Physiological interactions of phytochromes A, B1 and B2 in the control of development in tomato

Authors

  • James L. Weller,

    1. Laboratory of Plant Physiology, Graduate School of Experimental Plant Sciences, Wageningen University, Arboretumlaan 4, NL6703 BD Wageningen, The Netherlands,
    2. Laboratory of Genetics, Graduate School of Experimental Plant Sciences, Wageningen University, Dreijenlaan 2, NL6703 HA Wageningen, The Netherlands, and Biology Department, University of Leicester, Leicester LE1 7RH, UK
    Search for more papers by this author
  • Mariëlle E.L. Schreuder,

    1. Laboratory of Plant Physiology, Graduate School of Experimental Plant Sciences, Wageningen University, Arboretumlaan 4, NL6703 BD Wageningen, The Netherlands,
    Search for more papers by this author
  • Harry Smith,

  • Maarten Koornneef,

    1. Laboratory of Genetics, Graduate School of Experimental Plant Sciences, Wageningen University, Dreijenlaan 2, NL6703 HA Wageningen, The Netherlands, and Biology Department, University of Leicester, Leicester LE1 7RH, UK
    Search for more papers by this author
  • Richard E. Kendrick

    Corresponding author
    1. Laboratory of Plant Physiology, Graduate School of Experimental Plant Sciences, Wageningen University, Arboretumlaan 4, NL6703 BD Wageningen, The Netherlands,
    Search for more papers by this author

*For correspondence (fax +31 317484740; e-mail Dick.Kendrick@algem.pf.wau.nl).

Summary

The role of phytochrome B2 (phyB2) in the control of photomorphogenesis in tomato (Solanum lycopersicum L.) has been investigated using recently isolated mutants carrying lesions in the PHYB2 gene. The physiological interactions of phytochrome A (phyA), phytochrome B1 (phyB1) and phyB2 have also been explored, using an isogenic series of all possible mutant combinations and several different phenotypic characteristics. The loss of phyB2 had a negligible effect on the development of white-light-grown wild-type or phyA-deficient plants, but substantially enhanced the elongated pale phenotype of the phyB1 mutant. This redundancy was also seen in the control of de-etiolation under continuous red light (R), where the loss of phyB2 had no detectable effect in the presence of phyB1. Under continuous R, phyA action was largely independent of phyB1 and phyB2 in terms of the control of hypocotyl elongation, but antagonized the effects of phyB1 in the control of anthocyanin synthesis, indicating that photoreceptors may interact differently to control different traits. Irradiance response curves for anthocyanin synthesis revealed that phyB1 and phyB2 together mediate all the detectable response to high-irradiance R, and, surprisingly, that the phyA-dependent low-irradiance component is also strongly reduced in the phyB1 phyB2 double mutant. This is not associated with a reduction in phyA protein content or responsiveness to continuous far-red light (FR), suggesting that phyB1 and phyB2 specifically influence phyA activity under low-irradiance R. Finally, the phyA phyB1 phyB2 triple mutant showed strong residual responsiveness to supplementary daytime FR, indicating that at least one of the two remaining phytochromes plays a significant role in tomato photomorphogenesis.

Introduction

Light has profound effects on the development of flowering plants (Kendrick and Kronenberg, 1994; Neff et al., 2000; Thomas and Vince-Prue, 1997) The most widespread and obvious influences of light are in the regulation of hypocotyl and stem elongation, cotyledon and leaf expansion, and chlorophyll synthesis. In many species, other processes such as seed germination, anthocyanin synthesis, and the induction of flowering are also responsive to light. Three known families of photoreceptor proteins interact to control these developmental responses. These are the phytochromes, which incorporate a bilin chromophore and absorb mainly red (R) and far-red (FR) light (Neff et al., 2000), and the cryptochrome and phototropin photoreceptor families which absorb blue light (BL) by virtue of flavin chromophores (Briggs and Huala, 1999; Cashmore et al., 1999).

