shygrl1 Is a Mutant Affected in Multiple Aspects of Photomorphogenesis

Abstract

We have used a counter-selection strategy based on aberrant phytochrome regulation of an Lhcb gene to isolate an Arabidopsis mutant designated shygrl1 (shg1). shg1 seedlings have reduced phytochrome-mediated induction of the Lhcb gene family, but normal phytochrome-mediated induction of several other genes, including the rbcS1a gene. Additional phenotypes observed in shg1 plants include reduced chlorophyll in leaves and additional photomorphogenic abnormalities when the seedlings are grown on medium containing sucrose. Mutations in the TATA-proximal region of the Lhcb1*3 promoter that are known to be important for phytochrome regulation affected reporter gene expression in a manner similar to the shg1 mutation. Our results are consistent with the possibility that the mutation either leads to defective chloroplast development or to aberrant phytochrome regulation. They also add to the evidence of complex interactions between light- and sucrose-regulated pathways.

Plants require light not only as an energy source, but also as a cue for growth and development. In the absence of light, plants follow a distinct program of development that is characterized by elongated growth, closed cotyledons, lack of chloroplast development, and low transcriptional activity. Upon exposure to light, seedlings undergo photomorphogenesis, a developmental program that is visually characterized by short hypocotyls and green, open cotyledons. Transcription of nuclear- and chloroplast-encoded genes, many of which have functions related to photosynthesis, is induced as greening commences. Light is perceived by different photoreceptors. Three types that have been characterized are the phytochromes, cryptochromes, and phototropins (Briggs and Huala, 1999; Deng and Quail, 1999).

Components of the phytochrome-signaling pathways have been identified in Arabidopsis using genetic, biochemical, or molecular biological approaches (Neff et al., 2000). PKS1 (Fankhauser et al., 1999) and NDPK2 (Choi et al., 1999) are kinases that interact with phytochromes. CCA1 (Wang et al., 1997), PIF3 (Ni et al., 1998), and HFR1/REP1 (Fairchild et al., 2000; Soh et al., 2000) are transcription factors for which some target genes have been identified (Wang et al., 1997; Martinez-Garcia et al., 2000). FAR1 (Hudson et al., 1999), FIN219 (Hsieh et al., 2000), PAT1 (Bolle et al., 2000), and SPA1 (Hoecker et al., 1998; 1999) have been characterized at the genetic and molecular levels, but their biochemical functions remain a puzzle. Like phytochrome mutants, a number of phytochrome-signaling mutants have long hypocotyls when grown under white, red, or far-red light. Some, such as cr88 (Lin and Cheng, 1997) and ndpk2 (Choi et al., 1999), exhibit a pale-green phenotype in addition to long hypocotyls. In many cases, the mutants display only a subset of the phenotypes associated with loss of a specific phytochrome, indicating the existence of branched pathways. The emerging picture indicates a complex network of pathways, some overlapping and some independent, leading from phytochrome perception to a physiological response (Casal, 2000).

Phytochrome regulation of transcription of the Lhcb gene family may be used as a system to identify additional loci that function in photomorphogenesis (Karlin-Neumann et al., 1991). This nuclear gene family encodes light-harvesting chlorophyll (Chl) a/b-proteins associated with photosystem II, and it is induced within one-half of an hour of phytochrome activation in dark-grown seedlings (Silverthorne and Tobin, 1984). In addition to phytochrome, Lhcb genes are regulated, at both transcriptional and posttranscriptional levels, by such factors as developmental stage, blue light, abscisic acid, cytokinins, plastid development, and circadian rhythms (Flores and Tobin, 1988; Taylor, 1989; Brusslan and Tobin, 1992; Gao and Kaufman, 1994; Weatherwax et al., 1996; Kubasek et al., 1998; Millar, 1999). The Lhcb promoter has been used previously to identify mutants in light and circadian clock regulation (Li et al., 1995; Millar et al., 1995; Lopez-Juez et al., 1998). Various efforts have identified the TATA-proximal region of Lhcb genes to be important for phytochrome regulation of transcription, and two motifs, CCAAT and GATA, have been demonstrated to be crucial to the phyto-chrome-mediated activity of promoter fragments (Anderson et al., 1994; Kehoe et al., 1994; Anderson and Kay, 1995; Degenhardt and Tobin, 1996). In vitro- binding assays showed that many proteins in plant extracts associate with these promoter fragments (Terzaghi and Cashmore, 1995).

To identify components of the phytochrome signal transduction pathways, we have designed a counter-selection strategy that takes advantage of the phytochrome responsiveness of the Lhcb1*3 gene. We used transgenic plants transformed with a 1.3-kb promoter fragment of the Lhcb1*3 gene, which is strongly induced in etiolated seedlings by brief, saturating red illumination (Karlin-Neumann et al., 1988), fused to the coding region of the Nia1*2 nitrate reductase gene. Nitrate reductase (NR), an enzyme that reduces nitrate to nitrite, can also convert the nontoxic compound chlorate to chlorite, which is toxic to plants. We had previously found that the growth of the transgenic line was severely inhibited when grown on chlorate under conditions in which the phytochrome system was activated by intermittent red (IR) light (Heimer et al., 1995). This strategy has led to the identification of a locus that we have designated shygrl1 (shg1) because when mutated, it causes short stature, yellow-green leaves, and reduced accumulation of Lhcb genes.

