Table of Contents

Phenotype of the Tomato high pigment-2 Mutant Is Caused by a Mutation in the Tomato Homolog of DEETIOLATED1

  1. Chris Bowler4
  1. Molecular Plant Biology Laboratory, Stazione Zoologica Anton Dohrn, Villa Comunale, I-80121 Naples, Italy
  1. 4 To whom correspondence should be addressed. E-mail chris{at}alpha.szn.it; fax 39-081-764-1355.

Abstract

Tomato high pigment (hp) mutants are characterized by their exaggerated photoresponsiveness. Light-grown hp mutants display elevated levels of anthocyanins, are shorter and darker than wild-type plants, and have dark green immature fruits due to the overproduction of chlorophyll pigments. It has been proposed that HP genes encode negative regulators of phytochrome signal transduction. We have cloned the HP-2 gene and found that it encodes the tomato homolog of the nuclear protein DEETIOLATED1 (DET1) from Arabidopsis. Mutations in DET1 are known to result in constitutive deetiolation in darkness. In contrast to det1 mutants, tomato hp-2 mutants do not display any visible phenotypes in the dark but only very weak phenotypes, such as partial chloroplast development. Furthermore, whereas det1 mutations are epistatic to mutations in phytochrome genes, analysis of similar double mutants in tomato showed that manifestation of the phenotype of the hp-2 mutant is strictly dependent upon the presence of active phytochrome. Because only one DET1 gene is likely to be present in each of the two species, our data suggest that the phytochrome signaling pathways in which the corresponding proteins function are regulated differently in Arabidopsis and tomato.

INTRODUCTION

Light is a critical environmental signal controlling many aspects of plant development. For example, dark-grown plants display a typical etiolated morphology with elongated hypocotyls, closed apical hooks, and unexpanded cotyledons, whereas plants grown in the light have short hypocotyls, opened apical hooks, and expanded photosynthetically active cotyledons. Light is perceived by a series of photoreceptors that can detect light within a wide spectral range. The phytochromes are the best characterized of these photoreceptors and are able to intercept light primarily within the red and far-red regions of the spectrum (Furuya and Schäfer, 1996). They exist as multigene families, for example, PHYA to PHYE in Arabidopsis, and each phytochrome is likely to have a specific photoperceptory function during plant development (Quail et al., 1995). In addition, plants contain blue/UV-A–absorbing cryptochromes and UV-B–absorbing photoreceptors.

Several models for light signal transduction in plants have been proposed. One has been deduced largely by microinjection experiments with tomato and involves G proteins, calcium, and cGMP (Bowler et al., 1994b; Mustilli and Bowler, 1997). Others are based on the genetic analysis of Arabidopsis mutants, such as deetiolated (det) and constitutively photomorphogenic (cop), which display characteristics of light-grown plants when grown in complete darkness, for example, reduced hypocotyl length, cotyledon opening and expansion, chloroplast development, and expression of light-induced genes (Chory et al., 1989, 1996; Deng et al., 1991; Chamovitz and Deng, 1996).

Although several COP and DET genes have been identified, it is not clear how the activities of their gene products are regulated by phytochrome or by the calcium- and cGMP-dependent pathways identified by microinjection (Millar et al., 1994). One problem is that the cop and det mutants display their strongest phenotypes in the absence of light, and double mutant analyses with phytochrome-deficient mutants indicate that their phenotypes are independent of phytochrome function. Furthermore, many are allelic to seedling-lethal fusca mutants (Miséra et al., 1994). Therefore, it is possible that these molecules encode global regulators involved in many plant responses, not only those mediated by light. A more targeted genetic approach to identify components of signal transduction pathways specific for phytochrome could be to isolate mutants with altered dynamics of light responses rather than mutants with constitutive phenotypes in the absence of light. Only recently have such mutants been described in Arabidopsis (e.g., spa1 [Hoecker et al., 1998] and psi2 [Genoud et al., 1998]). In contrast, several light-hypersensitive mutants have been described in tomato (Solanum lycopersicum) over the last decades, for example, high pigment (hp-1 and hp-2), atroviolacea (atv), and Intense pigmentation (Ip) (Kendrick et al., 1994).

The monogenic recessive nonallelic hp-1 and hp-2 mutants do not display any obvious phenotypes in darkness but are characterized by their exaggerated light responsiveness. For example, they display higher anthocyanin levels, shorter hypocotyls, and more deeply pigmented fruits when compared with wild-type plants. This latter characteristic is due to elevated levels of flavonoids and carotenoids and has resulted in considerable interest from the tomato industry to use these mutations in commercial varieties. hp-1 was discovered as a spontaneous mutant in 1917 at the Campbell Soup Company farms (Riverton, NJ) (Reynard, 1956) and has been the subject of most studies of hp mutants, whereas the hp-2 mutant was reported in the Italian San Marzano variety in 1975 (Soressi, 1975). Despite some initial confusion, it is now clear that hp-1 and hp-2 mutations map to different chromosomes and therefore encode different proteins (Van Tuinen et al., 1997; Yen et al., 1997), even though their phenotypes appear to be identical. Two mutant alleles have been identified at each locus, hp-1 and hp-1w, and hp-2 and hp-2j (Kerckhoffs and Kendrick, 1997; Van Tuinen et al., 1997).

Phenotypes of hp mutants appear to be identical to those obtained by ectopic expression of PHYA in tomato (Boylan and Quail, 1989). Conversely, double mutant analysis of hp-1 with PhyA- and PhyB-deficient tomato mutants has demonstrated that the hp-1 mutation can amplify responses mediated by both phytochromes and that the amplification phenotype is critically dependent upon the presence of an active phytochrome (Peters et al., 1992; Kerckhoffs et al., 1997b). It appears, therefore, that the hp mutations affect fairly specifically the responses mediated by phytochrome. The fact that no phenotypic counterparts of hp mutants have been isolated thus far in Arabidopsis indicates that the cloning of the HP genes is likely to provide important new information about the regulation of photomorphogenesis, because it should identify molecules that act specifically as negative regulators of phytochrome signal transduction.