Five phytochromes (phyA to phyE) are present in Arabidopsis, including two closely related phyB-type phytochromes, phyB and phyD (Clack et al., 1994; Sharrock and Quail, 1989). Tomato also contains five phytochromes, including homologues of Arabidopsis phyA, C and E, and two closely related phyB-type phytochromes, phyB1 and phyB2 (Hauser et al., 1995; Mathews and Sharrock, 1997; Pratt et al., 1995). However, tomato phyB1 and phyB2 are not orthologous to phyB and D in Arabidopsis, but represent an independent duplication within the Solanaceae (Pratt et al., 1995). In general, the genes encoding phytochrome types B–E are expressed at a relatively low level and are subject to relatively weak regulation by light (Clack et al., 1994; Hauser et al., 1998). By contrast, phytochrome A has acquired several distinct features, including a high level of PHYA gene expression in darkness, a strong down-regulation of gene expression by light, and light-dependent degradation of the phyA holoprotein (Clough and Vierstra, 1997; Quail, 1994).

The physiological roles of phyA and phyB have been characterized by transgenic and mutant analyses in a number of species (Kendrick et al., 1997; Weller et al., 1995; Weller et al., 1997; Whitelam and Devlin, 1997). Phytochrome B is considered to function in a relatively simple manner as a R/FR reversible switch, with a level of activation determined by the ratio of R to FR photons in the incident light. The action of phyB can be seen in seedling responses to continuous R (cR) or pulses of R, and in the response of older, white-light-grown plants to end-of-day or supplementary daytime FR (Smith, 1995). Phytochrome A, in contrast, acts in two quite different physiological modes (Shinomura et al., 1996; van Tuinen et al., 1995a; Whitelam et al., 1993). These are the very-low-fluence response (VLFR) mode, in which responses are saturated by extremely low fluences of light at wavelengths throughout the spectrum and are photo-irreversible (Mancinelli, 1994), and the FR high-irradiance response (FR-HIR) in which responses are induced by exposure to continuous high-irradiance FR in an irradiance-dependent manner (Mancinelli, 1994).

Less is known about the properties and functions of the remaining phytochromes, and they have generally been assumed to act in a mode similar to that of phyB. However, promoter fusion analyses (Goosey et al., 1997) and the recent identification of phyD and phyE mutants in Arabidopsis (Aukerman et al., 1997; Devlin et al., 1999) have shown that there are differences in their patterns of expression and physiological roles. In Arabidopsis seedlings, phyB strongly inhibits hypocotyl elongation and promotes cotyledon expansion under cR (Koornneef et al., 1980; Reed et al., 1994). Phytochrome D has a similar but less pronounced role and appears to act together with phyB in a largely redundant manner (Aukerman et al., 1997). Both phyB and phyD contribute to shade-avoidance responses, including the promotion of flowering and petiole elongation (Devlin et al., 1999; Halliday et al., 1994). In contrast, phyE has a negligible effect on hypocotyl elongation, but has a striking role in inhibiting the elongation of rosette internodes under WL (Devlin et al., 1998). Phytochrome C-deficient mutants have yet to be isolated, and the only evidence about the function of this photoreceptor comes from over-expression studies in Arabidopsis and tobacco, which indicate a role in cotyledon and leaf expansion (Halliday et al., 1997; Qin et al., 1997) and flowering (Halliday et al., 1997).

In tomato, phyB1 contributes strongly to de-etiolation under cR and thus has a similar role to Arabidopsis phyB (van Tuinen et al., 1995b). However, this role appears somewhat less dominant than that of phyB in Arabidopsis, since tomato phyB1 mutants still respond strongly to cR, end-of-day FR and supplementary daytime FR (van Tuinen et al., 1995b). Clear roles for the remaining phytochromes in tomato have yet to be distinguished, although differences in their spatial, temporal and diurnal patterns of expression suggest that they may have somewhat specialized functions (Hauser et al., 1997; Hauser et al., 1998). In addition, the differences in expression of the tomato PHY genes do not correspond clearly to those observed for the homologous genes in Arabidopsis (Clack et al., 1994), raising the possibility that additional differences may exist in the roles and relative importance of the phytochrome family members in these two species.

To identify mutants that might potentially carry lesions in other phytochrome genes, we screened M2 progeny of mutagenized phyA phyB1 seedlings for mutants showing impaired de-etiolation under white light (WL) (Kendrick et al., 1997). One of the allelic groups identified was recently shown to carry mutations in PHYB2 (Kerckhoffs et al., 1999). Here we use these phyB2 mutants to examine the role of phyB2 and interactions among phyA, phyB1 and phyB2 in control of seedling de-etiolation, shade avoidance and other responses in the mature plant.