RESULTS

Isolation of Mutants with Reduced Lhcb1*3::Nia1*2 Expression

Seeds of the transgenic line A4 (Heimer et al., 1995), which were homozygous for the Lhcb1*3::Nia1*2 chimeric gene construct depicted in Figure 1 were mutagenized. Plants grown from these seeds were selfed, and the progeny of these plants (the M2 generation) were used to select potential mutants of interest. Seedlings were germinated and grown under IR on medium with Gln as the nitrogen source and chlorate as the selection substrate. Because endogenous NR genes are not expressed when plants are grown on Gln, this growth condition ensures NR expression predominantly from the introduced construct. Therefore, seedlings that had normal phytochrome-induced expression of NR from the Lhcb promoter fragment were severely growth inhibited. Seedlings that grew to a height greater than 6 mm under these conditions were expected to have impaired transcription of the Lhcb1*3::Nia1*2 construct and low levels of NR activity. These seedlings were transferred to soil and allowed to produce seed. These M3 progeny were then assayed for NR activity. To eliminate mutations that affected endogenous NR directly, lines that had little or no NR activity when grown on nitrate under white light were not studied further.

Figure 1.

The Lhcb1*3::Nia1*2 chimeric gene construct. Part of the Lhcb1*3 (1.3 kb) promoter, which extends through +14 of transcription start, was fused to 3.5 kb of Nia1*2 genomic DNA containing its own transcription termination sequences. Construction of the transformation vector and selection of the stably transformed A4 line was described in detail in Heimer et al. (1995).

Two mutant lines had characteristics expected of a mutant defective in phytochrome signaling. In Table I, their NR activities are compared with the untransformed wild-type (WT) parent Wassilevskija (WS) and the transgenic parental line A4 when grown with Gln or nitrate. One mutant line, 60, had levels of NR activity that were approximately one-third of the parental A4 line when grown on Gln under IR light, which suggests that the phytochrome induction of the introduced transgene was greatly reduced. When grown on nitrate under continuous white light, the NR activity of line 60 was 1.5-fold higher than WS, the parental line of the A4 transgenic plants; thus, the mutant is not likely to be defective in NR proteins or cofactors and most probably is able to produce additional NR from the introduced transgene under these growth conditions. The NR activity of line 60 was 75% of the parental A4 line under continuous white light, suggesting that the mutation led to limited expression of the introduced construct even under continuous light. Line 58, the second mutant line, behaved similarly. Lines 60 and 58 also had a similar appearance, exhibiting short stature and yellow-green stems and leaves. Crosses of each mutant line to the untransformed parent ecotype WS showed that the mutants were recessive, and crosses of lines 60 and 58 to each other showed that they fell into one complementation group that we have designated shg1 (data not shown). Therefore, we renamed lines 60 and 58 shg1-1 and shg1-2, respectively.

Table I.

Nitrate reductase activity of seedlings

Line	Growth Conditions
	Glutamine and IR	Nitrate and CWL
A4	100	100
WS	9.8 (9.3–10.2)	49.7 (48.5–50.8)
58	25.6 (25.0–26.3)	64.0 (63.4–64.6)
60	34.7 (32.7–36.7)	74.8 (71.2–74.5)

Seedlings were grown for 5 d on medium containing glutamine under intermittent red light (IR), or on medium containing nitrate under continuous white light (CWL). We used the transgenic wild-type A4, the untransformed parent of A4 (WS), and the mutant lines 58 and 60. Nitrate reductase activity was measured according to Heimer et al. (1995) and is expressed as the percentage of the activity found in A4 seedlings for each growth condition. The averages of duplicate assays are given. The ranges for each average value are denoted in parentheses.

Phytochrome Regulation of Endogenous Lhcb1*3 Is Reduced in shg1-1 Mutants

To demonstrate that the shg1 mutation affected the transcription of the Lhcb1*3 gene in response to phytochrome, we tested the phytochrome induction of the endogenous Lhcb1*3 gene. Dark-grown seedlings express Lhcb1*3 at a very low level; a 1-min exposure to red light (R) causes an increase in this mRNA level that has previously been shown to be due to increased transcription (Karlin-Neumann et al., 1991). Figure 2 shows that the phytochrome-induced increase of Lhcb1*3 mRNA in shg1-1 was considerably less than in the transgenic WT A4 plants. The initial dark level of this transcript was also lower than that found in A4. These results are consistent with the idea that the mutation affects endogenous Lhcb1*3 promoter activity, and not just the promoter activity of the introduced construct. It also suggests that shg1-1 may be mutated in a component of the phytochrome signal transduction pathway leading to increased Lhcb transcription.

Figure 2.

Time course of Lhcb1*3 mRNA accumulation in the transgenic WT line A4 (WT) and mutant line 60 (shg1-1) seedlings after a 1-min R treatment (R). A, RNase protection assay (RPA) of total RNA from 5-d-old dark-grown seedlings given R and harvested at the indicated time points. The fragments protected by the Lhcb1*3 and ubiquitin (ubq3) riboprobes are indicated. Quantification of Lhcb1*3 mRNA relative to ubq3 mRNA levels is shown in B.

The yellow-green phenotype of shg1-1 cosegregated with the reduced phytochrome induction of Lhcb1*3 mRNA. The phenotypes of WT and shg1-1 are shown in Figure 3, A and B. We grew F₂ plants from the selfed F₁ generation of a backcross between WT and shg1-1. We selected 10 green F₂ plants that yielded 100% green offspring and 10 yellow-green F₂ plants that yielded 100% yellow-green offspring for further analysis. Etiolated F₃ seedlings from these plants were assayed for R-induced Lhcb1*3 mRNA accumulation. In all cases, yellow-green F₂ lines yielded seedlings that had a reduction in R-induced accumulation of Lhcb1*3 mRNA compared with that found in seedlings from the green F₂ plants (data not shown). The appearance of the two traits, yellow-green and reduced phytochrome-mediated Lhcb1*3 mRNA accumulation, in the same lines indicates that they cosegregate, and, unless due to mutation in a closely linked gene, are caused by the same mutation. The yellow-green phenotype of shg1-1 can be observed from the seedling stage. As shown in Figure 3, C through E, the cotyledons and first leaves of shg1-1 are noticeably paler than WT. In addition, all leaves of shg1 are serrated (data not shown). This is in contrast to WT, which develops serrated leaves at a later stage. The flowering time of shg1-1 did not differ noticeably from WT, and it had no marked fertility defect.