Furthermore, tomato has become an excellent model system for studying phytochrome function. The phytochrome gene family has been characterized (Pratt et al., 1997), and mutant studies are revealing that the individual tomato phytochromes have functions similar but not identical to their Arabidopsis homologs (Kendrick et al., 1997). Consequently, tomato is now a comparative experimental system to Arabidopsis for studying the roles of individual phytochrome members, and, in addition, tomato promises to reveal the roles of phytochrome in fruit ripening.

In this report, we have analyzed the phenotype of hp-2 mutants and have cloned the HP-2 gene. Surprisingly, HP-2 encodes the tomato homolog of Arabidopsis DET1. Given the very different characteristics of the two mutants, we propose that the signal transduction pathways controlling photomorphogenesis may be wired differently in tomato and Arabidopsis.

RESULTS

Anthocyanin Pigment Production in hp-2 Mutants

Although the phenotype of the hp-2 mutant is generally considered to be identical to that of the hp-1 mutant, it has not been examined in great detail. Therefore, we performed an analysis of hp-2 mutant characteristics by using hp-2 and hp-2j mutant seedlings grown in the light and in the dark.

Consistent with the light-hypersensitive hp-2 phenotype, both hp-2 and hp-2j seedlings displayed a light-dependent reduction in hypocotyl length and increase in anthocyanin content (Kerckhoffs et al., 1997a; Figures 1A and 1B). For both responses, the hp-2j mutant appears to be the stronger of the two alleles.

To determine the tissue distribution of anthocyanin pigments in hp-2 and hp-2j mutants, we performed light microscopy with hand-cut sections prepared from a range of plant material. In hypocotyls of young hp-2 seedlings, anthocyanin pigments were present specifically within subepidermal cells, as was previously reported for hp-1 mutants (Von Wettstein-Knowles, 1968; Figure 2B). This cell specificity for anthocyanin biosynthesis also was found in wild-type seedlings, although in contrast to hp-2 seedlings, anthocyanin accumulation was only transient and was diluted out after 3 to 4 days (Figure 2A).

In leaves of 3-week-old seedlings, anthocyanin accumulation also was localized to specific cell types in hp-2 and hp-2j mutants. Most notably, pigmentation was observed in the collenchyma and within the basal cells of trichomes (Figures 2D, 2F, and 2H). In wild-type plants grown under the same conditions, anthocyanin pigmentation was not evident (Figures 2C, 2E, and 2G). However, wild-type plants exposed to extreme conditions, such as high light in combination with low temperatures, displayed identical anthocyanin localization patterns (data not shown). This observation indicates that although anthocyanin pigments are overproduced significantly in hp-2 and hp-2j mutants, cell-type specificity nonetheless is not affected.

Chloroplast Biogenesis in hp-2 Mutants

Immature fruits of hp-2 and hp-2j mutants display very high levels of chlorophyll pigments compared with wild-type fruits (Figures 2I and 2J). In hp-1 mutants, this darker pigmentation appears to be due to a higher number of chloroplasts per cell (Yen et al., 1997). In the leaves of hp-2 and hp-2j mutants, chloroplast number did not appear to be affected; indeed, chlorophyll content was not higher than in wild-type leaves (data not shown). However, whereas roots of wild-type seedlings were white, the roots of hp-2 and hp-2j seedlings were green, indicating the presence of chloroplasts. Indeed, confocal imaging of root sections revealed chlorophyll fluorescence within the cortex and pericycle cells of hp-2 and hp-2j mutant roots but not in wild-type roots (Figures 3A and 3B).

Figure 1.

Changes in Hypocotyl Length and Anthocyanin Content in hp-2 and hp-2j Seedlings Are Light Dependent.

(A) Hypocotyl length (in centimeters) of wild-type (wt), hp-2, and hp-2j tomato seedlings grown at 25°C for 5 days in absolute darkness (Dark) or in a 16-hr-light and 8-hr-dark photoperiod (Light). Error bars indicate ±se.

(B) Anthocyanin content (per milligram fresh weight [FW]) from cotyledons and hypocotyl of wild-type (wt), hp-2, and hp-2j light-grown tomato seedlings.

Values are the means of 10 seedlings, and the experiments were repeated three times.

Electron microscopic observations of wild-type roots revealed the presence of only a few small amyloplasts in most cell types (Figure 3C). By contrast, in hp-2 and hp-2j roots, cortical cells contained large numbers of plastids with very well-developed thylakoid membrane systems similar to those observed in leaves (Figure 3E), whereas plastids within pericycle cells were less developed and fewer in number (Figure 3D).

hp-2 Mutants Display Only Minor Dark Phenotypes

Neither hp-1 nor hp-2 mutants have been reported to display dark phenotypes, such as reduced hypocotyl length, opened apical hooks, or enlarged cotyledons, except for very small reductions in hypocotyl length in hp-1w and hp-2j seedlings that are not statistically significant (Kerckhoffs et al., 1997a; Figure 1A). Nonetheless, electron microscopic observations revealed the presence of partially developed chloroplasts in both hp-2 and hp-2j mutants (Figure 4), a well-known characteristic of Arabidopsis cop and det mutants (Chory et al., 1989; Deng et al., 1991). Prolamellar bodies were generally absent (hp-2j) or reduced in size (hp-2), and prothylakoid membranes were apparent (particularly in hp-2j). The presence within the same plastid of both prothylakoid membranes and prolamellar bodies has not been reported previously for Arabidopsis cop and det mutants and may be a result of hypersensitivity in the hp mutant background to the active phytochrome normally present in mature seeds.

Gene Expression Is Deregulated in the Light

To examine the effects of hp-2 and hp-2j mutations on gene expression, we performed RNA gel blot analysis of light-regulated gene expression by using CHS (encoding chalcone synthase) and CAB (encoding the chlorophyll a/b binding protein) gene fragments as probes. In agreement with the weak phenotypes of hp-2 and hp-2j seedlings grown in darkness, we did not observe any dramatic alteration of CHS and CAB gene expression. Nonetheless, CHS and CAB mRNA levels were slightly higher in plants carrying the hp-2j allele when compared with hp-2 and wild-type seedlings (Figure 5). Interestingly, CHS mRNA appeared to be of a slightly smaller size than that found in light-grown material, perhaps indicating differential light-dependent splicing.

