Abstract
Ambient light controls the development and physiology of plants. The Arabidopsis thaliana photoreceptor phytochrome B (PHYB) regulates developmental light responses at both seedling and adult stages. To identify genes that mediate control of development by light, we screened for suppressors of the long hypocotyl phenotype caused by a phyB mutation. Genetic analyses show that the shy (short hypocotyl) mutations we have isolated fall in several loci. Phenotypes of the mutants suggest that some of the genes identified have functions in control of light responses. Other loci specifically affect cell elongation or expansion.
PLANTS adjust their development in response to ambient wind, temperature, water, and light. Such adjustments allow plants to grow in a variety of sites and to adapt to seasonal changes in external conditions. Light is among the most relevant environmental signals because plants use light for photosynthesis and because light conditions reflect both the local growth environment and diurnal and seasonal time (Smith 1994). Plants sense the light environment using a battery of photoreceptors that specifically control development. These include red/far-red light photoreceptors called phytochromes, blue light photoreceptors called cryptochromes, and unnamed photoreceptors that mediate phototropism and UV light responses (von Arnim and Deng 1996; Fankhauser and Chory 1997).
The phytochromes are the most extensively characterized developmental photoreceptors in plants (Quail 1991; Furuya 1993). They are soluble dimeric proteins, and each ~120-kD monomer has a covalently attached linear tetrapyrrole chromophore. Phytochromes are synthesized in the dark in a red light–absorbing form called Pr. Upon absorption of red light, they are converted to a far-red light absorbing form called Pfr. On the basis of physiological, genetic, and biochemical studies, Pfr is thought to be the active form. Studies of seed germination have suggested that Pr may also have an activity that counteracts the activity of the Pfr form (Reedet al. 1994; Shinomuraet al. 1994).
Although the mechanisms of phytochrome signal transduction are uncertain, several models have been proposed. A cyanobacterial phytochrome homolog signals by a phosphorelay mechanism (Yehet al. 1997), suggesting that higher plant phytochromes might also signal in this fashion. Pharmacological studies have suggested that phytochromes may act through branched signaling pathways involving G proteins, cyclic GMP, and calcium (Shacklocket al. 1992; Neuhauset al. 1993; Bowleret al. 1994). More recently, it has been reported that phytochrome B migrates to the nucleus in the light, suggesting that phytochromes may signal in the nucleus as well as in the cytoplasm (Sakamoto and Nagatani 1996).
The genetics of plant light responses have been studied most extensively in Arabidopsis thaliana. Arabidopsis has five genes that encode phytochrome apoproteins (Sharrock and Quail 1989; Clacket al. 1994), and mutations are known in three of these, PHYA (Deheshet al. 1993; Whitelamet al. 1993; Reedet al. 1994), PHYB (Koornneefet al. 1980; Reedet al. 1993), and PHYD (Aukermanet al. 1997). Analyses of the phenotypes of these mutants have shown that these three phytochromes each mediate overlapping subsets of light responses, but often do so under distinct light conditions. For example, phyA mutant seeds germinate poorly in response to very low fluence light over a wide spectral range (Bottoet al. 1996; Shinomuraet al. 1996), whereas phyB mutant seeds germinate poorly in response to red light (Reedet al. 1994; Shinomura et al. 1994, 1996). phyA mutants fail to inhibit hypocotyl elongation in response to far-red light (Nagataniet al. 1993; Parks and Quail 1993; Whitelamet al. 1993), whereas phyB mutants fail to inhibit hypocotyl elongation in response to red light (Koornneefet al. 1980). Finally, phyA mutants flower later than wild-type plants in response to night breaks or day length extensions (Johnsonet al. 1994; Reedet al. 1994), whereas phyB mutants flower earlier than wild-type plants under a variety of conditions (Gotoet al. 1991; Reedet al. 1993; Bagnallet al. 1995; Koornneefet al. 1995). Effects of phyA and phyD mutations on the inhibition of hypocotyl elongation by red light can be seen in a phyB mutant background (Reedet al. 1994; Aukermanet al. 1997). These results reveal that the multiplicity of phytochromes serves in part to increase the versatility of the plant in responding to different light environments. These phenotypic analyses also suggest that the signaling pathways initiated by the different phytochromes may overlap. hy4 mutants deficient in the blue light photoreceptor cryptochrome 1 have some phenotypes in common with the phytochrome-deficient mutants (Koornneefet al. 1980; Younget al. 1992; Ahmad and Cashmore 1993), suggesting that blue light signaling pathways also converge with phytochrome signaling pathways. Physiological analyses of photoreceptor mutants have also suggested that blue and red light systems interact functionally (Casal and Boccalandro 1995; Ahmad and Cashmore 1997).
Mutations that identify possible downstream components of phytochrome signaling have been isolated in long hypocotyl screens, in screens for early flowering mutants, and in screens for seedlings with characteristics of light-grown plants in the dark. The fhy1 and fhy3 mutants have long hypocotyls in far-red light, suggesting that they may have lesions in a PHYA-specific signaling pathway (Whitelamet al. 1993; Barneset al. 1996). hy5 mutants have long hypocotyls under all light conditions, suggesting that HY5 may act downstream of the convergence of different photoreceptor pathways (Koornneefet al. 1980). Consistent with this idea, HY5 encodes a basic leucine zipper transcription factor (Oyamaet al. 1997). elf3 mutants have an elongated hypocotyl under all light conditions, and they flower early (Zagottaet al. 1996). The cr88, pef1, pef2, pef3, and red1 mutants have long hypocotyls in red light (cr88, pef2, pef3, and red1) or in both red and far-red light (pef1; Ahmad and Cashmore 1996; Lin and Cheng 1997; Wagneret al. 1997). These genes may encode positive regulators of light signaling.
