernalizing temperatures of winter, flower early in the spring. A requirement for vernalization in winter annuals ensures that flowering does not occur prematurely.
In winter-annual accessions of Arabidopsis, the block to flowering in plants that have not been vernalized is largely due to the action of two genes, FLOWERING LOCUS C (FLC) and FRIGIDA (FRI) (Clarke and Dean, 1994; Koornneef et al., 1994; Lee et al., 1994). FLC encodes a MADS domain-containing transcription factor which is a repressor of flowering, and FRI acts to increase FLC expression to levels that inhibit flowering (Johanson et al., 2000; Michaels and Amasino, 1999; Sheldon et al., 1999). Vernalization results in the repression of FLC expression, and this FLC repression is mitotically stable in the absence of cold. Rapid-flowering accessions of Arabidopsis appear to have arisen from winter annuals due to allelic variation at FRI and/or FLC. Specifically, many summer-annual types flower rapidly without vernalization because they lack an active allele of FRI and thus have low levels of FLC expression (Johanson et al., 2000; Michaels and Amasino, 2001) or have an allele of FLC that is not upregulated by FRI (Michaels et al., 2003). Three genes have been identified, VERNALIZATION 1 and 2 (VRN1 and VRN2) and VERNALIZATION INSENSITIVE 3 (VIN3), that are required to maintain the vernalization-mediated repression of FLC. VRN1 encodes a B3 domain DNA-binding protein (Levy et al., 2002) and VRN2 encodes a protein related to polycomb-group proteins (Gendall et al., 2001). VIN3 encodes a PHD domain-containing protein that is involved in the initiation of vernalization-mediated modifications to FLC chromatin (Sung and Amasino, 2004).
In addition to the VRN genes, another group of genes that act to repress FLC expression are the autonomous-pathway genes LUMINIDEPENDENS (LD), FLOWERING LOCUS D (FLD), FPA, FVE, FY, and FCA (Koornneef et al., 1998a,b; Michaels and Amasino, 1999; Sheldon et al., 1999). However the autonomous-pathway genes are not likely to be involved in the vernalization response because autonomous-pathway mutants, which are delayed in flowering, are fully vernalization responsive (Michaels and Amasino, 2000). Autonomous-pathway mutations cause rapid-flowering types of Arabidopsis to behave like winter annuals because, although FRI activity is absent, the loss of any autonomous-pathway gene permits FLC to be expressed to levels that strongly inhibit flowering (Michaels and Amasino, 1999; Sheldon et al., 1999).
The first positive regulator of FLC, FRI, was identified in studies of natural variation in flowering time as discussed above. Recently four additional positive regulators of FLC have been identified; mutations in these genes attenuate FLC expression. FRIGIDA-LIKE 1 (FRL1) is a gene related to FRI that is required for FRI to upregulate FLC (Michaels et al., 2004). PHOTOPERIOD INDEPENDENT 1 encodes an ATP-dependent, chromatin-remodeling protein of the SWI2/SNF2 family (Noh and Amasino, 2003). VERNALIZATION-INDEPENDENT 4 (VIP4) encodes a hydrophilic protein with repeated Trp-Asp (WD) motifs which is related to proteins in yeast and animals (Zhang and van Nocker, 2002; Zhang et al., 2003). Mutations in EARLY IN SHORT DAYS4 (ESD4), a gene that encodes a small protease of the SUMO class, cause a modest decrease in FLC levels (Murtas et al., 2003; Reeves et al., 2002).
Additional positive regulators of FLC are likely to exist and thus we initiated a screen for such regulators. In this paper, we report the characterization of several mutations in ABSCISIC ACID HYPERSENSITIVE 1 (ABH1) that suppress the high levels of FLC expression typically found in an FRI-containing, winter-annual genetic background. Thus an abh1 lesion converts a winter-annual type of Arabidopsis into one with rapid-flowering behavior. ABH1 encodes the large subunit of the mRNA cap-binding complex. The loss of ABH1 causes additional phenotypes such as serrated leaves and ABA hypersensitivity as previously reported (Hugouvieux et al., 2001, 2002) and, in non-inductive photoperiods, an increase in the number of cauline leaves (i.e. bracts) on the elongating stem after the initiation of bolting.