Figure 3.

WT (A) and shg1-1 (B) grown for 6 weeks under continuous white light, bar = 1 cm. Cotyledons (C), primary (D), and secondary (E) leaves of WT (left) and shg1-1 (right) seedlings grown for 2 weeks on MS2S plates under continuous white light.

SHG1 Maps to Chromosome 2

We used PCR-based markers and a mapping population of 75 yellow-green F₂ siblings to map shg1 to the middle of chromosome 2, approximately 4 cM from the er marker. This map position places shg1 in the vicinity of hy1, a mutant defective in heme oxygenase 1 (Davis et al., 1999; Muramoto et al., 1999), which is pale green and has reduced Lhcb RNA accumulation (Chory et al., 1989a; Parks and Quail, 1991). Complementation analysis demonstrated that shg1 is not allelic to hy1 (data not shown). None of the cue mutants, which are pale-green mutants that exhibit reduced Lhcb gene expression (Li et al., 1995; Lopez-Juez et al., 1998), map to chromosome 2. Other photomorphogenic mutants such as cr88 and fin219, which have some phenotypes similar to shg1, map to positions distinct from the region to which shg1 map. serrate (se), which is mutated in a putative single 2Cys-2His zinc finger transcription factor (M.J. Prigge and D.R. Wagner, personal communication), map to the vicinity of SHG1. Because shg1 has serrated leaves like se, we crossed the se and shg1-1 mutants to each other. Complementation tests showed that they are not allelic to each other (data not shown). Thus, SHG1 is a new photomorphogenic locus.

shg1-1 Has Reduced Chl Levels

To characterize the pale-green phenotype of shg1, we measured levels of Chl and Chl intermediates in shg1-1. Figure 4A shows that cotyledons of shg1-1 had eight to 10 times lower Chl levels than those found in WT cotyledons. However, leaves from 2-month-old shg1-1 plants grown in the greenhouse showed only a small difference. At the seedling stage, the Chl a/b ratio of shg1-1 was only slightly higher than WT, whereas at the adult stage, the Chl a/b ratios of shg1-1 and WT were similar (Fig. 4B).

Figure 4.

Levels of Chl and Chl intermediates in WT and shg1-1. A, Total Chl levels from cotyledons of seedlings grown for 7 d on MS plates under long-day conditions, and from rosette leaves of plants grown for 2 months on soil in the greenhouse. B, Chl a/b ratios of samples from A. Accumulation of Protoporphyrin IX (ProtoIX), Mg-protoporphyrin IX (Mg-protoIX), and Protochlorophyllide (Pchlide) in WT and shg1-1 seedlings after ∂-aminolevulinic acid (ALA) feeding. A schematic diagram of the Chl synthetic pathway from Glu (Glu) to Pchlide is shown and each graph above ProtoIX (C), Mg-protoIX (D), and Pchlide (E) represents levels of that intermediate. Error bars represent se and representative data from two independent experiments are presented.

The pale phenotype of shg1 suggested that it might be a Chl biosynthesis mutant. ALA feeding of etiolated seedlings showed a 4-fold reduction in protoporphyrin IX, Mg-protoporphyrin IX, and protochlorophyllide, compared with WT (Fig. 4C). We tested for accumulation of ALA and porphobilinogen (PBG) in shg1-1, and found that both of these intermediates were undetectable (data not shown). Previous work had shown that treatment of Chlorella vulgaris with levulinic acid, an inhibitor of ALA dehydratase, resulted in a large increase in ALA, equivalent to the molar loss of Chl (Beale, 1970). An increase of this size would have been easily detected in our assay; thus, we can conclude that neither ALA nor PBG are accumulating in shg1. Together, these results suggest that the pale phenotype of shg1 is a result of reduced flux through the Chl biosynthetic pathway, and not likely to be a result of a leaky mutation in one of the early enzymes of Chl biosynthesis.

Phytochrome Regulation of Other Lhcb Genes, But Not of rbcS1a, CCA1, or ATHB-2, Is Affected in shg1-1

The low levels of Lhcb1*3 mRNA and Chls in the seedling stage prompted us to check whether other members of the Lhcb gene family are similarly affected. Figure 5, A and B, shows the phytochrome-induced increase of Lhcb1*1 and Lhcb1*2 mRNAs in dark-grown WT and shg1-1 seedlings. Because the sequences of these two Lhcb genes are so similar, they give a single band on the RNA gel blot. As seen for Lhcb1*3 mRNA, the phytochrome induction of Lhcb1*1 and Lhcb1*2 mRNAs was reduced greatly. Therefore, different Lhcb genes are regulated similarly in shg1-1.

Figure 5.

The shg1-1 mutation affects phyto-chrome-regulated mRNA accumulation of other Lhcb genes but not of rbcS1a, CCA1, or ATHB-2 genes. A, Total RNA was prepared from 6-d-old, dark-grown WT and shg1-1 seedlings harvested after R. RNA gel blots were successively hybridized with riboprobes specific for the CCA1, ubq3, and both Lhcb1*1 and Lhcb1*2 genes. Quantification of Lhcb1*1,2 (B) and CCA1 (C) mRNA levels in WT and shg1. D, RPA using an rbcS1a ³³P-labeled RNA probe which protects 160 bases. E, Quantification of rbcS1a mRNA normalized to ubq3 mRNA levels. F and G, ATHB-2 RNA gel-blot analysis and quantification. Representative data from two independent experiments are shown.