Results

Isolation of monogenic phyB2 mutants

Two mutant PHYB2 alleles were identified previously in a phyA phyB1 background (Kerckhoffs et al., 1999). Triple mutant seedlings grown under WL were in general substantially more etiolated in appearance than the double mutant progenitor, having longer hypocotyls with reduced anthocyanin content, and paler cotyledons with retarded expansion (Kerckhoffs et al., 1999). The loss of phyB2 also had a strong effect in older plants, with the triple mutant having longer internodes, paler stems, leaves and fruit trusses. To analyse the role of phyB2 in more detail, we crossed the original triple mutant lines to the wild-type cv. Moneymaker and selected monogenic and double mutant combinations by a combination of phenotypic and PCR-based screening.

The phyB2 mutation was found to have very little effect on either a wild-type or a phyA background in plants grown under glasshouse conditions. However, a strong effect of phyB2 was revealed in the absence of phyB1 (Figure 1). Relative to wild-type or either single mutant, phyB1 phyB2 double mutants grown in the glasshouse had longer internodes and a lower chlorophyll content in the stem and leaves (Figure 1).

Figure 1.

PhyB2 acts redundantly with phyB1 to control development under white light.

The photograph shows 5-week-old wild-type cv MM, phyB1, phyB2 and phyB1 phyB2 plants grown from sowing under natural daylight in the glasshouse during summer.

Phytochromes A, B1 and B2 interact to control de-etiolation under continuous red light

Previous reports have shown that both phyA and phyB1 contribute to the inhibition of hypocotyl elongation by cR (van Tuinen et al., 1995a; van Tuinen et al., 1995b). Hypocotyl elongation in the phyA phyB1 double mutant has not yet been reported. However, plants of this genotype do not synthesize any detectable anthocyanin in response to 24 h irradiation with cR (Kerckhoffs et al., 1997). To examine the roles and interactions of phyA, phyB1 and phyB2 in the response to cR, we first grew the eight possible homozygous allelic combinations under cR (3 µmol m−2 sec−1) for 12 days after sowing, and measured hypocotyl length, cotyledon mass and the content of anthocyanin and chlorophyll in the hypocotyl. Figure 2 compares these plants to dark-grown wild-type control plants. Dark-grown control plants of the other genotypes did not differ significantly from wild-type for any of the three characteristics measured (data not shown).

Figure 2.

Interactions between phyA, phyB1 and phyB2 in the control of de-etiolation under red light.

Wild-type cv MM (WT) and mutant seedlings were grown from sowing under continuous broad-band red light of 3 µmol m−2 sec−1. All measurements were taken 12 days after sowing. Values are means ± SE for n = 20–30 (hypocotyl length) or n = 4–6 (all other characteristics). The experiment was repeated three times with qualitatively similar results.

The loss of phyB1 caused a substantial reduction in the inhibition of hypocotyl elongation, while the loss of phyA had a smaller, yet clearly significant effect (Figure 2). As in the case of WL-grown plants, loss of phyB2 had no measurable effect in a wild-type or phyA background but had a strong effect in a phyB1 background. Phytochromes A and B1 had a clearly additive effect on hypocotyl growth inhibition, but the phyA phyB1 double mutant still showed substantial inhibition relative to the dark control. Loss of all three phytochromes resulted in a complete loss of hypocotyl growth inhibition (Figure 2). It is clear that the relative degree of inhibition by phyA depends little on which other phytochromes are present. However, the partial redundancy of phyB1 and phyB2 is clearly shown by the fact that the relative effect of each of these phytochromes is much greater in the absence of the other. The results for cotyledon mass and hypocotyl chlorophyll content were very similar to those for hypocotyl elongation (Figure 2). phyB1 also had the strongest effect for the stimulation of anthocyanin production. When phyB1 was present, the loss of phyB2 had no effect, whereas the loss of phyA resulted in a marked increase in anthocyanin content. This negative effect of phyA reflects a specific interference with a phyB1 response, because in the absence of phyB1, phyA contributed positively to anthocyanin accumulation (Figure 2). The apical hook opened in all mutant combinations with the exception of the phyA phyB1 phyB2 triple mutant, which retained an apical hook similar to wild-type seedlings grown in complete darkness (data not shown).