A lesion in ABH1 suppresses the late-flowering phenotype conferred by FRI and FLC
The Columbia (Col) accession of Arabidopsis thaliana is rapid flowering because it lacks a functional allele of FRI. Introgression of an active FRI allele from the San-Feliu (Sf-2) accession into Col (to create the Col FRI-Sf2 line as described in Lee et al., 1994) increases expression of FLC and thus converts Col into a winter annual that is severely delayed in flowering in the absence of vernalization. (This Col FRI-Sf2 line will be referred to here as Col FRI.) In a screen of both fast-neutron and T-DNA-mutagenized lines of Col FRI, five mutants that were early-flowering without vernalization, and which exhibited the additional phenotype of serrated leaves, were identified (Figure 1a,b and Table 1). Thus, these mutations cause a conversion of the winter-annual phenotype of Col FRI into that of a rapid-flowering line. Complementation tests (including the F2 generation) revealed that all five mutants were allelic (data not shown) and, as discussed below, the mutants were designated abh1-2, -3, -4, -5, and -6 (Table 1).
Table 1. Leaf number (rosette and cauline leaves formed by the primary meristem) of long-day and short-day grown plants. Each value is the mean ± 1 SD. A minimum of 10 plants from each genotype was analyzed
ND, not determined.
aAll genotypes are in the Col background.
ABH1 FRI (Col FRI)
71.3 ± 1.5
8.6 ± 0.6
18.0 ± 1.4
7.6 ± 0.6
50.5 ± 4.4
19.0 ± 1.1
17.3 ± 1.2
7.0 ± 1.8
52.8 ± 2.6
21.6 ± 2.7
17.3 ± 1.6
7.0 ± 1.0
44.8 ± 6.1
21.5 ± 1.8
14.0 ± 0.9
3.5 ± 1.4
53.0 ± 4.2
23.7 ± 2.1
18.5 ± 0.5
7.8 ± 0.8
53.0 ± 3.4
24.5 ± 0.6
ABH1 fri (Col)
14.6 ± 1.0
3.8 ± 1.0
54.2 ± 3.8
9.2 ± 0.4
12.3 ± 0.8
2.7 ± 0.8
32.2 ± 2.6
15.6 ± 1.6
se ABH1 FRI
27.8 ± 2.5
15 ± 1.3
50.7 ± 3.0
20.3 ± 4.0
A region of DNA flanking the T-DNA insertion of one allele (abh1-5) revealed that the T-DNA was inserted into a gene encoding the large subunit of the mRNA cap-binding protein complex (At2g13540). This gene is present in a single copy in the Arabidopsis genome, and has been previously named ABH1 because the abh1-1 mutant is abscisic acid (ABA) hypersensitive (Hugouvieux et al., 2001, 2002). abh1-2, -5 and -6 were identified in T-DNA mutant populations. abh1-2, which was the allele used for most of the studies described below, did not contain a T-DNA insertion but rather had a G–A transition which destroys a splicing site for the first intron (Figure 2). abh1-5 has a T-DNA insertion in the middle of exon 8 and abh1-6 has a 14 base pair duplication in the last exon, which causes a C-terminal frameshift (Figure 2). abh1-3 and -4 were identified in fast neutron-mutagenized populations and these alleles were not further characterized.
Effect of the abh1 lesion on rosette and cauline leaf number
In long days (LD; 16 h light/8 h dark) the abh1 mutants in Col FRI were much earlier flowering than the parental Col FRI line (Table 1). The mutants initiated the elongation of an inflorescence stem (bolting) with only a slightly higher rosette leaf number compared with the rapid-flowering Col accession. However, the elongated inflorescence stem (bolt) contained many additional nodes at which cauline leaves were formed [cauline leaves can be considered as bracts subtending coflorescence meristems (Ratcliffe et al., 1998)]. The number of cauline leaves formed by the mutants in the Col FRI background in LD was more similar to the Col FRI parent than to Col (Table 1). To determine whether the increase in cauline leaf number was due to an interaction of FRI and abh1, abh1-2 was introduced into Col, which naturally lacks an active FRI allele (Johanson et al., 2000). In the Col background, the abh1 lesion did not increase cauline leaf number; rather the abh1-2 mutant (abh1-2 fri) flowered with slightly fewer rosette and cauline leaves compared with Col (Figure 1c and Table 1).