Screens for mutants with short hypocotyls and leaf development in the dark have yielded candidate negative regulators. A series of cop, det, and fus mutants have short hypocotyls, develop leaves, and express light-induced genes in the dark (reviewed in Wei and Deng 1996; Fankhauser and Chory 1997). The biochemical functions of the products of these genes remain unclear. However, the DET1, COP1, COP9, and FUS6 proteins localize to the nucleus, suggesting that they may repress gene expression in the dark (Pepperet al. 1994; von Arnim and Deng 1994; Chamovitzet al. 1996; Staubet al. 1996). Double-mutant plants that carry photoreceptor mutations and det/cop/fus mutations have phenotypes consistent with the DET/COP/FUS gene products functioning downstream of photoreceptor pathways (Chory 1992; Ang and Deng 1994; Miséraet al. 1994; Wei et al. 1994a,b). More detailed analysis of cop/det/fus mutant phenotypes suggests that these genes may play a more general role in regulating gene expression in response to a variety of stimuli in addition to light. For example, det1 seedlings express LHCB (encoding light harvesting chlorophyll a/b–binding protein) and other genes in roots (Chory and Peto 1990); and cop1, det1, and cop9 mutants overexpress genes that are normally activated by pathogen infection, hypoxia, or developmental signals, as well as genes normally activated by light (Mayeret al. 1996). Finally, several other mutants with subsets of the phenotypes described above have been described, including det2, det3, cop2, cop3, cop4, and doc1 (Choryet al. 1991; Cabreray Pochet al. 1993; Houet al. 1993; Liet al. 1994).
Although many genes involved in light signaling have been identified in these screens, it is likely that numerous other relevant loci remain to be discovered. In other systems, screens for suppressors and enhancers of mutations in a pathway have identified important new genes (for example, Karimet al. 1996). These genetic methodologies have been used less frequently in plants (Koornneefet al. 1982; Niyogiet al. 1993; Carolet al. 1995; Cernacet al. 1997; Silverstoneet al. 1997), but promise to become very useful for dissecting light responses. A screen for suppressors of a hy2 phytochrome chromophore–deficient mutation (Kimet al. 1996) and a screen for suppressors of a det1 mutation (Pepper and Chory 1997) have recently revealed more candidate light-signaling mutations. To identify other genes involved in light signaling, we have conducted a screen for mutations that suppress the long hypocotyl phenotype caused by a phyB mutation. In this report, we describe the results of our initial screen. Following the precedent established by Kim et al. (1996), we have called these new mutations shy (for short hypocotyl and suppressor of hy). We have characterized various phenotypes of the mutants, allowing us to identify those that affect light signaling.
MATERIALS AND METHODS
Mutagenesis and genetic methods: Mutant phyB-1 (previously called hy3-Bo64) is in the Landsberg erecta background and has a stop codon in the PHYB coding sequence (Koornneefet al. 1980; Reedet al. 1993; Quailet al. 1994). We incubated 5000 phyB-1 M1 seeds overnight in 0.3% ethyl methane sulfonate and then rinsed several times with water. We collected M2 progeny in eight batches from ~600 M1 plants per batch. We screened ~14,000 M2 seedlings on MS/sucrose/agar plates (see below) under white light for short hypocotyl variants. To detect revertants of the starting phyB-1 mutation and to follow the phyB-1 mutation in subsequent genetic manipulations, we assayed for the presence (wild-type allele) or absence (phyB-1 mutant allele) of an AlwNI restriction site in a PHYB-specific PCR product amplified from chromosomal DNA of the plant being tested (Reedet al. 1993). Mutants were judged to be independent if they came from different batches or if they had clearly distinguishable phenotypes.
To separate the shy mutations from the starting phyB-1 mutation, we crossed the phyB-1 shy strains to wild-type Landsberg erecta, allowed these F1 plants to self-fertilize, and identified PHYB/PHYB shy/shy progeny in the F2 or F3 generation. For shy mutations that confer an obvious phenotype (such as dwarfism) in a wild-type background we identified PHYB/PHYB shy/shy plants in the F2 generation. For mutations that confer a phenotype similar to the wild-type phenotype in a phyB-1 mutant background and that confer no dramatic phenotype in the wild-type PHYB background, we identified PHYB/phyB-1 heterozygous F2 plants from the outcross to wild type that did not segregate tall (phyB-1/phyB-1 SHY/SHY or phyB-1/phyB-1 SHY/shy) F3 progeny. The failure to segregate tall F3 plants indicated that these F2 plants were homozygous for the shy mutation (or that the shy mutation was linked to the phyB-1 mutation, a possibility tested in the mapping experiments described below). From the F3 progeny populations, we identified PHYB/PHYB individuals of genotype PHYB/PHYB shy/shy. In these genetic manipulations, we distinguished the PHYB and phyB-1 alleles by the PCR-based assay described above.
We mapped the shy mutations using PCR-based SSLP and CAPS markers polymorphic between Landsberg and Columbia ecotypic backgrounds (Konieczny and Ausubel 1993; Bell and Ecker 1994). We crossed phyB-1/phyB-1 shy/shy plants (Landsberg erecta background) with phyB-9/phyB-9 plants (Columbia background), and assayed DNA from individual F2 progeny for Landsberg- or Columbia-specific polymorphisms. Both the phyB-1 and phyB-9 mutations create stop codons in the PHYB-coding sequence (Reedet al. 1993). Since suppressor mutations isolated in this screen should bypass the requirement for PHYB, we expected the phyB-1 and phyB-9 alleles to behave equivalently in the mapping populations. In cases where we suspected that a mutation was allelic to a previously known mutation, we performed complementation tests between the shy mutation and the previously described mutation, and we looked for lack of segregation of F2 plants with a wild-type phenotype, indicating that the two mutations mapped to the same location. In cases where we found allelism to a previous locus, we have given the new mutation an allele designation that incorporates the established gene name.