One possibility for the reduction in R-induced mRNA accumulation of Lhcb genes is the loss or reduction of a positive-acting factor for Lhcb transcription. CCA1 is such a transcription factor. It binds to a conserved motif in the Lhcb1*3 promoter in vitro (Wang et al., 1997), and loss of CCA1 in a null mutant caused reduced phytochrome induction of Lhcb1*3 (Green and Tobin, 1999). In dark-grown seedlings, CCA1 itself is regulated by phytochrome (Wang and Tobin, 1998), and the kinetics of its induction are consistent with a role in activating Lhcb genes. However, phytochrome induction of CCA1 was not affected in shg1-1 (Fig. 5, A and C).

To determine whether shg1-1 is also affected in the signaling pathway for other phytochrome-regulated genes, we tested the phytochrome regulation of rbcS1a and ATHB-2. Like Lhcb genes, rbcS1a is a nuclear-encoded gene for a chloroplast protein. Of the small family of genes encoding the small subunit of Rubisco, rbcS1a is the most strongly induced by phytochrome (Dedonder et al., 1993). Figure 5, D through E, shows that the increase in rbcS1a mRNA after R treatment of shg1 seedlings approximated that found in WT. Thus, the phytochrome regulation of rbcS1a is not significantly affected in the mutant line. We also tested phytochrome regulation of ATHB-2. The ATHB-2 gene, which encodes a homeodomain protein (Schena and Davis, 1992) not thought to be related to chloroplast function (Schena et al., 1993; Steindler et al., 1999), is down-regulated by phytochrome (Carabelli et al., 1993; 1996). As shown in Figure 5, F and G, ATHB-2 was similarly regulated in WT and shg1-1. Thus, SHG1 affects only a subset of phytochrome-regulated genes. Furthermore, the fact that the expression of the CCA1 gene is not affected, although the Lhcb genes to which it can bind are, suggests that either SHG1 acts downstream of CCA1 or it acts in a different pathway.

Mutation of Conserved Motifs in the Lhcb1*3 Promoter Affects β-Glucuronidase (GUS) Reporter Gene Expression in a Way Similar to shg1-1

The reduced induction of the endogenous Lhcb1*3 gene in shg1 mutants was the molecular phenotype predicted to be found in a mutant involved in phytochrome signaling. To identify promoter elements that might be targets of the signal transduction pathway affected in shg1 mutants, we introduced nucleotide changes in the context of a full-length (1.3-kb) Lhcb1*3 promoter. These mutations were targeted to three different motifs implicated previously as important for phytochrome responsiveness. These motifs and the mutant constructs are shown in Figure 6A. The Mu1 construct has a 1-bp change in each of the two imperfect repeats of the motif AA^A/_CAATCT. This motif is conserved among Lhcb promoters and is part of a site to which the transcription factor CCA1 binds. A DNA fragment containing the Mu1 mutation competed less effectively for CCA1 binding in vitro than the WT equivalent, though more effectively than a fragment containing a 9-bp mutation (Wang et al., 1997). The Mu2 construct has two bases changed in the CCAAT sequence and the Mu3 construct has two bases changed in each of two GATA boxes. The WT and mutant promoters were fused to the GUS reporter gene, and the constructs were stably transformed into Arabidopsis. F₃ seedlings from multiple transgenic lines were analyzed for phytochrome regulation of the introduced gene for each construct.

Figure 6.

Mutating either CCAAT or both GATA promoter motifs in Lhcb1*3::GUS sharply reduces phytochrome-responsive and constitutive dark GUS expression. A, Diagram of the TATA-proximal regions of WT and mutant Lhcb1*3::GUS constructs. The Lhcb1*3 promoter (−1,150 to +34) was fused upstream of the GUS gene in pBI101 (CLONTECH, Palo Alto, CA). The sequences of WT nucleotide motifs are shown in uppercase letters. The motif that is mutated in each mutant reporter is enclosed in a box, with the changed base denoted in lowercase. B, RPA of total RNA from 5-d-old seedlings given no light treatment (D) or 1 min R 3.5 h before harvest. Three independent lines transformed with either WT or mutant constructs Mu1, Mu2, and Mu3 were used; the fragments protected by GUS and ubq3 riboprobes are indicated. C, Comparison of the activities of WT and mutant promoters from the RPA above. For each set of lines, fold-induction of GUS mRNA levels, shown in parentheses above each R column, were obtained by dividing relative GUS mRNA from R samples by relative GUS mRNA from D samples, which gives an indication of the phytochrome inducibility of the transgenic construct. For each line, the value for relative GUS mRNA was obtained by dividing the absolute GUS mRNA level with the absolute ubq3 mRNA level and the average relative value for three lines per construct is shown. Error bars represent se. D, GUS activity found in 7-d-old dark-grown transgenic lines containing either WT or mutant constructs. Six to eight independent lines were tested per construct. E, Phytochrome inducibility of Lhcb::GUS is abolished when expressed in the shg1-1 mutant background. The WT Lhcb::GUS construct was crossed into the shg1-1 mutant background. Five-day-old dark-grown WT Lhcb::GUS and shg1-1/Lhcb::GUS seedlings were given no (D) or R treatment, 4 h before harvesting for RNA extractions. GUS and ubq3 mRNA levels were analyzed by RPA. Calculation of fold-induction levels, shown in parentheses, was done as explained in C. Error bars represent se from three independent lines.

We found that Mu2- and Mu3-transformed lines exhibited reduced dark expression and phytochrome induction of GUS. The results from one representative experiment, using three independent lines for each of the WT, Mu1, Mu2, and Mu3 constructs, are shown in Figure 6, B and C. In different WT Lhcb::GUS lines, the increase in levels of GUS mRNA after R treatment ranged from 9- to 18-fold, and in Mu1 lines the average increase was approximately 12-fold. Thus, it does not appear that the nucleotide changes in the Mu1 construct affected phytochrome responsiveness of that promoter. In contrast, both the Mu2 and Mu3 lines had consistently low increases in response to the R treatment, ranging from no induction to 5-fold induction. We have extended previous studies by showing that mutating these motifs in the context of the full-length promoter has a strong effect on expression. Our results also demonstrate that the loss of signaling through these motifs results in a reduction in phytochrome inducibility of the Lhcb1*3 promoter such that it should be detectable in our promoter-mediated mutant screen.