The phyA-mediated response to very low irradiances of continuous red light is dependent on the presence of phyB1 or phyB2

The irradiance response curve for the accumulation of anthocyanin in response to 24 h irradiation with cR is clearly biphasic in wild-type tomato cv Moneymaker, with low- and high-irradiance components (Figure 3). It has previously been shown that these two components are dependent on the presence of phyA and phyB1, respectively (Kerckhoffs et al., 1997). Loss of phyB2 alone had no significant effect on anthocyanin synthesis throughout the irradiance range tested. Loss of phyB1 alone selectively impaired the high-irradiance component (Figure 3; Kerckhoffs et al., 1997). The phyB1 phyB2 double mutant was even less responsive than the phyB1 single mutant in the high-irradiance range and was also strongly impaired in the low-irradiance component (Figure 3). Loss of the low-irradiance response was also seen in another phyB1 phyB2 double mutant containing a different phyB2 allele (phyB2-2) (data not shown). One reason for the loss of this response could be that level of phyA was reduced in the phyB1 phyB2 double mutant. However, this is not the case, because wild-type and phyB1 phyB2 seedlings exposed to 0.02 µmol m−2 sec−1 cR for 24 h did not differ substantially in their level of immunoblot-detectable PHYA (data not shown). It therefore appears that this phyA response is dependent on the presence of a phyB-type phytochrome, even though these phytochromes are not capable of inducing a response to R in this irradiance range.

Figure 3.

Irradiance response relationship for induction of anthocyanin synthesis by red light in the phyB1, phyB2 and phyB1 phyB2 double mutants.

Wild-type cv MM (WT) and mutant seedlings were grown in complete darkness for 4 days, and then given a 24 h irradiation with narrow-band red light (660 nm), or maintained in darkness for a further 24 h before harvest. Values are means ± SE for three independent experiments.

To characterize further the interaction of phyA and the phyB-type phytochromes, we also examined anthocyanin accumulation under longer wavelengths. The results for 657 nm in Figure 4 confirm the previous report (Kerckhoffs et al., 1997) that phyA is essential for expression of a response to low irradiance at this wavelength. Despite a somewhat higher overall level of anthocyanin due to the use of a different narrow-band filter, the results in Figure 4 are consistent with those in Figure 3, showing that this phyA-mediated response is largely, but not entirely, dependent on the presence of either phyB1 or phyB2. Phytochromes B1 and B2 together also mediate a high-irradiance component that is independent of phyA. The phyA-mediated, phyB-dependent low-irradiance response is still seen under 687 nm light, where both phyA and the phyB-type phytochromes contribute to the response in the high-irradiance range. Under 708 nm and 729 nm light, the low-irradiance component is no longer apparent, the phyB-type phytochromes are not active, and the response is essentially a phyA-mediated HIR. We conclude that the phyB1 phyB2 double mutant is not impaired in response to FR and therefore that phyB1 or phyB2 are only required for the action of phyA at low irradiances of cR, and not for the FR-HIR.

Figure 4.

The phyB1 phyB2 double mutant shows normal anthocyanin synthesis in response to far-red light.

Seedlings were grown in complete darkness for 4 days and then given a 24 h irradiation with narrow-band light of various wavelengths, or maintained in darkness for a further 24 h before harvest and anthocyanin quantification. Values are means ± SE for three independent experiments.