Phenotypic tests: Seeds were surface sterilized and plated on Murashige and Skoog (MS)/agar plates [1× MS salts (GIBCO, Grand Island, NY), 0.8% phytagar (GIBCO), 1× Gamborg's B5 vitamin mix (Sigma, St. Louis, MO)] with or without 2% sucrose, stored overnight at 4°, and moved to the appropriate light condition. For hypocotyl length tests in red light, we used LED red light sources emitting light with a peak at 670 nm and a half bandwidth of 25 nm (Quantum Devices, Inc., Barneveld, WI). For far-red light, we used LED sources emitting light with a peak at 730 nm and a half bandwidth of 25 nm (Quantum Devices). For blue light, we used cool white fluorescent bulbs filtered through Schott blue glass filter No. 5-57 (Newport Industrial Glass, Costa Mesa, CA). For fluence rate/response experiments, light was filtered through various thicknesses of bronze plexiglass No. 2412 (Golden Rule Plastics, Haw River, NC). This filter causes minimal distortion of the light spectrum (data not shown). Light levels were measured with an LI-189 quantum radiometer (Li-Cor, Lincoln, NE), or extrapolated based on numbers of layers of plexiglass. For red and far-red fluence rate–response experiments, we grew seedlings on MS/agar plates (without sucrose) that were placed vertically behind various thicknesses of plexiglass, with the light source placed so as to project horizontally. After 5 days, we took digital images of the plates with a CCD camera and measured the hypocotyl lengths using image analysis software (NIH Image, Bethesda, MD). For root length experiments, we grew seedlings in red light (30–50 μmol · m−2 · sec−1) on vertical MS/sucrose/agar plates for 5 days and measured roots against a ruler. In other experiments and for all blue light experiments, we grew plants on plates placed horizontally and measured hypocotyl lengths against a ruler.
phyB-1 shy seedlings grown in white light. Seedlings were grown for 8 days on MS/sucrose plates. Genotypes shown are (A) phyB-1 (starting strain), (B) Landsberg erecta (wild-type parent), (C) phyB-1 shy2-2, (D) phyB-1 shy2-3, (E) phyB-1 amp1-4, (F) phyB-1 shy-115, (G) phyB-1 pom1-15, (H) phyB-1 bot1-5, (I) phyB-1 shy3-1, (J) phyB-1 shy4-1, (K) phyB-1 shy5-1, and (L) phyB-1 shy6-1. shy4-1 and shy4-2 seedlings appeared similar to each other, and pom1-14 and pom1-15 seedlings appeared similar to each other. Therefore, only one mutant at each locus is shown.
For flowering time determinations, we grew seedlings on MS/sucrose/agar plates for 10–14 days and then transplanted them to soil (Pro-Mix BX; Hummert, St. Louis, MO). Experiments were performed in a Conviron growth chamber at 21°. Light was provided on a 9 h:15 h day:night cycle from 12 fluorescent (F72T12/CW/VHO, 160 W) and six incandescent (60 W) bulbs, and had an intensity at plant height of 100–230 μmol · m−2 · sec−1, depending on the experiment. We repeated the experiment six times, testing each genotype in between one and five different experiments (average = 2.5).
RESULTS
Isolation and genetic analysis of shy mutants: From our screen, we isolated 13 independent shy (short hypocotyl, or suppressor of hy) variants with mutations at sites distinct from the starting phyB mutation (see materials and methods). As described below, we have named those that turned out to be alleles of previously known loci according to the established gene names. Figure 1 shows phyB-1 shy mutant seedlings grown for 8 days in white light. They fall into several distinct phenotypic classes, as summarized in Table 1 and described in more detail below.
Classes of> shy mutations
To assess the degree of dominance or recessiveness of the shy mutations in the phyB-1 mutant background, we measured hypocotyl lengths after growth in constant red light, as this parameter is sensitive and easy to score. We compared hypocotyl lengths of phyB-1/phyB-1 SHY/shy F1 plants with those of the phyB-1/phyB-1 SHY/SHY and phyB-1/phyB-1 shy/shy parents. As shown in Table 2, the amp1-4, shy-115, shy-802, pom1-15, bot1-5, shy4-1, shy4-2, shy5-1, and shy6-1 heterozygous seedlings were the same height as SHY/SHY seedlings, indicating that these mutations are each recessive for hypocotyl length. In contrast, the shy2-2, shy2-3, pom1-14, andshy3-1heterozygous seedlings were significantly shorter than the SHY/SHY seedlings and significantly taller than the corresponding shy/shy seedlings, indicating that these mutations are each partially dominant.
To determine whether the short hypocotyl phenotype of these mutants was caused by mutation at a single locus, we checked the segregation of the short hypocotyl phenotype in the F2 generation of these backcrosses. For each mutant, the phenotype segregated in a manner consistent with a mutation at a single locus (Table 3).
We found three allelic pairs in our screen, at the SHY2, POM1, and SHY4 loci. As the shy2 mutations are partially dominant, our assessment that they are alleles of SHY2 (and of each other) is based on their conferring similar phenotypes as shy2-1 does (Kimet al. 1996), as well as mapping to the same location (Table 4). In addition, we have recently obtained independent molecular evidence that shy2-1, shy2-2, and shy2-3 are allelic (Q. Tian and J. W. Reed, unpublished data). As we identified single mutations at the remaining seven loci described here, the screen has not been saturated, and there are probably several other loci that can be mutated to give a shy phenotype.
We mapped the shy mutations by outcrossing them to a different ecotype and assaying polymorphic markers or by establishing allelism with previously mapped mutations (see materials and methods). Mapping results are summarized in Table 4. We obtained linkage to a polymorphic marker for all the loci except for the shy5-1 mutation, for which different mapping populations failed to show consistent linkage (data not shown). The 13 mutations fall in three previously known genes, AMP, SHY2, and POM1 (Chaudhuryet al. 1993; Hauseret al. 1995; Kimet al. 1996); two probable DWF genes (mutated in shy-115 and shy-802, but not assigned by complementation—see below; Feldmannet al. 1989); and five new genes, BOT1 (BOTERO1; H. Höfte, personal communication), SHY3, SHY4, SHY5, and SHY6.