Consistent with the exaggerated photoresponsiveness of hp-2 and hp-2j mutants, CHS mRNA levels were significantly enhanced compared with wild-type seedlings after light irradiation for 48 hr (Figure 5). hp-2j seedlings contained higher levels of CHS transcripts than did hp-2, and CHS mRNA was particularly abundant in hypocotyls. Detailed time-course experiments indicated that CHS gene expression was both anticipated and stronger in hp-2 and hp-2j mutants compared with wild-type seedlings (A.C. Mustilli and C. Bowler, manuscript in preparation). In contrast to those of CHS, CAB mRNA levels were higher in cotyledons than in hypocotyls and were slightly lower in light-exposed hp-2 and hp-2j mutant seedlings when compared with the wild type (Figure 5).

Figure 2.

Phenotypes of hp-2 Seedlings.

(A) and (B) Transverse sections of hypocotyls of wild-type (A) and hp-2 (B) seedlings.

(C) and (D) Transverse sections of young leaves of wild-type (C) and hp-2 (D) seedlings.

(E) and (F) Longitudinal sections through vascular tissue in leaves of wild-type (E) and hp-2 (F) seedlings.

(G) and (H) Trichomes on the adaxial side of leaves of wild-type (G) and hp-2 (H) seedlings.

(I) and (J) Immature fruits of wild-type (I) and hp-2 (J) plants.

In hp-2 seedlings, anthocyanin pigments are located within the subepidermal cells of hypocotyls (B) and within the collenchyma ([D] and [F]) and basal cells of trichomes (H) on the upper surface of leaves. Note that trichomes directly above the vasculature do not contain anthocyanin (F). Hand-cut sections were derived from 3-week-old greenhouse-grown seedlings. hp-2j mutants displayed the same phenotypes as did hp-2 plants, but the hp-2j mutant phenotypes were slightly more severe (data not shown). Bars in (A) to (D) = 100 μm; bars in (E) and (F) = 20 μm; bars in (G) and (H) = 50 μm.

Phenotype of the Tomato hp-2 Mutant Is Caused by Mutation in DET1

The hp-2 mutation has been mapped previously using restriction fragment length polymorphism (RFLP) analyses of a segregating population derived from a cross between the hp-2 mutant (S. lycopersicum) and S. pennellii. Mapping data with a second backcross (BC-2) population indicated a position close to the centromere of chromosome 1, within a cluster of several RFLP markers (Balint-Kurti et al., 1995; Broun and Tanksley, 1996; Van Tuinen et al., 1997). We discovered that the CT151 RFLP marker included in this cluster had a high degree of similarity to the Arabidopsis DET1 gene, mutations in which were previously identified as causing the deetiolated phenotype of det1 mutants (Pepper et al., 1994). Therefore, we examined whether the hp-2 mutant phenotype was a result of mutation in the tomato DET1 (TDET1) gene.

Comparison of CT151 with Arabidopsis DET1 revealed that CT151 has homology to the 3′ end of DET1. We isolated a full-length cDNA encoding TDET1 by using 5′ rapid amplification of cDNA ends (Loh et al., 1989). Subsequently, the TDET1 gene was isolated from genomic libraries in λ DASH and λ FIXII. Alignment of the TDET1 cDNA and genomic sequences revealed the presence of 10 introns, which is similar to the number found in Arabidopsis (Figure 6A). Nine introns are located in the same positions as those of the Arabidopsis DET1 gene, whereas intron 2 of TDET1 is not present in the Arabidopsis homolog (Figures 6A and 6B; Pepper et al., 1994). Comparative protein sequence analysis between Arabidopsis DET1 and TDET1 shows 81.3% similarity and 74.8% identity (Figure 6B). There are no major differences between the sequences, except for a small deletion of 16 amino acids at the center of TDET1, suggesting that the tomato and Arabidopsis genes are true homologs.

The function of DET1 in Arabidopsis is not known, although the protein is largely hydrophylic and contains a bipartite nuclear localization signal (NLS) (Figure 6B). As expected, DET1 is found predominantly in the nucleus in Arabidopsis cells (Pepper et al., 1994), and the high conservation of TDET1 within the NLS (Figure 6B) indicates that TDET1 is likely to be nuclear localized in tomato as well. Indeed, fusions of TDET1 with the green fluorescent protein (GFP) confirmed that the tomato protein also can mediate nuclear translocation (Figure 6C).

A DET1 homolog is not present in the yeast genome or in any prokaryotic genome sequenced to date (data not shown). However, mouse and human expressed sequence tags with homology to DET1 have been identified (Figure 6B), as well as a Drosophila homolog (S. Pimpinelli, personal communication).

TDET1-encoding cDNAs from the wild type and hp-2 and hp-2j mutants were amplified with specific primers from leaf mRNA (see Methods). In the hp-2j TDET1 cDNA, we found a C-to-T mutation in exon 11 (nucleotide 1640), which gave rise to the substitution of a conserved proline for a serine residue in the C-terminal region of the protein (amino acid 498) (Figures 6A and 6B). In the hp-2 mutant, we found an alternative splicing of intron 10, which led to a deletion of the first three amino acids (Gly, Pro, and Glu) of exon 11 within the second putative NLS (amino acids 478 to 480), presumably resulting in mislocalization of the protein (Figures 6A and 6B). To determine the site of the mutation that caused alternative splicing, we directly sequenced the 3′ end of the TDET1 gene isolated from the hp-2 mutant. We found an AG-to-TG substitution in the consensus 3′ splice junction of intron 10 (Figure 6D). These results demonstrate that the phenotype of the hp-2 and hp-2j mutants is caused by mutation in the TDET1 gene.