When the abh1 mutants in Col FRI were grown in non-inductive short days (SD; 8 h light/16 h dark), conditions under which the parental Col FRI line flowers after the primary meristem has formed greater than 100 leaves, the mutants flowered with a primary rosette leaf number similar to Col (Table 1). However, as was the case in LD, the number of cauline leaves formed by the mutants in SD was much greater than the number formed by Col (Table 1). Thus in both LD and SD the abh1 lesion suppresses most of the effect of FRI on rosette leaf number, but does not suppress the effect of FRI on the number of cauline leaves.
In SD in the Col background (i.e. lacking FRI activity), abh1 formed fewer rosette leaves but had more cauline leaves than Col (Table 1). Thus in the Col background in SD, but not in LD, abh1 increases the number of cauline leaves formed independent of FRI, and abh1 affects rosette and cauline leaf numbers in opposite ways. This increase in cauline leaf number results in an inflorescence in which the flowers open later after bolting in abh1 than in Col wild type (Figure 1d). The prolonged cauline leaf-forming phase of abh1 inflorescence development under SD, but not LD, in the Col background is similar to that previously reported for serrate (se), another mutant with serrated leaves (Clarke et al., 1999). SE encodes a zinc-finger protein that may be involved in chromatin modification (Prigge and Wagner, 2001).
Serrate can partially suppress the delayed flowering caused by FRI
The similarity of the phenotypes of abh1 and se mutants led us to determine whether the se mutation could suppress the late-flowering of Col FRI. Accordingly, the se lesion was introduced into Col FRI. In LD, the rosette leaf number of Col FRI is reduced by the presence of se, but this suppression of FRI-mediated late flowering conferred by se mutation is not as strong as the suppression caused by abh1 (Table 1). In addition, the se FRI line produced even more cauline leaves than the Col FRI parent in LD which was not observed in the abh1 alleles in an FRI background. In SD, the se FRI line behaved similar to the abh1 FRI lines with respect to leaf numbers at flowering (Table 1).
Effect of abh1 on leaf emergence and phase change
The se mutant exhibits a slower rate of leaf production (Clarke et al., 1999; Prigge and Wagner, 2001). Thus it was of interest to evaluate this parameter in abh1. In both LD and SD, there was little difference in the rate of leaf emergence in the parental Col FRI line compared with abh1 with or without the presence of FRI (Figure 3a,b). In these conditions, se in Col FRI had a slower rate of leaf production similar to previous reports of the effect of se in Col (Clarke et al., 1999; Prigge and Wagner, 2001).
SERRATE is also involved in phase change; in se mutants the number of juvenile leaves (i.e. leaves that lack trichomes on their abaxial surface) is reduced (Clarke et al., 1999; Prigge and Wagner, 2001). Thus, whether phase change was altered by the abh1 lesion was also examined. The abh1 lesion does not affect the number of juvenile leaves formed (Figure 3c). Moreover, the presence of FRI did not affect the number of juvenile leaves formed in wild type, abh1 or se. Thus although the presence of FRI causes a substantial delay in flowering in wild-type Col, FRI does not alter phase change. In addition, leaf phyllotaxy is altered in se (Clarke et al., 1999; Prigge and Wagner, 2001), but it is normal in abh1-2 in the presence or absence of FRI (data not shown).