The hp1 mutation is expressed in the absence of phyA, phyB1 and phyB2

The hp1 mutation confers a substantial enhancement of de-etiolation under both R and WL (Kerckhoffs et al., 1997; Peters et al., 1992). However, the mutation has no detectable morphological effects in dark-grown seedlings, suggesting that the action of the HP gene depends on activation of phytochrome. The hp1 mutation clearly amplifies both phyA- and phyB1-mediated responses (Kerckhoffs et al., 1997; Peters et al., 1992). We investigated whether hp1 could also amplify residual responses in the phyA phyB1 and phyA phyB1 phyB2 backgrounds. In WL-grown seedlings, the effects of the hp1 mutation are clearly expressed even in the absence of phyA, phyB1 and phyB2 (Figure 5a). At the seedling stage, the hp1 mutation partially overcame many aspects of the phyA phyB1 phyB2 phenotype, causing a reduction in length and an increase in anthocyanin content of the hypocotyl. The hp1 mutation was also expressed in older plants, but did not affect all aspects of the phyA phyB1 phyB2 phenotype to an equivalent extent (Figure 5b). The loss of phyB2 in a phyA phyB1 background resulted in a striking reduction in fruit chlorophyll content and a marked increase in truss length as a result of increased distance between fruits on the inflorescence axis. Introduction of the hp1 mutation had a strong effect on fruit chlorophyll content in both phyA phyB1 and phyA phyB1 phyB2 backgrounds, conferring a sixfold increase (Figure 5b, data not shown). In contrast, hp1 had little effect on truss length (Figure 5b).

Figure 5.

The hp1 mutation is strongly expressed in a phyA phyB1 phyB2 background under white light.

(a) Two-week-old seedlings. (b) Third fruit truss. Plants were grown from sowing under natural daylight in the glasshouse during summer.

Although hp1 clearly enhances de-etiolation under both R and blue light (BL) (Peters et al., 1992), it is not clear whether the effects in BL are due to amplification of the phytochrome signal alone or also to amplification of a signal from a BL receptor or receptors. The enhanced anthocyanin synthesis and inhibition of hypocotyl elongation seen in the phyA phyB1 phyB2 hp1 quadruple mutant under WL (Figure 5a) could thus reflect amplification of the response to one of the two remaining phytochromes, or to a BL receptor such as cryptochrome 1, which is clearly active in WL-grown tomato seedlings (Ninu et al., 1999; Weller et al., unpublished results). It was therefore of interest to examine whether any phytochrome-controlled responses were still present in the phyA phyB1 phyB2 or phyA phyB1 phyB2 hp1 mutants.

Residual phytochrome responses in the phyA phyB1 phyB2 triple mutant

The results in Figure 2 superficially suggest no role for phytochromes other than phyA, phyB1 and phyB2 in the control of seedling de-etiolation under cR. However, the irradiance used in these experiments is unlikely to be saturating for phytochrome action (Kerckhoffs et al., 1997), and it is possible that the action of an additional phytochrome might only be seen at higher irradiances. The results in Figure 6(a) show that, even under a 10-fold higher R irradiance, the phyA phyB1 phyB2 mutant showed no significant response for any of the characteristics measured. On a phyA phyB1 background, the hp1 mutation caused a dramatic enhancement of response for all characteristics, implying that hp1 can enhance responsiveness to phyB2. By contrast, on a phyA phyB1 phyB2 background, hp1 had no significant effect on hypocotyl elongation or anthocyanin synthesis. However, a significant effect of hp1 on hypocotyl chlorophyll content and cotyledon mass was seen (Figure 6a).

Figure 6.

Residual red-light responsiveness in the phyA phyB1 phyB2 triple mutant.

(a) Seedlings were grown from sowing in complete darkness or under continuous red light (R) (30 µmol m−2 sec−1). (b) Seedlings were grown from sowing under continuous blue light (BL) (15 µmol m−2 sec−1), either with or without continuous supplementary R (15 µmol m−2 sec−1). All measurements were taken 14 days after sowing. Values are means ± SE for n = 20–30 (hypocotyl length) or n = 4–6 (all other characteristics). The experiments were repeated with qualitatively similar results.

The absence of residual response to cR in the phyA phyB1 phyB2 triple mutant might simply reflect the possibility that phytochromes E and F do not play a substantial role in tomato photomorphogenesis. Alternatively, a more complex situation could be envisaged in which the function of these phytochromes might be dependent on phyA, phyB1, phyB2 or BL receptor activation. We therefore examined whether the phyA phyB1 phyB2 triple mutant or the phyA phyB1 phyB2 hp1 quadruple mutant showed any residual phytochrome responses under conditions where BL receptors would also be activated. To do this, we grew these genotypes together with their PHYB2 isolines under continuous BL and examined the effect of adding R. The results in Figure 6(b) show clearly that the addition of R to a background of BL caused a small but significant inhibition of elongation and an increase in anthocyanin synthesis in the hypocotyl of phyA phyB1 phyB2 triple mutant seedlings. These responses were considerably greater in the phyA phyB1 phyB2 hp1 quadruple mutant. This result suggests that at least one of the two remaining phytochromes is capable of contributing to the control of de-etiolation under R, although the action of other R-absorbing pigments cannot be excluded. It also suggests that BL photoreceptors must be activated for this contribution to be detected, and confirms that hp1 can amplify this additional R response.