To determine whether any of the shy mutations completely quench PHYB signaling, we tested whether the shy mutations were epistatic to the phyB-1 mutation. We compared the hypocotyl lengths of phyB-1 shy double mutants with that of the phyB-1 single mutant and with those of the corresponding shy single mutants. For each mutant, we identified PHYB shy plants among progeny of an outcross to a wild-type Landsberg erecta plant (see materials and methods). Table 5 shows hypocotyl lengths in red light of each of the PHYB shy and phyB-1 shy seedlings. In most cases, the phyB-1 shy double mutant is significantly taller than the PHYB shy single mutant, indicating lack of epistasis (Table 5). For pom1-14, the difference between PHYB and phyB-1 genotypes was not significant, indicating that pom1-14 is epistatic to phyB-1 for this phenotype.
Hypocotyl lengths of phyB-1/phyB-1 shy/shy and phyB-1/phyB-1 SHY/shy seedlings in red lighta
Phenotypic analyses of shy mutant plants: To assess whether the shy mutations affect genes involved in PHYB signal transduction or, more generally, in light signaling, we checked several phenotypes. Because a number of photomorphogenic mutations affect dark growth, we examined the morphology of shy seedlings in the dark. We also examined whether the shy mutations affect light-dependent phenotypes. As described above, phyB mutants have long hypocotyls in red light, have short roots, and flower early. We expected that mutations specifically limiting hypocotyl cell enlargement or elongation would suppress only the long hypocotyl phenotype (and perhaps other elongation phenotypes such as elongated root hairs or bolting stems), but not the flowering time or short root phenotypes. In contrast, mutations that affect a general control function might suppress multiple phenotypes caused by the phyB-1 mutation. These phenotypic criteria have indeed allowed us to distinguish mutations that affect elongation from those that affect putative control functions. Within this broad classification, many of the mutants have unique characteristics that define distinct roles in development (Table 1). We describe the mutants briefly here to help clarify the presentation below.
Segregation of shy mutant phenotypes among F2 progeny of backcrosses
Mapping of shy loci
The shy2-2 and shy2-3 mutants have leaves that curl up at the edges. The shy2-2 and shy2-3 mutations were each semidominant (Table 2). In these respects, these mutants resemble the shy1 and shy2-1 mutants isolated as suppressors of a hy2 mutation (Kimet al. 1996) and the axr3 mutants isolated as having auxin-resistant root growth (Leyseret al. 1996). The shy2-2 mutation conferred more extreme phenotypes than the shy2-3 mutation (see below), suggesting that it is the stronger allele. As described below, these mutants may identify an important control function in light-regulated development.
Hypocotyl lengths of phyB-1 shy and PHYB shy seedlings in red lighta
Mutations in the AMP (altered meristem program) gene confer altered phyllotaxy and partially de-etiolated growth in the dark (Chaudhuryet al. 1993; Chin-Atkinset al. 1996). Like the previously described amp mutants, the amp1-4 mutant we isolated appears dwarfed and has pale leaves and decreased apical dominance. In the light, amp1-4 seedlings formed more leaves than wild-type plants (Figure 1). In the dark, they showed partial leaf development (see below).
The shy-115 and shy-802 mutants are brassinosteroid-deficient dwarfs. These resemble the previously characterized bri1, cbb, det2, dim, and dwf mutants in having small dark green leaves (Feldmannet al. 1989; Choryet al. 1991; Takahashiet al. 1995; Clouseet al. 1996; Liet al. 1996; Szekereset al. 1996). Such mutants have been shown to have deficiencies in brassinosteroid synthesis or response. Both shy-115 and shy-802 mutants responded to exogenous brassinolide, suggesting that they are deficient in brassinosteroid synthesis (data not shown). shy-115 plants were fully fertile, whereas shy-802 plants were almost sterile and produced few seed. Both shy-115 and shy-802 complemented det2-1, and they also complemented each other. These equations therefore represent distinct loci. On the basis of their map positions, shy-115 may be allelic with DWF1 (also called DIM), and shy-802 may be allelic with DWF4. Because we obtained few shy-802 mutant seeds, it was difficult to subject the shy-802 mutant to extensive phenotypic analyses. Moreover, light-related phenotypes of brassinosteroid-deficient mutants have been described extensively. Therefore, we omitted the shy-802 mutant from many of the experiments described below.
pompom1 (pom1) mutants were first isolated as having abnormal root elongation (Hauseret al. 1995). The pom1-14 and pom1-15 mutants have similarly deficient root growth (see below). These mutants also have unusual hypocotyl morphology in the dark, and may be deficient in some aspect of cell elongation (see below).
The bot1-5 mutant has morphological characteristics suggestive of a general deficiency in cell enlargement. Cotyledons, leaves, and flower parts were all foreshortened, and the mutant produced very few seed (Figure 1; data not shown). By complementation analyses, we determined that bot1-5 is allelic to the botero1-1 mutation (data not shown). This mutation confers similar phenotypes and maps to the same vicinity as bot1-5 (H. Höfte, personal communication).
Finally, in contrast to the other mutants, the shy3, shy4, shy5, and shy6 mutants have no unusual morphological characteristics; instead, they exhibit phenotypes within the normal range of wild-type growth patterns. Mutations at three loci, shy4, shy5, and shy6, are recessive, and one mutation, shy3-1, is semidominant (Table 2). As described below, shy3-1 and shy5-1 mutants have quantitative phenotypes that suggest that they may affect PHYB signaling.
Mean hypocotyl lengths of PHYB shy seedlings after 6 days of dark growth. Seedlings were grown on MS plates containing 2% sucrose, and hypocotyl lengths were normalized to the hypocotyl length of wild-type seedlings. Data from three experiments were pooled. Error bars indicate standard deviations. In the experiment in which shy2-3, bot1-5, and shy4-2 data were gathered, the wild-type measurement had a relative standard deviation of 2.5 times that shown. In each case, the shy mutant hypocotyl lengths were statistically significantly shorter than the wild-type hypocotyl length (P < 0.05 for shy4-2, P < 0.001 for the rest).