Expression analysis of TDET1 mRNA in wild-type seedlings revealed no major differences between dark- and light-grown material (Figure 6E and data not shown). Conversely, no major differences were observed in TDET1 steady state mRNA levels between the wild type and the two mutants (data not shown). However, the presence of transcripts deriving from the mutated gene in the hp-2 mutant could be readily detected in reverse transcriptase–polymerase chain reaction (RT-PCR) products because the mutation results in the generation of a PvuII restriction site (CAGCTG) at the junction between exons 10 and 11 (Figures 6D and 6E). RT-PCR and subsequent restriction enzyme digestion revealed that TDET1 mRNA in the hp-2 mutant was a mixture of mutated and wild-type transcripts (Figure 6E), indicating that at a reduced efficiency the mutated gene can be spliced normally and that this mutant is therefore not a null. RT-PCR under quantitative conditions followed by DNA sequencing revealed that the hp-2 mutant contains ~10% correctly spliced TDET1 mRNA (data not shown).

Comparison of Tomato hp-2 and Arabidopsis det1 Mutants

As is clear from the results presented in Figures 1, 2, 3, 4 and 5, hp-2 mutants are phenotypically different from det1 mutants. Most conspicuously, they do not display dark phenotypes, such as reduced hypocotyl length, opened apical hooks, or enlarged cotyledons, whereas these were selection criteria for the isolation of det and cop mutants in Arabidopsis (Chory et al., 1989; Deng et al., 1991). Conversely, det1 mutants can even develop true leaves and floral buds in prolonged darkness. No such phenotypes are observed in hp tomato seedlings grown in the dark.

Arabidopsis det1 mutants also are characterized by high-level expression of light-regulated genes such as CHS and CAB in the dark, whereas in the light, CHS and CAB gene expression is similar to that observed in wild-type plants (Chory et al., 1989). In tomato hp-2 mutants, dark expression of CHS and CAB genes is only very slight, and deregulation of gene expression is essentially light dependent (Figure 5).

Also in the light, det1 mutant phenotypes contrast with those of hp mutants. Although both are small and have reduced apical dominance, det1 mutants are pale green (Chory and Peto, 1990), whereas hp mutants are deeply pigmented (Figure 2).

Despite the above-mentioned differences, plastid development in darkness in the cotyledons of hp-2 and hp-2j seedlings (Figure 4) is similar to that observed in Arabidopsis det1 mutants (Chory et al., 1989). The development of chloroplasts in the roots of these mutant seedlings grown in the light (Figure 3) also is consistent with the phenotype of det1 mutants (Chory and Peto, 1990) and is even more pronounced. However, whereas in Arabidopsis det1 mutants this observation was interpreted as indicating an alteration in cell-type specificity, in tomato this is not the case because, unlike in Arabidopsis, old roots of wild-type tomato plants commonly display some chloroplast development (data not shown). Similarly, whereas in Arabidopsis det1 mutants anthocyanins are ectopically produced in all cell types (Chory and Peto, 1990; Miséra et al., 1994), the cell-type specificity of anthocyanin production is maintained very clearly in tomato hp-2 mutants (Figure 2). Consequently, the phenotypes of hp-2 and hp-2j mutants can be interpreted as being due to exaggerated responses to light rather than to alterations of cell-type specificity resulting from ectopic expression.

Figure 3.

Roots of Light-Grown hp-2 Seedlings Contain Fully Developed Chloroplasts.

(A) and (B) Confocal images from representative roots of wild-type (A) and hp-2 (B) seedlings. Images are three-dimensional reconstructions of Z optical sections 1.5 to 2 μm thick. Cell wall autofluorescence is shown in green pseudocolor, and chlorophyll autofluorescence is indicated in red.

(C) to (E) Electron microscopic images from roots of wild-type (C) and hp-2 ([D] and [E]) seedlings. Electron microscopic images of plastids were taken from cortical cells in (C) and (E) and pericycle cells in (D).

hp-2j roots displayed the same phenotypes as shown here for hp-2 mutants, but the hp-2j phenotypes were slightly more severe (data not shown). Bars in (A) and (B) = 100 μm; bars in (C) to (E) = 1 μm.

det1 seedlings have been reported to display a strongly enhanced expression of stress-related genes such as those encoding pathogenesis-related (PR) proteins and glutathione reductase (Mayer et al., 1996). To examine whether this was also the case in hp-2 and hp-2j mutants, we hybridized RNA gel blots with a probe encoding tomato PR-1b1 (Tornero et al., 1997). Although hypocotyls of hp-2 seedlings reproducibly displayed PR-1b1 gene expression at slightly higher levels compared with wild-type seedlings, this effect was very weak compared with that observed in Arabidopsis det1 mutants and has never been observed in hp-2j mutants (Figure 5). Nonetheless, this result indicates that, as with det1, the hp-2 mutation does not affect only light-regulated genes. An alternative explanation is that there is considerably more cross-talk between different regulatory pathways than is currently appreciated. In Arabidopsis, CAB gene expression recently has been shown to be regulated by low temperature in a light-independent manner (Capel et al., 1998).

Cytokinin Can Phenocopy the hp Mutant

Previous observations in Arabidopsis have shown that a det1 mutant phenotype can be phenocopied by the exogenous application of cytokinin (Chory et al., 1994). To determine the effects of cytokinin treatment in tomato, we treated wild-type seedlings with different concentrations of cytokinin in the dark and in the light (Figure 7). Interestingly, cytokinin does not phenocopy the hp-2 mutation in tomato. In dark-grown seedlings, although hypocotyls are shorter in the presence of cytokinin, apical hook opening, cotyledon expansion, and anthocyanin biosynthesis were not observed (Figures 7A and 7B), even after prolonged periods (up to 3 weeks; data not shown). However, in the light, cytokinin can phenocopy the hp-2 mutation: seedlings displayed shorter and thicker hypocotyls and accumulated high levels of anthocyanin (Figures 7A to 7C). These results therefore indicate that at least as far as the effects of cytokinin are concerned, the hp mutation in tomato is equivalent to the det1 mutation in Arabidopsis.

Figure 4.

Partial Plastid Development in Dark-Grown hp-2 and hp-2j Mutant Seedlings.

Electron microscopy of representative plastids from cotyledons of 5-day-old dark-grown wild-type (wt), hp-2, and hp-2j tomato seedlings is shown. Bars = 1 μm.