We also tested whether the mutations affected dark Lhcb1*3 expression. Accumulation of GUS mRNA in the dark appeared to be reduced in the Mu 1-3 lines, but it was difficult to ascertain the accuracy of our measurements because the mRNA levels were so low. Therefore, we assayed for GUS activity rather than GUS mRNA levels in the dark. We used at least six independent lines per construct for these measurements. The results from one such experiment are shown in Figure 6D. The WT and Mu1 lines showed a similar range of GUS activity levels, whereas Mu2 and Mu3 lines had considerably lower levels. Some Mu2 lines had GUS activities that were lower than the WT average by 10-fold. GUS mRNA levels from light-grown Mu2 and Mu3 lines were also consistently lower than those from WT lines (data not shown). Thus, the CCAAT and GATA motifs in the Lhcb1*3 promoter are not only important for phytochrome-induced expression, but also for maintenance of expression. Mutations that affect any of the proteins that mediate signaling through these motifs thus are likely to affect Lhcb1*3 expression in dark and in light.

The results presented in Figure 6E suggest that SHG1 may act in such a signaling pathway. We crossed one transgenic line containing the WT Lhcb::GUS construct with shg1-1. By introducing a WT Lhcb::GUS into the shg1-1 mutant background and analyzing different siblings from the cross, we found that phytochrome responsiveness of the construct was abolished. In addition, the level of GUS mRNA in the dark was reduced. This result confirms other experiments in which we observed reduced dark accumulation of the endogenous Lhcb mRNA in shg1-1 and shg1-2 (Figs. 2 and 5; data not shown). The shg1 mutation causes changes in Lhcb1*3 expression that are similar to those caused by mutating the CCAAT or GATA motifs. Therefore, it is possible that the SHG1 locus encodes a protein that acts as a component of a phytochrome-signaling pathway that terminates at one of the DNA-binding sites on the Lhcb1*3 promoter, possibly at the CCAAT or GATA site.

Suc Causes Additional Photomorphogenic Abnormalities in shg1

shg1 seedlings exhibit additional mutant phenotypes when grown in light and on medium containing Suc. As shown in Figure 7, A and B, one effect of a shg1 mutation that is unmasked by growth on Suc is a long hypocotyl phenotype. The characteristic increase of hypocotyl elongation on Suc-containing medium (Kurata and Yamamoto, 1998) was exhibited by WT seedlings, which were 50% taller on MS medium supplemented with 2% (w/v) Suc (MS2S) than on Murashige and Skoog (MS) medium while maintaining the same morphology. However, shg1-1 seedlings were 200% taller on MS2S than on MS alone. In addition, their petioles remained tightly closed in an upright position, and their cotyledons were only half open. This is in sharp contrast to shg1-1 seedling morphology on MS, which is essentially a paler and smaller version of the WT morphology. Thus, growth on Suc-containing medium causes shg1 seedlings to retain aspects of etiolation when grown in light.

Figure 7.

shg1-1 seedlings exhibit additional photomorphogenic abnormalities when grown on medium containing Suc. A, Photomorphogenic growth of shg1-1 seedlings is altered by the presence of Suc in the growth medium. WT and shg1-1 seedlings were grown on MS plates containing no (− Suc) or 2% Suc (+ Suc) for 7 d under a short-day photoperiod. Mean hypocotyl lengths of 15 to 20 seedlings are shown with error bars denoting se. Photographs of representative WT (left) and shg1-1 (right) seedlings grown on medium without (B) or with (C) Suc (bar = 1 mm). D, Anthocyanin content of shg1-1 seedlings is higher than WT when grown on medium containing Suc. WT and shg1-1 seedlings were grown on MS plates containing no (− Suc) or 2% Suc (+ suc) for the indicated number of days under a short-day photoperiod. Anthocyanins from 20 seedlings were extracted and the difference between A₅₃₅ and A₆₅₀ represents anthocyanin content. Three different sets of 20 seedlings were measured for each treatment. Error bars denote se.

Figure 7D shows that shg1 seedlings accumulated high levels of anthocyanin only when the growth medium was supplemented with Suc. Light-grown WT and shg1-1 seedlings accumulated similar levels of anthocyanin when grown on plates with MS medium only or on soil (data not shown). On MS2S, there was a substantial increase in anthocyanin levels. This response to Suc has been described previously for WT petunia (Petunia hybrida; Weiss, 2000) and Arabidopsis (Tsukaya et al., 1991). In shg1-1 seedlings, however, the anthocyanin increase was similar to WT seedlings during the first 3 d of growth, but it continued to rise for a longer period, reaching a peak that reflected a 15-fold increase. This burst of anthocyanin accumulation peaked 4 to 5 d after germination, and then levels dropped to slightly above WT levels. By the 10th d, shg1-1 levels were only twice those of WT plants. Dark-grown shg1 seedlings were indistinguishable from WT when grown on medium with or without Suc (data not shown).

DISCUSSION

We have used phytochrome regulation of the Lhcb1*3 gene as a system to identify potential phytochrome signaling intermediates. By generating a stably transformed line containing an Lhcb1*3 promoter fused to a selectable marker and carrying out the genetic screen using IR, a light regime that supports high phytochrome-specific Lhcb1*3 promoter activity while preventing full de-etiolation (Tobin, 1981; Karlin-Neumann et al., 1991), we aimed to select mutants that affect phytochrome regulation of gene expression. We identified shg1, a yellow green mutant which exhibited reduced phytochrome induction of Lhcb1*3.