White-light-grown phyA phyB1 phyB2 triple mutant plants retain a strong shade avoidance response to supplementary far-red light

Another approach to identifying potential roles for residual phytochromes in tomato would be to examine whether WL-grown seedlings of the phyA phyB1 phyB2 mutant show any phytochrome responsiveness. Such an approach has two additional advantages. Firstly, the activation of residual phytochromes and any BL receptors would likely be much higher under high-irradiance WL than under the conditions used in the R and BL + R experiments, and this would give a better chance of detecting any residual response. Secondly, seedlings could be grown beyond the hypocotyl stage, increasing the chance of detecting phytochromes with action restricted to later stages of development. A well-recognized way to examine phytochrome responses in de-etiolated plants is to manipulate the R:FR ratio by adding FR to a background of WL containing a high proportion of R (Smith, 1982). Figure 7 shows that WL-grown wild-type tomato plants respond to supplementary FR both by increased hypocotyl elongation and by reduction of anthocyanin accumulation. Triple mutant plants lacking phyA, phyB1 and phyB2 also responded strongly for all characters. The residual response was particularly striking in the first internode of triple mutant plants, which showed a two- to threefold increase in length (Figure 7). These results clearly demonstrate a substantial role for phyE and/or phyF in the control of hypocotyl and stem elongation and of anthocyanin synthesis in the hypocotyl.

Figure 7.

Shade-avoidance responses in phytochrome-deficient tomato mutants.

Wild-type and phytochrome-deficient mutant plants were grown under an 18 h photoperiod given as white light of approximately 150 µmol m−2 sec−1 either with (WL + FR; R:FR = 0.08) or without (WL; R:FR = 6.28) supplementary far-red irradiation for the duration of each daily photoperiod. (a) Representative 24-day-old wild-type (WT) and phyA phyB1 phyB2 triple mutant plants grown for 10 days from sowing under WL before transfer to differential R:FR conditions for 14 days. (b) Hypocotyl length and anthocyanin content in 2-week-old seedlings grown from sowing under differential R:FR conditions. Values are means ± SE for n = 20–30 (hypocotyl length) or n = 4–6 (anthocyanin content). The length of the first internode was measured in seedlings grown for 10 days from sowing under WL before transfer to differential R:FR conditions for 14 days. Values are means ± SE for n = 7.

Discussion

The full complement of expressed PHY genes has only been described for two higher plant species; Arabidopsis and tomato (Clack et al., 1994; Hauser et al., 1995; Sharrock and Quail, 1989). Although phylogenetic analyses show orthology between phyA, phyE and phyC/F gene pairs in these two species, it is clear that the phyB lineage has undergone independent duplication (Pratt et al., 1995), raising questions about the extent of functional differentiation within the phyB family in these species. Mutants for both phyB-type phytochromes have now been identified in Arabidopsis (Aukerman et al., 1997; Somers et al., 1991) and tomato (Kerckhoffs et al., 1999; van Tuinen et al., 1995b), which has provided the first opportunity for exploring this question. In addition, there is growing interest in the interactions between photomorphogenic photoreceptors (Casal, 2000), and the isolation of phyB2 mutants in tomato has made it possible to investigate interactions among the phytochromes in this species, and to compare these with interactions identified in Arabidopsis.

Roles of phyA, phyB1 and phyB2 in de-etiolation under red light

Roles for phyA and phyB1 in the mediation of de-etiolation responses to R in tomato have been demonstrated previously (van Tuinen et al., 1995a; van Tuinen et al., 1995b). In the control of anthocyanin biosynthesis under R, phyA acts predominantly at low irradiances, and phyB1 at higher irradiances (Kerckhoffs et al., 1997). Distinct low- and high-irradiance components of the response to cR have also been identified in Arabidopsis, and, as in tomato, these are associated with phyA and phyB action, respectively (Mazzella et al., 1997). Although the activity of phyA under low-irradiance cR has been considered to be a VLFR (Mazzella et al., 1997), it saturates at a phytochrome photo-equilibrium two orders of magnitude higher than classic VLFRs (Mancinelli, 1994). In contrast, the phyA-dependent induction of seed germination by R in Arabidopsis has the characteristics of a classic VLFR (Botto et al., 1996; Shinomura et al., 1996). The exact nature of the relationship between these two modes of phyA-mediated response to R is not yet clear.