Double Mutant Analyses

Double mutant analysis with det1 and phytochrome-deficient mutants in Arabidopsis indicates that the det1 mutation is completely epistatic to photoreceptor mutations (Chory, 1992). This has been interpreted as meaning that DET1 acts downstream of phytochrome. To determine the relationship between TDET1 and phytochrome function, we examined the effects of the hp-2 mutation in an aurea (au) mutant background, a phytochrome chromophore-deficient mutant (Terry and Kendrick, 1996). In contrast to its counterpart double mutant in Arabidopsis, au hp-2 is similar to the single mutant au. For example, hypocotyls are elongated and anthocyanin accumulation is limited (Figures 8A to 8C). Although in genetic terms this result therefore infers that au is epistatic to hp-2, we nonetheless consider it more likely that TDET1 acts downstream of phytochrome, as proposed for Arabidopsis, but that its activity as a negative regulator is strictly dependent upon the presence of active phytochrome. This requirement is clearly not observed in Arabidopsis (Chory, 1992). The small but significant reduction in hypocotyl length and the small increase in anthocyanin found in the au hp-2 double mutant compared with au (Figure 8) are likely, therefore, to be a result of hypersensitivity caused by the hp-2 mutation toward the low amounts of functional phytochrome present in the au mutant, which is estimated to be ~3% of wild-type levels (Parks et al., 1987; Adamse et al., 1989). Similar results have been found in the au hp-1 double mutant (Peters et al., 1992; Kerckhoffs et al., 1997b).

Only One DET1 Homolog in Tomato

The fact that the hp-2 mutant phenotype is strictly light and phytochrome dependent could suggest that in tomato there is either a redundancy in DET1 gene function or that the regulatory networks controlling photomorphogenesis are wired differently compared with those in Arabidopsis. To identify whether there may be other DET1 homologs in tomato, DNA gel blot analysis was performed with tomato genomic DNA at medium stringency. The results indicate that, as in Arabidopsis (Pepper et al., 1994), there is only one DET1 gene in tomato (Figure 6F).

DISCUSSION

Ever since the photobiology of hp mutants was first described (Adamse et al., 1989), it has been thought that their exaggerated responsiveness to light must be a result of a mutation in a key negative regulator of the amplification steps in phytochrome signal transduction. Cloning of HP genes therefore promised to identify new components functioning specifically in phytochrome signal transduction pathways. In this report, we have demonstrated that the HP-2 gene encodes the tomato homolog of DET1. This is surprising, given the dramatic differences between Arabidopsis det1 and tomato hp-2 phenotypes. Most conspicuously, hp-2 mutants do not display any strong visible phenotypes in darkness, whereas deetiolation in the dark was the selection scheme used to isolate det1 mutants. Furthermore, whereas det1 mutations are epistatic to phytochrome mutations, analysis of the au hp-2 double mutant reveals that the appearance of the hp-2 mutant phenotype is strictly dependent upon the presence of active phytochrome.

The fact that mutation of TDET1 in tomato is not manifested strongly in the dark or in the absence of functional photoreceptors may be related to the observation that in tomato, anthocyanin biosynthesis is strictly light dependent, whereas in Arabidopsis, it can be induced in the absence of light (Chory et al., 1989; Deng et al., 1991; Peters et al., 1992; Miséra et al., 1994; Pepper et al., 1994; Kwok et al., 1996; Kerckhoffs et al., 1997b). This may also explain why no constitutive deetiolated mutants, such as cop and det, have ever been reported in tomato. Conversely, seedling-lethal mutants that display high levels of anthocyanin in the embryos of developing seeds, such as Arabidopsis fusca mutants, also have never been observed in tomato, even though they are commonly found in mutant screens in Arabidopsis and often represent strong alleles of cop and det mutants (Miséra et al., 1994; Kwok et al., 1996). The brown seed (bs) mutants are perhaps the most analogous in tomato (Rick, 1990). Although one bs mutation has been found to map to the vicinity of the HP-2 gene on chromosome 1 (Balint-Kurti et al., 1995; Broun and Tanksley, 1996; Van Tuinen et al., 1997), the brown coloration of bs seeds is due to enhanced pigmentation of the endosperm and not of the embryo, and bs is not mutated in the TDET1 gene (data not shown).

Figure 5.

Gene Expression in Wild-Type, hp-2, and hp-2j Seedlings.

Seedlings were grown at 25°C for 5 days in absolute darkness (D), followed by 2 days in continuous white light (L). RNA was extracted from whole seedlings (dark) or from cotyledons (C) and hypocotyls (H) (light). Modifications in gene expression in hp-2 and hp-2j seedlings compared with wild-type seedlings (wt) are shown to be principally light dependent. Ten-microgram samples of total RNA were loaded on gels and analyzed for expression of CHS1 (CHS), CAB6 (CAB), and PR1-1b1 (PR, for pathogenesis related) genes after RNA gel blotting. The 30S rRNA is shown as a control for loading.

Figure 6.

Structure of TDET1 and Sites of hp-2 and hp-2j Mutations.

(A) Exon–intron structure of tomato TDET1 gene and sites of mutations in the hp-2 and hp-2j alleles. Stippled boxes indicate exons, and slashes indicate introns. The scale bar is indicated.

(B) Alignment of the deduced amino acid sequences of S. lycopersicum TDET1, Arabidopsis DET1, and the mammalian expressed sequence tags (ESTs) with homology to DET1 (determined using the Clustal method [Higgins and Sharp, 1988]). The putative bipartite NLS is overlined. The amino acids missing in the hp-2 mutant are indicated with asterisks, and the amino acid substitution in hp-2j is denoted by a plus sign. The mammalian sequence is a compilation of derived amino acid sequences from mouse and human ESTs (GenBank accession numbers AA756238, AA236057, AA050184, and W64359). Boxed residues indicate conserved amino acids; dashes indicate arbitrary insertions.

(C) Targeting of the GFP–TDET1 fusion protein to nuclei of tobacco BY-2 suspension-cultured cells. A non-nuclear-targeted fusion protein, PHYA–GFP, is shown as a control. See Methods for more details. Scale bars = 20 μm.