FLC mRNA levels in an FRI background are reduced by abh1
The delayed flowering of the non-vernalized Col FRI line is due to the upregulation of FLC by FRI (Michaels and Amasino, 1999; Sheldon et al., 1999). Thus it was of interest to determine whether the early-flowering phenotype of abh1 in Col FRI was due to an effect on FLC mRNA levels. Indeed, abh1-2 in Col FRI has reduced FLC mRNA levels when compared with the parental Col FRI line, similar to that of Col (Figure 4); se is also a suppressor of the late-flowering phenotype, and the se FRI line also showed reduced levels of FLC (Figure 4). Thus, the suppression of the late-flowering effect of FRI by abh1 and se appears to be due, at least in part, to the effect of these lesions on FLC expression.
ABA hypersensitivity is observed in all abh1 alleles and in se
Hugouvieux et al. (2001) reported that abh1-1 exhibits ABA hypersensitivity. The additional abh1 alleles presented in this study were also ABA hypersensitive with respect to the inhibition of germination (Figure 5a). Curiously, one allele, abh1-6, which has a C-terminal frameshift (Figure 2) was not as ABA hypersensitive as the other alleles in tests with two different lots of seeds, yet abh1-6 was similar to other abh1 alleles with respect to the suppression of FRI-mediated late flowering. The se mutant also exhibited some degree of ABA hypersensitivity although it was not as pronounced as that of abh1. (Figure 5b). The presence of FRI did not alter the ABA sensitivity of abh1 alleles or se (Figure 5a).
Because abh1 has reduced FLC levels, the ABA germination response of an flc null mutant was also examined (Figure 5a). The flc null behaved similar to wild type; thus the effect of abh1 on ABA sensitivity is independent of its effects on FLC.
In a screen for mutants that cause a winter-annual type of Arabidopsis to exhibit a rapid flowering phenotype, we identified five alleles of abscisic acid hypersensitive 1 (abh1). ABH1 was first identified as a modulator of ABA sensitivity, and it encodes the large subunit of the eukaryotic nuclear mRNA cap-binding complex (Hugouvieux et al., 2001). We find that loss of ABH1 suppresses the late flowering associated with dominant alleles of FRI. Specifically, the abh1 mutation causes a FRI-containing line (Col FRI) to initiate bolting after forming a number of rosette leaves comparable with a line that lacks FRI activity (Col wild type). FRI causes late flowering by increasing FLC expression (Michaels and Amasino, 1999, 2001; Sheldon et al., 1999), and the abh1 lesion suppresses the FRI-mediated increase in FLC mRNA levels. Preliminary results indicate that abh1 does not suppress the late flowering of all autonomous-pathway mutants as effectively as it suppresses FRI-mediated late flowering.
The abh1 suppression of the effects of FRI on rosette leaf number occurs in both inductive (LD) and non-inductive (SD) photoperiods. However in LD most of the abh1 alleles in the FRI background produced almost as many cauline leaves as the FRI-containing parental line, whereas abh1-2 in a Col background (i.e. lacking FRI) did not form an increased number of cauline leaves. Therefore, the abh1 lesion per se does not cause an increased cauline leaf number in LD, but rather the lesion does not suppress the effect of FRI on cauline leaf number. FRI-containing lines produce high numbers of rosette and cauline leaves because FRI increases expression of the floral repressor FLC (Michaels and Amasino, 1999, 2001; Sheldon et al., 1999). This differential effect of abh1 on rosette and cauline leaf production might be due to differential effects on FLC expression at different stages of development and/or in different tissues. It is interesting to note that in SD the abh1-2 mutation in a Col background actually forms more cauline leaves than Col. Thus in SD the loss of ABH1 affects cauline leaf number independently of FRI.
The increase in cauline leaf number of the abh1 mutant in SD, as well as the serrated leaf morphology first described by Hugouvieux et al. (2001), is reminiscent of the phenotype of the serrate (se) mutant (Clarke et al., 1999). Thus, we evaluated whether se could act as a suppressor of FRI-mediated late flowering. Indeed, se partially suppressed the increase in rosette leaf number due to the presence of FRI, although the suppression was not as strong as that of abh1, and, like abh1, se did not reduce cauline leaf numbers in LD. The se mutant exhibits altered phyllotaxy, a slower rate of leaf emergence and a shortening of the juvenile phase (Clarke et al., 1999; Prigge and Wagner, 2001). These alterations are maintained when se is introduced into an FRI-containing line. However, abh1 mutants do not share these characteristics of the se mutant; abh1 mutants exhibit wild-type phyllotaxy, rate of leaf emergence and length of the juvenile phase.