Dark phenotypes: Mutations that activate light-response pathways constitutively might be expected to cause phenotypes in the dark. In fact, a number of genes thought to play important roles in photomorphogenesis were identified in screens for mutants that make leaves in the dark (see above). After 6 days in the dark, wild-type seedlings had a long hypocotyl, an unopened apical hook, small unexpanded cotyledons, and no leaf primordia (Figures 2 and 3). After 23 days, the apical hooks had opened, but the cotyledons were unexpanded and very few seedlings had visible leaf primordia (Figure 4, Table 6). We grew shy seedlings in the dark, and found that after 6 days all of them had significantly shorter hypocotyls than the wild-type seedlings (Figure 2). Some of the shy mutants also had open cotyledons or leaf development, morphological characteristics that are normally limited to light-grown plants (Table 6, Figure 4).
The shy2-2 and brassinosteroid-deficient shy-802 seedlings resembled light-grown seedlings most closely, in that after 23 days they had structures resembling true leaves. Petioles were very short, and trichomes were visible (Figure 4). In this respect, these two mutants resembled the de-etiolated mutants, det1, det2, and cop1, described previously (Choryet al. 1989; Denget al. 1991). Mutants carrying the weaker allele of SHY2, shy2-3, usually had expanded cotyledons after 23 days, but no visible leaf primordia (Figure 4, Table 6). The weaker brassinosteroid-deficient mutant shy-115 appeared similar to the wild type in the dark.
PHYB shy seedlings after 6 days of dark growth. Seedlings were grown on MS/sucrose plates. Genotypes shown are (A) Landsberg erecta, (B) shy2-2, (C) shy2-3, (D) amp1-4, (E) shy-115, (F) pom1-15, (G) bot1-5, (H) shy3-1, (I) shy4-1, (J) shy5-1, and (K) shy6-1. shy4-1 and shy4-2 seedlings appeared similar to each other, and pom1-14 and pom1-15 seedlings appeared similar to each other. Therefore, only one mutant at each locus is shown.
Several mutants had phenotypes that superficially resembled a de-etiolated phenotype, but diverged from a normal light growth pattern. amp1-4 seedlings had the most extreme phenotype after six days, having an open apical hook, expanded cotyledons, and appearance of leaf primordia. However, after 23 days, the organs in the positions where leaves would normally arise resembled elongatedpetioles or stems more than leaves, and yellow leaf blade material was absent or extremely abbreviated (Figure 4, Table 6). Similarly, the shy4-1, pom1-14, pom1-15, and bot1-5 mutants frequently developed branches resembling petioles (Figure 4, Table 6). This curious phenotype may represent a partial commitment to leaf development in these mutants. Alternatively, enhanced petiole or stem development may arise in conditions where the hypocotyl would otherwise elongate but is prevented from doing so by some physiological limitation caused by a shy mutation.
As well as making numerous petiole-like structures, the pom1-14 and pom1-15 mutants had a unique dark hypocotyl morphology. The hypocotyls were quite crooked (Figures 3 and 4), and upon closer inspection, they appeared somewhat disorganized, with a rough surface (data not shown).
The remaining mutants (shy3-1, shy4-2, shy5-1, and shy6-1) had short hypocotyls, but otherwise looked similar to wild-type in the dark. shy3-1 seedlings had a slightly open apical hook after 6 days. All had open apical hooks after 23 days, but the cotyledons were unexpanded and few seedlings had leaves (Figure 4, Table 6). Each of the mutants had a slightly higher incidence of leaf primordium formation than the wild type (Table 6). We do not know the reason for this quantitative difference, but it may conceivably be an indirect consequence of having shorter hypocotyls.
PHYB shy seedlings after 23 days of dark growth. Seedlings were grown on MS/sucrose plates. Genotypes shown are (A) Landsberg erecta, (B) shy2-2, (C) shy2-3, (D) amp1-4, (E) shy-115, (F) shy-802, (G) pom1-15, (H) bot1-5, (I) shy3-1, (J) shy4-1, (K) shy5-1, and (L) shy6-1.shy4-1 and shy4-2 seedlings appeared similar to each other, and pom1-14 and pom1-15 seedlings appeared similar to each other. Therefore, only one mutant at each locus is shown.
Hypocotyl elongation responses to red, blue, and far-red light: If a SHY protein activity is normally modulated by a light-signaling pathway, then mutation of the gene encoding that protein might cause the pathway to be more constitutive and less regulatable. In such a case, PHYB shy mutant seedlings would have a decreased response to light because of their “partially responding” baseline state, and they should be less sensitive to additional red light than wild-type seedlings. To examine this possibility, we tested the fluence rate–response behavior of the PHYB shy plants for hypocotyl elongation in constant red light. Figure 5, A and B, show the hypocotyl lengths of 6-day-old shy mutant seedlings grown under different fluence rates, normalized in each case to the hypocotyl length of the same strain in the dark. Most of the shy mutants responded to the same range of fluence rates as the wild type did, having hypocotyl lengths of ~30% of their hypocotyl lengths in the dark at the highest fluence rate tested (Figure 5, A and B). As expected, the phyB-1 mutant was less sensitive to red light, showing a minimum hypocotyl length of ~80% of its dark hypocotyl length.
Leaf formation by shy mutants after 23 days in the dark
The shy2-2, pom1-14, and pom1-15 mutants showed altered hypocotyl responses to red light. shy2-2 had a decreased response at the highest fluence rate tested (40–45 μmol · m−2 · sec−1), with a minimum hypocotyl length of ~60% of its dark hypocotyl length (Figure 5A). In contrast, the pom1-14 and pom1-15 mutants had longer hypocotyls at very low fluence rates (<0.1 μmol · m−2 · sec−1) than in the dark (Figure 5A). They thus showed the opposite response from the wild type in this fluence range. At higher fluence rates (>1 μmol · m−2 · sec−1), they had a normal response to red light, reaching a minimum hypocotyl length of ~70% of their dark hypocotyl length (Figure 5A). Because the shy2-2, pom1-14, and pom1-15 mutants are each extremely short in the dark (Figures 2 and 3, data not shown), these observations may reflect an inherent limitation in the degree to which red light can inhibit hypocotyl elongation. However, the bot1-5 mutant, responded proportionately similarly to the wild type despite having a very short hypocotyl in the dark (Figures 2 and 5B). This observation suggests that the shy2-2, pom1-14, and pom1-15 mutants do not reach a minimum attainable hypocotyl length in these experiments, and that the mutations may thus affect red light signaling or response pathways.