(D) Donor and acceptor splicing sites of intron 10 from the wild type and hp-2 mutant. Brackets indicate splice sites, dots indicate internal intron sequence (data not shown), and amino acids are shown in the one-letter code. The asterisk indicates the mutated nucleotide in the hp-2–derived sequence.

(E) Detection of mutated TDET1 mRNA sequences in dark (D)- and light (L)-grown hp-2 seedlings. A 0.9-kb TDET1 fragment was amplified by reverse transcriptase–polymerase chain reaction (RT-PCR) from the wild type (wt) and the hp-2 mutant and subsequently digested with PvuII, as described in Methods. The hp-2 mutant sequence can be detected by PvuII digestion, which yields 0.7- and 0.2-kb fragments. Wild-type and mutant sequences were verified by DNA sequencing. RT-PCR was performed under nonquantitative conditions. Molecular lengths are shown at left in kilobases.

(F) DNA gel blot analysis of wild-type tomato genomic DNA. DNA was digested with EcoRI (lane 1), EcoRV (lane 2), DraI (lane 3), HindIII (lane 4), and BamHI (lane 5). A 1.8-kb TDET1 cDNA fragment was used as probe, and hybridization patterns were fully consistent with the presence of a single TDET1 gene in tomato, based on the genomic sequence (GenBank accession numbers AJ224356 and AJ224357). The only additional hybridizing bands at lower stringencies were nonspecific organellar DNA fragments (data not shown). Molecular lengths are shown at left in kilobases.

Detailed analysis of hp-2 mutant phenotypes in darkness has revealed that, as in det1 mutants, plastids show partially developed morphologies. Furthermore, several light-dependent responses also have escaped attention previously. For example, hp-2 roots display some chloroplast development, and stress-related genes such as PR-1b1 are partially deregulated. However, the observation that the tissue specificity of anthocyanin biosynthesis is not affected in hp-2 mutants is in notable contrast to det1 mutants (Chory and Peto, 1990; Miséra et al., 1994).

It could be argued that the discrepancies observed between hp-2 and det1 mutants could result if the hp-2 and hp-2j mutant alleles were weaker than were the reported det1 alleles. Whereas the hp-2 mutation is certainly not null (Figure 6E), whether the hp-2j mutation is null awaits further analysis, although it should be noted that the mutated proline residue in the hp-2j mutant is conserved in both plant sequences. We believe, however, that the absence of several well-known det1 phenotypes in hp-2 and hp-2j mutants (as discussed above) is unlikely to be due to the severity of the det1 alleles compared with the hp alleles because (1) these are qualitative not quantitative differences (e.g., altered tissue specificities in det1 mutants); (2) weak alleles of det1 in Arabidopsis (e.g., det1-1) display visible dark phenotypes (Pepper et al., 1994); (3) even heterozygous det1 plants show clearly deregulated gene expression in the dark (Pepper et al., 1994); and (4) in tomato, cytokinin can phenocopy an hp mutant phenotype rather than a det1 mutant phenotype (Figure 7). Furthermore, some of the phenotypes of hp-2 and hp-2j mutants are actually stronger than those observed in det1 mutants, for example, chloroplast development in the dark and in the roots of light-grown seedlings. Consequently, whereas the role of DET1 as a specific negative regulator for photomorphogenesis can be questioned in Arabidopsis, the dark- and light-dependent phenotypes of hp mutants in tomato are consistent with a specific role for DET1 in regulating light responses in this species.

It is also possible that different threshold amounts of DET1 protein are needed for different light-regulated responses in tomato compared with Arabidopsis, because the mutant phenotypes clearly demonstrate that phytochrome action in etiolated seedlings is under constraint of the DET1 protein. The availability of clearly null mutant alleles in tomato should resolve these issues. We are currently in the process of generating such mutants by transposon mutagenesis.

Figure 7.

Cytokinin Can Phenocopy the hp-2 Mutant Phenotype in Light-Grown Tomato Seedlings.

(A) Phenotype of wild-type (wt) seedlings grown in the absence (−) or the presence (+) of 500 μg/L benzoaminopurine compared with hp-2 seedlings grown in the dark (left) and in the light (right).

(B) and (C) Hypocotyl length (in centimeters) and anthocyanin content (per mg fresh weight [FW]), respectively, of wild-type tomato seedlings grown in the presence of 0, 1, 5, 20, 100, and 500 μg/L benzoaminopurine (bars 1 to 6). Seedlings were grown at 25°C for 5 days in absolute darkness (Dark) or in a 16-hr-light and 8-hr-dark photoperiod (Light; [C]). In (B), values are the mean of 10 seedlings. Highly similar results were obtained with the cytokinin zeatin (data not shown). Error bars indicate ±se.

The most dramatic difference between hp-2 and det1 mutants is therefore the fact that hp-2 mutations are able to sensitize light responses, whereas det1 mutations cannot. For example, whereas light inducibility of CHS gene expression in hp-2 mutants is more than double that observed in wild-type seedlings (Figure 5), in det1 mutants, CHS gene photoregulation is essentially lost, which is the result of both higher expression levels and ectopic expression in all cell types in the dark (Chory and Peto, 1990; Miséra et al., 1994). Therefore, we propose that even though the function of the DET1 protein is likely to be identical in both tomato and Arabidopsis, the regulatory circuitry in which it operates is not the same in the two plants—in tomato, activation of photomorphogenesis is strictly dependent upon a light stimulus, whereas in Arabidopsis, photomorphogenic programs can be uncoupled from the light. Most simply, the signal transduction pathways regulating photomorphogenesis in Arabidopsis may have a low basal activity in darkness (i.e., a flux or dark current), whereas in tomato, they may be completely inactive under the same conditions. This hypothesis is reinforced further by the very different epistatic relationships of hp-2 and det1 mutations with mutations affecting phytochrome function.