We showed that shg1 exhibited reduced phytochrome regulation of Lhcb genes, but normal phytochrome regulation of the rbcS1a gene. Our finding that rbcS1a is regulated normally provides genetic evidence for the separate regulatory pathways between Lhcb and rbcS that have been inferred from physiological experiments (Batschauer et al., 1986) and from the divergent light-regulatory elements found in each promoter (for review, see Arguello-Astorga and Herrera-Estrella, 1998). Coordinate regulation of Lhcb and rbcS genes has been typically found for phytochrome and photomorphogenic mutants (for hy mutants, Chory et al., 1989a, Sun and Tobin, 1990; det, Chory et al., 1989b; gun, Susek et al., 1993; cop, McNellis and Deng, 1995; and sun, Djikwel et al., 1997). Thus, SHG1 may affect a step in phytochrome signaling that is either after a branch point or separate from the pathway regulating rbcS expression.

The two other nuclear genes we tested, CCA1 and ATHB-2, do not encode chloroplast proteins and were also normally regulated by phytochrome in shg1-1. CCA1 is a transcription factor that binds to a conserved motif in the Lhcb1*3 promoter. We found that CCA1 mRNA accumulation after R is identical in WT and shg1-1. The reduced Lhcb gene expression observed in shg1 thus is unlikely to be due to defective phytochrome signaling to CCA1. The phytochrome regulation of ATHB-2, a homeodomain protein, was examined as an example of a gene that is involved in neither chloroplast function nor Lhcb expression (Steindler et al., 1999). We found that the expression of ATHB-2 mRNA in shg1 was also very similar to WT. Thus, phytochrome signaling to the other genes outside the Lhcb gene family is intact, suggesting that the lesion in shg1 specifically affects Lhcb gene expression.

We considered that the reduction in Chl levels seen in shg1 might stem from an impairment in chloroplast function. Lopez-Juez et al. (1998) showed a correlation between reduced Chl levels and abnormal plastid structure. In addition, mutations in chloro-plast-localized proteins can cause Chl deficiency (recent examples include Sundberg et al., 1997; Jarvis et al., 1998). To address whether shg1 might be mutated in any of the enzymes in the Chl biosynthetic pathway, we measured the levels of some of the Chl intermediates. Because the rate-limiting step in Chl synthesis is the synthesis of ALA, accumulation in an intermediate pool after feeding dark-grown seedlings with ALA would indicate blockage at the subsequent enzymatic step. We found no such accumulation (Fig. 4). ALA and PBG contents were also reduced in light-grown shg1 (J.A. Brusslan, data not shown). Because shg1 does green, we can conclude that the pathway for the stepwise synthesis of Chl is likely to be intact. The reduced flow of intermediates through the pathway may be the result of aberrant regulation of the Chl synthesis pathway, similar to that found for the aurea (au) and yellow-green-2 (yg-2) mutants of tomato (Lycopersicon esculentum) (Terry and Kendrick, 1999).

Different aspects of greening such as Chl synthesis and transcription of nuclear-encoded photosynthetic genes are processes controlled by phytochrome (Mohr, 1984; Silverthorne and Tobin, 1984). Loss or reduction of phytochrome leads to reduced Chl accumulation as well as reduced phytochrome induction of gene expression (Chory et al., 1989a; Parks and Quail, 1991; Somers et al., 1991; Reed et al., 1993; Terry and Kendrick, 1996). Proteins that have been implicated in phytochrome signal transduction can cause a pale-green phenotype when mutated. For instance, mutation of diphosphate kinase 2, a protein that interacts with phytochrome, caused a pale-green phenotype (Choi et al., 1999). Further evidence that a pale-green phenotype arises concomitant with a deficiency in functional phytochrome was provided by Lagarias et al. (1997) using plants overexpressing biliverdin reductase, an enzyme that inactivates phytochrome chromophores. They suggested that the reduction in Chl observed in these plants was a consequence of the loss of phytochrome regulation of its biosynthesis. These results indicate that disrupting phytochrome regulation in the plants may lead to improper regulation of such chloroplast functions as Chl biosynthesis.

Aberrant phytochrome regulation also affects growth in the dark. There is physiological evidence for phytochrome control of the “capacity” for Chl formation (for review, see Mohr, 1984). The au mutant of tomato has reduced protochlorophyll and abnormal proplastids (Terry and Kendrick, 1996, 1999; Montgomery et al., 1999). Lhcb mRNA in dark-grown hy1 (Lopez-Juez et al., 1998) and au (Sharrock et al., 1988) seedlings is reduced. Our findings concerning the dark expression levels in shg1 and the mutant Lhcb1*3 promoter lines provide additional insights into regulation of gene expression. Dark expression of Lhcb mRNA is specifically affected in shg1. Moreover, disrupting either the CCAAT or both GATA boxes (Mu2 and Mu3, respectively) greatly reduced not only phytochrome responsiveness but also dark expression of Lhcb1*3 expression. Thus, the CCAAT and GATA motifs in the Lhcb1*3 promoter are not only important for phytochrome-mediated regulation, but also for maintenance of basal or dark expression. Mutations that affect any of the proteins that mediate signaling through these motifs are thus likely to affect Lhcb1*3 expression. It has been postulated that a signal from the plastids regulates the Lhcb gene family (Taylor, 1989) and a Chl intermediate can serve as one such signal for light-induced nuclear genes in Chlamydomonas reinhardtii (Kropat et al., 1997). It is tempting to speculate that shg1 and some of the cue mutants (cue3, cue6, and cue8; Lopez-Juez et al., 1998), which exhibit reduced Chl levels and reduced dark Lhcb1*3 expression, are mutated in a phytochrome-signaling pathway that includes the plastid and terminates at one of the DNA-binding sites on the Lhcb1*3 promoter.