A characteristic of the abh1-1 allele is ABA hypersensitivity (Hugouvieux et al., 2001). We find that the five additional abh1 alleles described in this study are also ABA hypersensitive in a seed germination assay. This ABA hypersensitivity as well as the serrated leaf morphology is independent of the effect of the abh1 lesion on FLC expression because loss of FLC does not affect either phenotype. Interestingly, the se mutant also exhibits ABA hypersensitivity, but the hypersensitivity of se to low levels of ABA (e.g. 0.1–0.2 μm; Figure 5b) is not as strong as that of abh1-2.
Whether the common features of the se and abh1 mutant phenotypes are due to a common molecular mechanism is not known. SE encodes a single C2H2-type zinc-finger protein that has been suggested to regulate in gene expression by modification of chromatin structure (Prigge and Wagner, 2001), whereas ABH1 encodes the large subunit of the mRNA cap-binding complex. However, certain zinc-finger proteins function as RNA-binding proteins (Friesen and Darby, 1998; Iuchi, 2001) so it is possible that a common feature of RNA metabolism is affected in both mutants.
In eukaryotes, transcripts produced by RNA polymerase II possess a 7-methylguanosine cap structure attached by a 5′-5′ phosphotriester linkage to the first encoded nucleotide of the transcript (Shatkin et al., 1976). In the nucleus, this cap can be bound by cap-binding complex (CBC) which is a heterodimer consisting of a small subunit cap-binding protein (CBP20), and a large subunit, CBP80. The presence of CBC bound to the cap is thought to promote pre-mRNA splicing (Ohno et al., 1987) and pre-mRNA 3′ end formation by cleavage and polyadenylation (Flaherty et al., 1997). Yeast mutants that lack the CBC do not have global defects in mRNA cleavage or polyadenylation (i.e. cbc mutants are not lethal), although such mutants are slower growing than wild type (Fortes et al., 1999). In ArabidopsisABH1/CBP80 is a single-copy gene and thus our work and that of Hugouvieux et al. (2001) demonstrate that, similar to yeast, a loss of CBC is not lethal.
The distinct phenotypes of the abh1 lesion in Arabidopsis may result from differential sensitivities of certain mRNAs to the lack of the CBC. Specifically certain mRNAs may not be properly processed in the absence of the CBC and these mRNAs might encode key regulators of certain pathways (such as the those for leaf development, ABA signal transduction, or flowering). Indeed, previous studies in yeast (Fortes et al., 1999) and plants (Hugouvieux et al., 2001) demonstrated that levels of certain mRNAs are more sensitive than others to loss of the CBC. A previous study of abh1 (Hugouvieux et al., 2001) did not reveal a change in FLC mRNA levels or a substantial effect on flowering time in the abh1/cpb80 mutant; that is because this study was performed in the rapid-flowering Col background which lacks FRI activity and thus has low levels of FLC expression. In the FRI-containing background, which has high levels of FLC expression, the effect of the abh1/cpb80 lesion on flowering time and FLC expression becomes apparent.
The mechanism by which loss of the CBC affects FLC mRNA levels is not known. It is possible that the effect is indirect and that the abh1 lesion alters the expression of a key regulator of FLC such as FRI. Alternatively, CBC bound to the FLC mRNA cap may be required for high levels of FLC expression. In animals, the presence of CBC increases the rate of recognition and splicing of the cap-proximal intron (Lewis et al., 1996). FLC contains a large first intron that may render the FLC pre-mRNA more sensitive to loss of CBC. Another possibility is that the specific regulation of FLC mRNA levels involves components that interact with a CBC-bound FLC mRNA or pre-mRNA. Among the regulators of FLC mRNA levels are three proteins with RNA-recognition motifs, FCA, FLK and FPA (Lim et al., 2004; Macknight et al., 1997; Schomburg et al., 2001), and the presence of the CBC may modulate an interaction between the FLC mRNA and these regulators of FLC. Recent studies have also revealed the involvement of RNA-binding proteins in ABA signal transduction (Koiwa et al., 2002; Lu and Fedoroff, 2000). Thus genes in pathways involving certain types of RNA-binding proteins may be most affected by loss of the CBC.