It will be extremely interesting to determine how TDET1 activity is desensitized by light and to examine how its activity is influenced by calcium and cGMP, which are the second messengers identified in tomato as being involved in phytochrome control of anthocyanin biosynthesis and chloroplast development, respectively (Bowler et al., 1994a). Based on our knowledge of the workings of these pathways, we could hypothesize that the main desensitizer of TDET1 activity in anthocyanin-producing cells of leaves and hypocotyls is cGMP, thus resulting in the enhanced anthocyanin biosynthesis and concomitant reduction in chloroplast development (via reciprocal control) in hp-2 mutants (Bowler et al., 1994b). In other tissues, such as in roots and immature fruits, TDET1 activity may be principally modulated by calcium, therefore resulting in enhanced chloroplast development in hp-2 mutants. Microinjection experiments should determine the validity of such hypotheses.

The function of DET1 remains a mystery. Although it is a nuclear protein, it has not been reported to bind DNA directly (Pepper et al., 1994), and its ability to influence the expression of a large number of genes indicates that it may regulate transcription at a higher level, perhaps by organizing repressive regions of chromatin in the dark (Chory et al., 1996; Mustilli and Bowler, 1997). The fact that DET1 homologs have now been found in mouse, human, and Drosophila but not in yeast or prokaryotes would indicate a role specific for multicellular organisms, perhaps as a general regulator of chromatin structure. Biochemical and cell biological experiments will be necessary to assess this possibility.

Based on the findings presented in this study, it is likely that only a few key regulators are responsible for controlling light responses in all higher plants. The isolation and characterization of other COP/DET/FUSCA homologs in tomato therefore will be of great interest in investigating whether alteration of their functions also results in light-hypersensitive phenotypes, such as those seen in hp-1 and atv mutants. The availability of such genes in tomato, as well as information from mutational analysis, would allow an evaluation of their role within the calcium- and cGMP-dependent phytochrome signaling pathways previously elucidated by microinjection in this species and may allow the construction of a single unified model for phytochrome signal transduction.

Figure 8.

The au Mutation Is Epistatic to hp-2.

(A) Light-grown phenotypes of wild-type (wt), au, hp-2, and au hp-2 tomato seedlings.

(B) and (C) Hypocotyl length (in centimeters) and anthocyanin accumulation (per seedling), respectively. Seedlings were grown at 25°C for 6 days under a 16-hr-light and 8-hr-dark photoperiod. Values are the mean of 15 seedlings, and the experiments were repeated three times. In (B), error bars indicate ±se.

METHODS

Plant Material and Growth Conditions

The hp-2 and hp-2j exaggerated photoresponse mutants, the au hp-2 double mutant, and the corresponding wild-type tomato seeds (Solanum lycopersicum cv Money Maker) were kindly provided by R.E. Kendrick and M. Koornneef (Wageningen Agricultural University, The Netherlands). Seeds were surface sterilized and directly sown in magenta boxes (Sigma) containing 4.3 g/L Murashige-Skoog salts (Sigma) and 0.8% agar. After 2 days pregermination in darkness, seedlings were grown at 25°C either in a 16-hr-light, 8-hr-dark photoperiod or in continuous dark, as appropriate. For cytokinin treatment, seeds were sown in the presence of different concentrations of benzylaminopurine (Sigma).

Anthocyanin Assays

Anthocyanins were extracted from cotyledons, hypocotyls, and whole seedlings with 0.5 mL of acidified (1% HCl) methanol for 48 hr in darkness with shaking. The extracts were separated by the addition of 0.4 mL of H2O and 1 mL of chloroform, followed by centrifugation for 5 min at 3000 rpm. The absorbance of the upper phase was determined spectrophotometrically at 535 nm (A535), and the anthocyanin content was calculated as (A535)/mg fresh weight or (A535)/seedling.

Isolation and Hybridization of Nucleic Acids

Total RNA was prepared from tomato leaves according to Verwoerd et al. (1989). RNA gels (10 μg per lane) were blotted onto Hybond N+ membranes (Amersham) and hybridized with random primed probes (see below). Hybridization was performed for 24 hr at 50°C in phosphate buffer (7% SDS, 0.5 M NaPO4, pH 7.2, and 1 mM EDTA), followed by 20-min washes in 40 mM NaPO4, pH 7.2, 1% SDS, and 1 mM EDTA. As probes, we used the tomato chalcone synthase gene CHS1 (O'Neill et al., 1990), the chlorophyll a/b binding protein gene CAB6 (Piechulla et al., 1991), and the pathogenesis-related protein gene PR-1b1 (Tornero et al., 1997). All RNA gel blots were repeated at least twice, using different samples.

DNA extraction from tomato leaves was performed as described by Dellaporta et al. (1983). Five-microgram samples were digested with a range of different restriction enzymes overnight and run on 0.8% agarose gels. Alkaline DNA gel blotting onto nylon Hybond N+ membranes was as described by the supplier (Amersham). The 1.8-kb fragment of tomato DEETIOLATED1 (TDET1) cDNA, labeled by random priming, was used as probe. Hybridization conditions were as described above.

Isolation and Sequencing of Genomic Clones, cDNAs, and Polymerase Chain Reaction Products

Reverse transcription of the 5′ sequence of the TDET1 mRNA was performed using the 5′ rapid amplification of cDNA ends system (Gibco BRL) with total tomato leaf RNA with the oligonucleotide 5′-CATCAACACTGCCAAAC-3′, derived from the sequence of the restriction fragment length polymorphism (RFLP) marker CT151. An 0.8-kb polymerase chain reaction (PCR) product containing the 5′ end of the TDET1 cDNA was obtained after two different amplification reactions with TaqI polymerase (Perkin Elmer) by using two CT151-derived nested primers (5′-GAAAGCAGCCGTTGCT-3′ and 5′-AGTTCATCATCTTCACGGC-3′, respectively) with the provided anchor primer. The 0.8-kb fragment was directly sequenced on both strands with Thermosequenase (Amersham).

Total cDNA from the wild type (cv Money Maker) and the corresponding hp-2 and hp-2j mutants were obtained by reverse transcription (RT) using avian myeloblastosis virus reverse transcriptase (Promega) of poly(A) mRNA isolated from leaves by using oligo(dT) Dynabeads (Dynal, Skøyen, Norway). From these total cDNAs, TDET1 cDNA sequences were PCR amplified with specific primers (5′-GTATGATTCACTAGTTTAATGCTGCTGAAAG-3′ and 5′-CCCATACTAGTCGTCTTGGCACTCTATCAAG-3′) by using the Expand High Fidelity system (Boehringer Mannheim) and subcloned in pBluescript SK+ (Stratagene, La Jolla, CA) as SpeI fragments; four independent clones were sequenced on both strands. The wild-type TDET1 cDNA and genomic sequences have GenBank accession numbers AJ224356, AJ224357, and AJ222798.