The most remarkable light-grown phenotype of shg1 is its response to Suc at the seedling stage. shg1 seedlings have elongated hypocotyls and partially closed cotyledons when grown on medium containing Suc. These alterations in shg1 seedling morphology, coupled with its pale-green color, make shg1 seedlings appear more like etiolated than light-grown seedlings, and suggests that Suc is acting to block full de-etiolation normally caused by light signals. The repressive effects of Suc on de-etiolation have been shown to be a phytochrome-dependent response in WT seedlings (Dijkwel et al., 1997). When grown in increasing light fluences and on medium containing Suc, phytochrome-deficient biliverdin reductase-overexpressing seedlings can be taller than dark-grown seedlings (Montgomery et al., 1999). We also found that the burst in anthocyanin accumulation seen in shg1 grown on Suc was more extensive than that of WT. Expression in transgenic Arabidopsis of a petunia chalcone synthase gene, encoding the enzyme for the first committed step in anthocyanin biosynthesis, was shown to be sugar-, particularly Suc-, dependent (Tsukaya et al., 1991). Furthermore, anthocyanin induction and the expression of chalcone synthase were demonstrated to be predominantly controlled by phy A (Kunkel et al., 1996). The increased hypocotyl length and anthocyanin accumulation in shg1 can be interpreted as an exaggeration of the effects normally induced by Suc in light-grown WT seedlings. Thus, Suc and SHG1 may act in a subset of the pathways normally regulated by phytochromes, similar to Suc-uncoupled (sun) mutants (Dijkwel et al., 1997). On the other hand, the heightened pigmentation that shg1 seedlings exhibit on Suc may constitute a stress response. In addition to light and Suc (for review, see Weiss, 2000), environmental stimuli such as stress can induce flavonoid biosynthesis (Shirley, 1996). Interpretation of the phenotypes induced by Suc in light-grown shg1 seedlings is complicated further by the antagonistic relationship of some phyA and phyB pathways (Casal, 2000). Furthermore, it has been reported recently that the dominant negative interference of phy A-mediated hypocotyl growth inhibition in far-red by overexpression of phyB (Wagner et al., 1996) is dependent on the availability of metabolizable sugars such as Suc (Short, 1999). Further experiments are necessary to determine whether shg1 is involved in the phyA photobiological pathway that is susceptible to the antagonistic action of phyB and Suc (Casal et al., 2000).

In summary, we have identified and characterized shg1, a pale-green mutant that shows reduced phytochrome induction of Lhcb genes as well as reduced Lhcb mRNA accumulation during growth in darkness. In these respects, shg1 is similar to other pale-green mutants. Phytochrome regulation of other genes that we tested, however, was normal. This differential regulation of Lhcb genes distinguishes shg1 from the other mutants. Because phytochrome is known to regulate the synthesis of Chl, expression of nuclear-encoded chloroplast proteins, and the development of plastids, it is difficult to ascertain whether the shg1 mutant phenotype is caused by defective plastid function or by impaired phytochrome signaling. We provide evidence that reduced Lhcb gene expression, which may be indicative of defective plastids, may also be caused by disruption of phytochrome signaling to Lhcb promoters. In addition, the light-grown phenotype of the shg1 mutant adds genetic proof of the complex interplay of signals between phytochrome, Suc, and photomorphogenic development.

MATERIALS AND METHODS

Plant Material and Growth Conditions

The mutant screen was conducted with Arabidopsis ecotype Wassilevskija (WS) seeds. Seedlings were grown at 22°C to 24°C on MS, MS2S, or on soil. Seedlings for hypocotyl length and anthocyanin content measurements were grown under short-day (8-h light/16-h dark) conditions with a fluence of approximately 90 μmol m⁻² s⁻¹ from a combination of fluorescent and incandescent bulbs. Seedlings for Chl measurements were grown on MS under long-day (16-h light/8-h dark) conditions with a fluence of approximately 45 μmol m⁻² s⁻¹. Etiolated seedlings analyzed for RNA were sown onto MS or MS2S plates, cold treated for 2 d, induced to germinate by a 30-min exposure to white light, then grown for another 5 d in the dark. Seedlings were given either no or 1 min R, then harvested for RNA extraction 3 to 4 h later. R treatments and sources were as described in Tobin (1981).

Mutant Screen

Seeds (600 mg, approximately 25,000) from the A4 transgenic line containing a chimeric Arabidopsis Lhcb1*3::Nia1*2 gene (Heimer et al., 1995) were mutagenized by incubation in 20 mm ethyl methanesulfonate for 8 h. Seeds from 100 M₁ plants were pooled to constitute one family, and approximately 1,000 seeds per family were screened under conditions similar to those described by Heimer et al. (1995). We screened 125,000 M₂ seeds. Approximately 330 seeds were sown onto a prerinsed filter on a plate containing nitrate-free MS medium, 10 mm l-Gln, 0.05% (w/v) MES (2-[N-morpholino]ethane-sulfonic acid, Sigma, St. Louis), 0.7% (w/v) phytagar (Gibco-BRL, Rockville, MD), and 2% (w/v) Suc (pH5.8), cold treated for 2 d, exposed to white light for 30 to 60 min, then transferred to IR light (1 min R every 2 h) for 5 d. Seedlings that were taller than 6 mm were transplanted to soil. One hundred twenty-seven individuals survived and set seed. M₃ seeds were used to grow tissue for NR assays, which were performed as described by Heimer et al. (1995).

Mapping

We generated a mapping population by crossing mutant line 60 to the Columbia ecotype and identifying 75 yellow-green F₂ offspring. We isolated DNA from these lines using either leaf and inflorescence tissue or pooled F₃ seedlings, according to the protocol described by Ausubel et al. (1994). We used markers for simple sequence length polymorphisms and cleaved amplified polymorphic sequences to obtain the approximate map location of shg1 (Konieczny and Ausubel, 1993; Bell and Ecker, 1994; Arabidopsis thaliana Database, Stanford University, Palo Alto, CA; http://www.Arabidopsis.org).