To test whether the shy mutations may affect response pathways downstream of other photoreceptors, we also performed fluence rate response experiments in blue and far-red light. We found that all the mutants responded to both types of light with the same threshold and saturation characteristics as the wild type (Figure 5, C and D; data not shown). In particular, the shy2-2 and pom1-15 mutants had normal responses to both blue light (Figure 5C) and far-red light (Figure 5D). Therefore, the mutations have not affected a blue or far-red light signaling pathway.
Root elongation: Rather than repressing cell expansion in all seedling tissues, PHYB causes a redistribution of growth away from the hypocotyl and toward the roots and cotyledons. Thus, light-grown phyB mutant seedlings have a longer hypocotyl but a shorter root and smaller cotyledons than wild-type seedlings (Reedet al. 1993; Neff and Van Volkenburgh 1994). To determine whether the shy mutations affect elongation in multiple tissues, we measured the rootlengths of phyB-1 shy seedlings grown in red light. This assay was particularly useful for distinguishing between mutants affected in regulatory functions, which should have roots longer than the phyB-1 seedlings, and mutants affected specifically in elongation functions, which would be expected to have shorter roots than phyB-1 seedlings. Mutants with roots the same length as phyB-1 seedlings could belong to either group. As shown in Figure 6, wild-type (PHYB SHY), amp1-4, shy2-2, and shy5-1 seedlings had significantly longer roots than phyB-1 seedlings, suggesting that these mutations may affect the redistribution of growth controlled by light. shy2-3, shy-802, pom1-14, pom1-15, bot1-5, shy4-2, and shy6-1 seedlings had significantly shorter roots than SHY seedlings, suggesting that these mutations affect functions specific to cell elongation. The remaining mutants (shy-115, shy3-1, and shy4-1) had roots of similar length as the phyB parent strain.
Flowering time: In addition to having elongation phenotypes, phyB mutants flower early (Gotoet al. 1991; Reedet al. 1993; Bagnallet al. 1995). Plants that overexpress PHYB also flower early (Bagnallet al. 1995). We assessed flowering times in short days, under which conditions the difference in flowering time between wild-type and phyB plants is greatest. We measured both the time the plants took to flower (DF, days to flowering) and the number of leaves at the time of flowering (LN, leaf number) in both PHYB and phyB-1 backgrounds. The results differed slightly from experiment to experiment, possibly because light conditions in our growth chamber varied (see materials and methods). Therefore, we consider as meaningful only results where we observed a statistically significant difference between SHY and shy plants in the majority of experiments. We present data for a subset of the mutants in Figure 7 and summarize our consensus results from six experiments in Table 7. Four of the mutations affected flowering time significantly.
Hypocotyl length response of shy seedlings to different fluence rates of red, blue, or far-red light. For each curve, data are normalized to the hypocotyl length of the same genotype in the dark. (A and B) Response to red light. Curves are split into two graphs for clarity, and data for wild type and phyB-1 are shown in both graphs. Each data point is the mean hypocotyl length of 20–40 seedlings, and standard deviations were generally 10–20% of the mean. Repetitions of this experiment gave similar results (data not shown). (C and D) Response to blue light (C) and far-red light (D). Shown are shy2-2, shy2-3, pom1-15, shy3-1, and shy5-1 mutants. amp1-4, shy-115, bot1-5, shy4-1, shy4-2, and shy6-1 mutants all responded similarly to wild type in both blue and far-red light (data not shown).
One shy mutation, shy5-1, delayed flowering in both PHYB and phyB-1 backgrounds (Figure 7, Table 7). In the PHYB background, shy5-1 also caused more leaves to form before flowering, whereas in the phyB-1 background, the leaf number was normal. These data suggest that the SHY5 gene product may normally antagonize the activity of PHYB.
Root lengths of phyB-1 shy seedlings in red light. shy seedling root lengths were normalized to the root length of phyB-1 seedlings. Data from four experiments were combined. Error bars indicate standard deviations of measurements. Normalized root lengths were tested for significant difference from phyB-1 root length by t-test. All genotypes except phyB-1 shy-115, phyB-1 shy3-1, and phyB-1 shy4-1 showed a difference from phyB-1 root lengths at 95% confidence or more. P values from t-tests for comparison with phyB-1 root lengths were wild type, P < 0.001; phyB-1 shy2-2, P < 0.001; phyB-1 shy2-3, P < 0.05; phyB-1 amp1-4, P < 0.001; phyB-1 shy-115, P < 0.5; phyB-1 shy-802, P < 0.001; phyB-1 pom1-14, P < 0.025; phyB-1 pom1-15, P < 0.001; phyB-1 bot1-5, P < 0.001; phyB-1 shy3-1, P = 1; phyB-1 shy4-1, P < 0.1; phyB-1 shy4-2, P < 0.001; phyB-1 shy5-1, P < 0.001; and phyB-1 shy6-1, P < 0.001.
The shy2-2 and shy3-1 mutations caused early flowering in the PHYB background, but had no effect on flowering time in the phyB-1 background. Both PHYB shy2-2 and PHYB shy3-1 mutant plants also flowered with fewer leaves than wild-type plants. Although they flowered at the normal time, phyB-1 shy3-1 plants had extra leaves. The early flowering of shy2-2 and shy3-1 mutants in the PHYB background suggests that SHY2 and SHY3 have significant regulatory functions.