Plant and growth conditions
The Col FRI line was previously described (Lee and Amasino, 1995); se seeds were obtained from ABRC (Ohio State University, CO, USA); fca-9 seeds were kindly provided by Caroline Dean; ld-1 seeds were previously described (Lee et al., 1994); fpa-7 was isolated in our laboratory from a T-DNA population. All genotypes used in this work are in the Columbia (Col) background. Plants were grown under cool-white fluorescent light (100 μmol m−2 sec−1; bulbs from Sylvania, Danvers, MA, USA) at 22 ± 1°C in Fafard Germination Mix (Fafard Co., Agawan, MA, USA) and were fertilized with Dyna-Grow 7-9-5 fertilizer (Dyna-Grow Corp., San Pablo, CA, USA) 2 weeks after planting. Long-days (LDs) consisted of 16 h light and 8 h darkness, and short days (SDs) consisted of 8 h light followed by 16 h darkness.
Measurement of rate of leaf emergence and juvenile phase length
The rate of leaf emergence was evaluated by counting visible leaves at a series of time points throughout development as described by Telfer et al. (1997). Only leaves formed by the primary meristem were counted. Juvenile phase length was measured by counting the number of juvenile leaves which are defined as leaves that do not possess abaxial trichomes (Telfer et al., 1997).
Fast-neutron and T-DNA mutagenesis
Fast-neutron mutagenized lines were obtained as previously described (Michaels and Amasino, 1999). For T-DNA mutagenesis, Col FRI plants were transformed using Agrobacterium tumefaciens containing the binary vector pSKI015 (Weigel et al., 2000) and T2 plants were screened for flowering time.
Cloning of ABH1 gene
The genomic DNA sequence flanking the T-DNA of abh1-5 was obtained by the thermal asymmetrical interlaced (TAIL)-PCR method as previously described (Schomburg et al., 2003). The location of the mutations on abh1 alleles was determined by sequencing PCR products obtained by using ABH1 (At2g13540)-specific gene primers.
RNA gel blot analysis
Total RNA was isolated from 15-day-old seedlings using TRIzol (Invitrogen, Carlsbad, CA, USA). Twenty micrograms of RNA was run in a 1.0% formaldehyde agarose gel (Sambrook et al., 1989). RNA was transferred to a Hybond-N+ membrane according to the manufacturer's instructions. The membrane was probed with 32P-dATP-labeled FLC cDNA fragment that did not contain the MADS-box domain (Michaels and Amasino, 2001). An 18S rRNA probe was used as RNA loading control.
ABA hypersensitivity assay
For seed germination assays on ABA plates, a minimal of two different seed lots were tested for each genotype and three replicates for each seed lot were performed. Approximately 100 seeds of each line were plated on minimal medium (0.25× Murashige and Skoog medium) containing 0.4 μm ABA. Plates were incubated for 4 days at 4°C and transferred to LDs. To avoid ABA breakdown, the light intensity was reduced by covering the plates with two layers of Kimwipes (Kimberly-Clark Co., Roswell, GA, USA). After 4 days, percentage germination was determined.
We are grateful to Michael Prigge for many useful discussions of serrated leaf mutants, to Colleen Bizzell for help in genotyping and to Sibum Sung and Robert Schmitz for help in RNA analyses. We thank the Salk Institute Genome Analysis Laboratory and the ABRC at Ohio State for providing knock-out pools containing alleles of se and abh1. This work was supported by the College of Agricultural and Life Sciences and the Graduate School of the University of Wisconsin, and by the United States Department of Agriculture National Research Initiative Competitive Grants Program and by grant 0133663 to R.M.A. from the National Science Foundation. The creation of insertion mutant lines was supported by National Science Foundation grant 0116945. I.C.B. was supported by a fellowship from CNPq in Brazil.