GUS Constructs

To construct the WT Lhcb1*3::GUS reporter construct, we introduced an XbaI site at +34 of Lhcb1*3 using the primer pair Atcab28 (−698 to −678 of the promoter) and JX77 (5′ cgctctAGATTTGTTGTTGTAAGCCAAG 3′, spanning +11 to +34 of the coding region). The amplified product was cut with BglII and XbaI, and used to replace the BglII/XbaI fragment from a pBlueScriptKS(-) vector containing an EcoRI/BamHI fragment spanning −1,150 to −1 bp of the Lhcb1*3 promoter used by Kenigsbuch and Tobin (1995); this plasmid was named pWTcabXba. The Lhcb1*3 promoter was excised from pWTcabXba as a HindIII/XbaI fragment and fused to the GUS gene by insertion into the polylinker region of pBI101 (CLONTECH).

The base substitutions in the Mu1 construct were created by amplifying two Lhcb1*3 fragments using the primer pairs Atcab28 and Atcab32 (5′-AGCAAACAAACAATCTAAACCCCAAAAAAAATTTATGACT-3′, spanning −129 to −90 of the upper strand of the Lhcb1*3 promoter), and Atcab33 (sequence complementary to Atcab32) and JX77. These fragments were combined to serve as template for another round of PCR using Atcab28 and JX77. One product around 732 bp, presumed to contain the mutated nucleotide, was cut with BglII and XbaI, yielding a 283-bp fragment that was used to replace the analogous fragment in pWTcabXba. Using the same strategy, Mu2 and Mu3 base substitutions were introduced using the primers: MO1 (5′ TGACTAGGCGAAAGCAACCTCAGAGATTGATATTTC 3′, spanning −94 to −55 of the upper strand); MO2 (sequence complementary to MO1) for Mu2 containing a mutated CCAAT box; MO3 (5′ CTCAGAGATTTTTATTTCAATTTAAGACAGTATTTAGATTTC 3′, spanning −72 to −35); and MO4 (sequence complementary to MO3) for Mu3 containing two mutated GATA boxes. Mutant Lhcb1*3 promoters were inserted into the HindIII/XbaI polylinker region of pBI101. All constructs were transformed into Arabidopsis ecotype Columbia by vacuum infiltration. For all the independent lines used in this study, the correct promoter::GUS junctions and TATA-proximal regions of the T2 lines were verified by sequencing.

Pigment Analysis

Anthocyanins were measured as described by Schmidt and Mohr (1981). Anthocyanins were extracted by immersing 20 seedlings in propanol:HCl:H₂0 (18:1:81 percent volume) and boiling for 3 min. Extracts were left in the dark overnight at room temperature, clarified by centrifugation, and analyzed spectrophotometrically. The difference between A₅₃₅ and A₆₅₀ represents anthocyanin content. Three different sets of 20 seedlings were measured for each treatment.

Chls were extracted from cotyledons or leaf tissue by overnight immersion in dimethylformamide. Chl content was calculated from absorbance readings using the Porra equation (Porra et al., 1989).

ALA Feeding

Seeds (approximately 250) were sown onto MS2S plates, cold treated for 2 d, exposed to white light for 20 min, and then grown in the dark for 4 d at 24°C. Three milliliters of 10 mm ALA, 5 mm MgCl₂, and 10 mm NaPO₄ (pH 7.0) were added to each plate under a green safelight. Seedlings were grown in the presence of ALA for 12 h, then homogenized in 5 mL 9:1 acetone:0.1 m NH₄OH, and centrifuged at 3,000 rpm for 5 min. Fluorescence measurements were performed on the recovered supernatants with an SPEX fluorometer (Metuchen, NJ) using Ex400:Em632 for Protoporphyrin IX, Ex440:Em633 for protochlorophyllide, and Ex420:Em595 for Mg protoporphyrin.

Hypocotyl Length Measurement

Seedlings were grown on plates and pressed down to the surface of the plate to display the hypocotyls on the plane of the plate. Photographs were taken using a digital camera and hypocotyl lengths were determined using National Institutes of Health image software (public domain; Bethesda, MD).

Analysis of RNA

Isolation of total RNA from etiolated seedlings and RPAs were performed according to Brusslan and Tobin (1992) with some modifications. Riboprobes were synthesized using ³³P-UTP (ICN, Irvine, CA) and purified using G-50 spin columns. RNA gel-blot analyses were done as described by Wang and Tobin (1998). Intensities of bands were quantitated using a Phosphorimager (Molecular Dynamics, Sunnyvale, CA). The ATHB-2 probe was synthesized by random priming the ATHB-2/360 plasmid, which contains the 360-bp HindIII insert of ATHB-2 (corresponding to nucleotides 93–453, I. Ruberti, personal communication). The Lhcb1*3 riboprobe was identical to the one designated cab1 by Karlin-Neumann et al. (1991); the riboprobe that recognized both Lhcb1*1 and Lhcb1*2 was identical to that designated cab 2,3 by Brusslan and Tobin (1992). The CCA1 and ubq3 riboprobes were identical to those from Wang and Tobin (1998). To produce a GUS riboprobe that protects a 176-bp product in the RPAs, we created a plasmid template containing a short fragment of GUS. An XbaI site was introduced at +12 of the GUS gene using the primer 5′ CGTCCTCTAGAAACCCCAACCCG 3′ and a BglII site at +174 using the primer 5′ CGAAGATCTGCATCGGCGAACTGATCG 3′ (+182 to +156). The PCR product was digested with XbaI and BglII, then ligated to pGEM112f(-) that had previously been digested with BamHI and XbaI. This plasmid was linearized with XbaI to serve as template for a T7 polymerase transcription reaction. All experiments were repeated with similar results.