As mentioned above, amp1-4 plants made leaves more quickly than AMP plants. This accelerated leaf production was accompanied by an acceleration of flowering in both PHYB and phyB-1 backgrounds. However, the effect of the amp1-4 mutation on the number of leaves made before flowering correlated poorly with its effect on flowering time. Thus, in some experiments, the leaf number was significantly greater than for the corresponding SHY plants, and in other experiments, the leaf number was significantly smaller (data not shown). This observation suggests that the accelerated flowering of amp1-4 plants may be a secondary consequence of its rapid leaf production.
The remaining shy mutations (shy2-3, shy-115, pom1-14, pom1-15, bot1-5, shy4-1, shy4-2, and shy6-1) had no effect on the number of days to flowering in either PHYB or phyB-1 backgrounds (Table 7). Three of these, shy2-3, bot1-5, and shy6-1, affected leaf number in one genetic background (Table 7). (The shy2-3 mutation caused extra leaf formation in the phyB-1 background, the bot1-5 mutation caused flowering with fewer leaves in the PHYB background, and the shy6-1 mutation caused extra leaf formation in the PHYB background.)
DISCUSSION
The phyB-1 mutation creates a stop codon and is probably a null allele (Reedet al. 1993). Therefore, the suppressor mutations described here most likely bypass the requirement for PHYB for inhibiting hypocotyl elongation. The shy mutations may identify downstream mediators of PHYB signaling, regulators of other environmental response pathways, or biochemical or metabolic functions needed for hypocotyl elongation. The occurrence of single mutations at several of the loci indicates that we have not saturated the screen. Additional known loci that can mutate to give a short hypocotyl phenotype include the AXR genes and the DET/COP/FUS genes (Fankhauser and Chory 1997).
Based on the mutant phenotypes, the best candidates for genes that regulate light responses are SHY2, SHY3, and SHY5. The recessive shy5-1 mutation suppresses all of the phyB-1 phenotypes we tested. phyB-1 shy5-1 plants have a shorter hypocotyl than the starting phyB-1 mutant, they have a longer root, and they flower late. For each phenotype, the suppression is partial in that the phyB-1 shy5-1 plants still differ from wild-type plants. This may indicate that the mutation is a partial loss-of-function allele or that the mutated function is encoded by more than one gene. Taken together, the results suggest that the SHY5 gene may encode a function that opposes the activity of PHYB, acting either downstream of PHYB in a light regulatory pathway or in a separate branch of the control network.
The shy2 mutants have several striking phenotypes that suggest that SHY2 may be an important mediator of red light responses. shy2-2 plants make leaves in the dark, respond only slightly to red light for control of hypocotyl elongation, have an elongated root in the phyB-1 background, and flower early in the PHYB background. The weaker shy2-3 mutation caused less profound effects on development than shy2-2, e.g., expanded cotyledons without obvious leaf formation in the dark. Like the shy2-2 and shy2-3 mutations described here, the previously described shy2-1 mutation is semidominant, causes cotyledon expansion in the dark, and causes upcurled leaves in the light (Kimet al. 1996). The semidominance of all three shy2 alleles is consistent with the mutations being hypomorphic, hypermorphic, or neomorphic. The frequency with which we have obtained shy2 alleles might indicate that the mutations are hypomorphic (decreased function), in which case SHY2 may normally repress phytochrome-mediated development in the dark. If the mutations are hypermorphic, SHY2 may normally activate de-etiolation in response to light. If the mutations are neomorphic, they may reveal otherwise cryptic effects of some other regulatory pathway on seedling development.
Flowering of selected PHY shy and phyB-1 shy plants in short days. (A) days to flower; (B) leaf number at time of flowering. Error bars indicate standard deviations. Shown are data for selected genotypes from three separate experiments. To facilitate comparison between different experiments, data have been normalized to the wild-type flowering time and LN in each case. In the experiments shown, the flowering times (A) for PHYB shy2-2, PHYB amp1-4, PHYB shy3-1, PHYB shy5-1, and phyB-1 shy5-1 were significantly different from the corresponding SHY plants (P < 0.05 or less);- and the leaf numbers, (B) for PHYB shy2-2, phyB-1 shy2-2, PHYB amp1-4, phyB-1 amp1-4, PHYB shy3-1, phyB-1 shy3-1, and PHYB shy5-1 were significantly different from the corresponding SHY plants (P < 0.05 or less). Flowering data from several experiments are summarized in Table 7.
Like shy2 mutants, mutants of the det/cop/fus class make leaves in the dark. However, the morphological phenotypes of shy2 mutants are quite distinct from the phenotypes of mutants of the det/cop/fus class, such as det1-1. For example, det1-1 mutant seedlings do not have curled leaves characteristic of shy2 seedlings, and shy2 mutant seedlings do not overproduce anthocyanin as det1 seedlings do. Thus, SHY2 probably regulates development in a manner different from the DET/COP/FUS gene products.
The last of the mutations with a substantial effect on both flowering and hypocotyl elongation is shy3-1. Interestingly, shy3-1 partially suppressed the flowering phenotype caused by a phyB-1 mutation, but in a PHYB background, it caused early flowering, as phyB-1 does. Thus, PHYB shy3-1 plants flowered with reduced leaf number compared to wild-type plants, whereas phyB-1 shy3-1 plants flowered with increased leaf number compared to phyB-1 plants. This dampening of the effect of PHYB on leaf number may indicate that SHY3 interacts with PHYB to control flowering (but see below). The semidominance of the shy3-1 mutation for hypocotyl length phenotypes is consistent with either a gain-of-function or loss-of-function type of allele, so the normal role of SHY3 in development could be either to transmit the PHYB signal or to antagonize it.