For the RT-PCR analysis presented in Figure 6E, TDET1 sequences were amplified from total cDNAs derived from dark- and light-grown wild-type and hp-2 seedlings by using the primers 5′-GTCTGCCTCAATCAGAAACTTCTTTCC-3′ and 5′-GTGATTTCTAGGTTGATTTCAATCTAGAG-3′. PCR products were subsequently purified, digested with PvuII, and run on 1% agarose gels.

Tomato genomic libraries in λ DASH and λ FIXII (kindly provided by, respectively, J. Giovannoni, Texas A&M University, College Station, TX, and the Tomato Genome Center, Rehovot, Israel) were screened using standard methods (Sambrook et al., 1989) with a 32P-labeled CT151 fragment as a probe. Overlapping fragments from different recombinant phages were subcloned in pBluescript SK+ for sequencing of both strands.

Oligonucleotides 5 ′-GAAGGTAATTTTATATTAAACATAGAATAGA-3′ and 5′-GTGATTTCTAGGTTGATTTCAATCTAGA-3′ were used for amplification of the 3′ end of the TDET1 gene from hp-2 genomic DNA. The 1.3-kb PCR product was directly sequenced on both strands beginning with the primer 5′-CAAATCGGTAACATAT-3′ by using a Thermosequenase kit (Amersham).

Microscopy

To observe anthocyanin pigment localization in leaves and hypocotyls of the wild type and hp mutants, we prepared hand-cut sections from 3-week-old greenhouse-grown plants and observed them by light microscopy using a Zeiss (Jena, Germany) Axiolab microscope.

For confocal microscopy, roots from 10-day-old seedlings were sectioned by hand, placed on a cover slip, and observed using a Zeiss LSM410 confocal microscope. For excitation, we used the 488-nm argon laser with a DBSP 488/543 dichroic mirror (Zeiss). Wall autofluorescence was detected with a BP 515- to 565-nm emission filter (Zeiss), and chlorophyll autofluorescence was visualized simultaneously with an LP 570-nm emission filter (Zeiss).

For electron microscopic analysis of cotyledons, seedlings were grown in darkness for 7 days, and samples were taken in absolute darkness. Cotyledons were fixed for 1 hr with 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.2, rinsed in buffer, and postfixed overnight in 2% osmium tetroxide at 4°C. Subsequently, samples were dehydrated in a graded ethanol series, rinsed in propylene oxide, and embedded in Epon resin (Electron Microscopy Sciences, Fort Washington, PA). Samples were then polymerized at 65°C overnight. Thin sections were stained with uranyl acetate and lead citrate and examined on a transmission electron microscope (model 400; Phillips, The Netherlands). For analysis of root samples, roots from 10-day-old seedlings were prepared for electron microscopy as described previously (Chory et al., 1989).

Green Fluorescent Protein Fusion Constructs and Transient Transfection

The full-length TDET1 coding region was cloned as a translational fusion downstream of a green fluorescent protein (GFP) gene variant, mGFP4, which contains modified codons engineered to mutate the cryptic intron splicing sites present within the wild-type GFP sequence (Haseloff et al., 1997). The fusion gene was under control of the 35S cauliflower mosaic virus promoter and nopaline synthase terminator in the pBI121-derived vector pBIN35S-mGFP4 (Haseloff et al., 1997). The construct PHYA–GFP was generated in the same binary vector and contains the full-length rice PHYA sequence (Kay et al., 1989) in a translational fusion upstream of mGFP4. This fusion protein is not targeted to the nucleus (F. Fenzi and C. Bowler, unpublished results).

Transient transfection of heterotrophic Nicotiana tabacum BY-2 suspension-cultured cells (Nagata et al., 1992) was performed using the Biolistic PDS-1000/He gene gun (Du Pont) and standard protocols recommended by the manufacturer. Twenty-four hours after particle bombardment, cells were observed in an epifluorescence microscope (Zeiss Axiolab) by using a ×20 objective and filter sets HQ:GFP (for TDET1–GFP) and GFP (for PHYA–GFP) from Chroma Technology Corp. (Brattleboro, VT).

ACKNOWLEDGMENTS

This work is dedicated to the memory of Professor Gaetano Salvatore, who passed away 25 June 1997, too early to see the fruits of his foresight. We thank Dick Kendrick, Maarten Koornneef, and members of their laboratories, in particular Ageeth Van Tuinen and Janny Peters, for plant material and for providing unpublished information, Jim Giovannoni and Robert Fluhr for providing tomato genomic libraries, Steve Tanksley for RFLP probes, Jim Haseloff for GFP constructs, Dirk Inzé for BY-2 cell suspension cultures, Luigi Frusciante and Renato Faraone Mennella for assistance with greenhouse and field space, Gennarino Iamunno for help with electron microscopy, Chiara Zurzolo and Danny Chamovitz for critical reading of the manuscript, and TOPLAB (Martinsried, Germany) for conducting some of the DNA sequencing under contract. This work was supported in part by grants from Human Frontier Science Programme (No. RG0362/1995M), Ministero per le Politiche Agricole (No. PM356), and the Consiglio Nazionale delle Ricerche Target Project on Biotechnology (No. 97.01259.PF49) to C.B.

Footnotes

  • 1 Current address: Department of Cell Biology, University of Freiburg, Schaenzlestrasse 1, Freiburg, Germany.

  • 2 Current address: Dipartimento di Patologia Vegetale, Facoltà di Agraria, Università degli Studi di Napoli, Via Università 100, I-80055 Portici, Italy.

  • 3 Current address: CNR–Istituto per il Miglioramento Genetico delle Piante da Orto e da Fiore, Via Università 133, I-80055 Portici, Italy.

  • Received September 8, 1998.
  • Accepted November 19, 1998.
  • Published February 1, 1999.

REFERENCES