At first blush, the early flowering conferred by the shy2-2 and shy3-1 mutations would seem counter to the expectation (fulfilled by shy5-1) that a regulatory suppressor mutation should affect any phenotypes in the opposite sense as the starting phyB-1 mutation. However, interpretation of the flowering phenotypes is complicated by the finding that when overexpressed, PHYB causes early flowering (Bagnallet al. 1995). This suggests either that overactivation of PHYB response pathways causes early (rather than late) flowering or that overexpressed PHYB can co-opt another pathway that normally promotes flowering. For example, phytochrome A normally inhibits elongation and also promotes flowering (Nagataniet al. 1993; Parks and Quail 1993; Whitelamet al. 1993; Johnsonet al. 1994; Reedet al. 1994). In either case, a shy mutation might cause early flowering by acting similarly to overexpressed PHYB. Another possible explanation for why these shy mutations cause early flowering could be that they affect flowering time indirectly through effects on elongation. phyB-1 plants make leaves at a slower rate than wild-type plants, perhaps as a result of extra elongation growth at the expense of new organ formation (Koornneefet al. 1995; data not shown). If flowering depends in part on formation of a threshold number of leaves, then the slower rate of leaf formation in phyB-1 plants may actually delay flowering and partially compensate for the propensity of such plants to flower with fewer leaves. In that case, a mutation that suppressed the elongation phenotypes of a phyB-1 mutant would cause earlier flowering, as the shy2-2 and shy3-1 mutations do.
Summary of developmental phenotypes of shy mutantsa
Our finding of an allele of AMP (amp1-4) and presumed brassinosteroid auxotrophs (shy-115 and shy-802) in this screen underscores the probable relevance of plant hormone signaling pathways to seedling de-etiolation. The amp mutant overproduces cytokinin (Chaudhuryet al. 1993; Chin-Atkinset al. 1996), and applications of cytokinin to dark-grown seedlings can induce aspects of a de-etiolated phenotype (Choryet al. 1994; Chin-Atkinset al. 1996). However, dark-grown amp mutant seedlings make organs resembling stems or petioles rather than leaves, and the altered phyllotaxy of amp mutants cannot be mimicked by altering light conditions. These observations suggest that the connection of AMP to photomorphogenesis may be indirect. Brassinosteroids have been implicated in repressing seedling de-etiolation because mutants isolated as having partially de-etiolated phenotypes in the dark have turned out to be brassinosteroid auxotrophs (Liet al. 1996; Szekereset al. 1996). Like shy-802, mutants such as det2-1 make leaves in the dark, although the plastids do not differentiateas they do in other de-etiolated mutants such as det1 (Choryet al. 1991). It remains to be determined whether these observations reflect direct regulation of brassinosteroid physiology by light. Other workers have reported evidence suggesting that auxin and light signaling are connected. For example, Nicotiana plumbaginifolia phytochrome mutants have been found to have increased auxin levels (Kraepielet al. 1995), and some auxin-resistant mutants are short and have partially de-etiolated phenotypes in the dark (Lincolnet al. 1990; Timpteet al. 1992; Leyseret al. 1996; Cernacet al. 1997; A. Sonawala and J. W. Reed, unpublished observations). In this regard, it is interesting that shy2 mutants share some phenotypes, such as curled leaves, with axr3 mutants (Leyseret al. 1996).
The remaining shy mutations suppress only the long hypocotyl phenotype of the starting phyB-1 mutant, but do not suppress the root elongation or flowering time phenotypes. The pom1-14, pom1-15, bot1-5, shy4-2, and shy6-1 mutations actually cause the root to be shorter than that of phyB-1 plants. This observation suggests that either these mutations primarily affect a cell elongation or enlargement function, or they affect cell division rates in the root. The morphological variety among these mutants suggests that the different mutations may affect distinct aspects of growth. Those with quantitative effects but having otherwise normal shape (e.g., shy4-2 or shy6-1) may affect a control function. Those conferring aberrant morphology (e.g., pom1-14 or bot1-5) may affect part of the cellular machinery that elongates cells, synthesizes cell walls, or determines cell polarity.
The twisted hypocotyls of dark-grown pom1 mutants resemble those of dark-grown procuste1 mutants (Desnoset al. 1996), suggesting that POM1 and PROCUSTE1 have related functions. Apparently, these genes are required for proper hypocotyl elongation in the dark. Whereas the procuste1 mutants show a biphasic hypocotyl elongation curve in blue light, however, our pom1 mutants have a normal blue light response but an aberrant response in red light. POM1 is also required for elongation and control of cortical cell enlargement in roots (Hauseret al. 1995) and for proper epidermal cell differentiation (Schneideret al. 1997). POM1 therefore plays a role in growth and/or shape determination of cells of various organs.
None of the putative regulatory mutations we have identified (shy2-2, shy2-3, shy3-1, or shy5-1) is epistatic to the phyB-1 mutation for the hypocotyl length phenotype. This indicates that these mutations either cause incomplete defects in the corresponding genes or affect redundant functions. The results reinforce our view of light signal transduction as a network of interacting components rather than as a collection of linear pathways (Reed and Chory 1994). As well, light signaling is surely intimately coupled to more general developmental control mechanisms. Further analysis of the shy mutants reported here may contribute to unraveling the complex regulatory pathways that mediate control of plant development by a variety of environmental factors.
Acknowledgments
We thank P. Nagpal and A. Pepper for helpful discussions; A. Chin-Atkins, M.-T. Hauser, H. Höfte, H.-G. Nam, and T. Wada for sending seeds; S. Whitfield for help with color figures; and an anonymous reviewer for a careful reading of the manuscript. This work was supported by National Institutes of Health (NIH) grant R29-GM52456 to J.W.R. Early stages of this work were supported by grants from the National Science Foundation and the International Human Frontier Science Program to J.C., and by a U. S. Department of Energy-Energy Biosciences postdoctoral fellowship from the Life Sciences Research Foundation (Baltimore, MD) to J.W.R. J.C. is currently supported by NIH grant R01-GM52413 and is an associate investigator of the Howard Hughes Medical Institute.
Footnotes
-
Communicating editor: V. Sundaresan
- Received August 18, 1997.
- Accepted November 17, 1997.
- Copyright © 1998 by the Genetics Society of America