Although the phyA phyB1 double mutant is effectively blind to low-irradiance R, it de-etiolates normally under WL. The phenotype of phyA phyB1 phyB2 mutants under natural daylight indicated an important role for phyB2 in this residual response (Kerckhoffs et al., 1999), and it is clear from the present study that phyB2 is also active in R-sensing. However, the strongly synergistic effects of phyB1 and phyB2 mutations indicate a high degree of functional redundancy between these phytochromes, as might be expected given their relatively recent divergence (Pratt et al., 1995). In seedling de-etiolation, effects of phyB2 were only seen in the absence of phyB1, whereas phyB1 still retained substantial function in the absence of phyB2. Phytochrome B1 can therefore compensate fully for the loss of phyB2, but the reverse is not so. This is similar to the interaction reported for phyB and phyD in the control of de-etiolation in Arabidopsis. Although the phyD mutation alone had limited effects on cotyledon expansion and anthocyanin accumulation in the Wassilewskija (WS) background, the effects of the phyB mutant were in general much stronger, and, in the Landsberg erecta (Ler) background, clear effects of phyD were only seen in the absence of phyB (Aukerman et al., 1997). Thus, although the phyB families in tomato and Arabidopsis have undergone independent duplications, one family member in each species is clearly more dominant. In tomato, it appears that the extent of this dominance can vary according to the light conditions. For example, when seedlings are de-etiolated under R pulses, the loss of phyB1 results in a complete insensitivity during the first few days of exposure (van Tuinen et al., 1995b), suggesting that phyB1 is initially the only phytochrome capable of sustaining a significant response under this regime. The gradual recovery of responsiveness in the phyB1 mutant shows the increasing contribution of other phytochromes, although it is not clear whether this reflects an increase in photoreceptor content or level of activity as the period of light treatment is extended.

Redundancy of phyB-type phytochromes

The absence of any striking functional differentiation between the two phyB-type phytochromes in tomato raises the question of why such similar photoreceptors have been maintained. One possibility may be that they have a cumulative function under some conditions, as seen for Arabidopsis phyB and phyD in the WS genetic background (Aukerman et al., 1997). Although we did not detect any effect of phyB2 in a PHYB1 background, both phyB1 and phyB2, and in fact most of the other photoreceptors studied to date, show a degree of haplo-insufficiency under some conditions or certain genetic backgrounds (Koornneef et al., 1980; van Tuinen et al., 1995b; J.L. Weller, unpublished data). This means that light responses during de-etiolation may be limited by total photoreceptor quantity. The selective advantage in duplication of PHY genes may thus have derived initially from an increased sensitivity to the low levels of light immediately below the soil surface, which would allow the germinating seedling to better anticipate emergence and the need for transition to photo-autotrophic growth.

However, it is also possible that phyB1 and phyB2 may have acquired partially distinct functions. An early step in functional divergence is likely to be the acquisition of unique expression patterns (Pickett and Meeks-Wagner, 1995), and an indication of any discrete roles for phyB1 and phyB2 might therefore be gained from their patterns of expression. The spatial patterns of expression of phyB-type phytochromes show substantial overlap (Goosey et al., 1997; Hauser et al., 1997), suggesting that they continue to be active in the same cells and tissues. In Arabidopsis, phyD expression is restricted to a subset of those tissues expressing phyB, mainly leaf tissue (Goosey et al., 1997). In tomato, both PHYB1 and PHYB2 are expressed at a similar level in most plant parts, but the expression of PHYB2 is substantially elevated relative to PHYB1 in fruits (Hauser et al., 1997). A clear difference is also seen in their diurnal rhythms of expression, which are out of phase by approximately 10 h (Hauser et al., 1998). These two divergent features of phyB1 and phyB2 could reflect a divergence of function that is not apparent in the present study. In Arabidopsis, it has recently been shown that phyB is intimately associated with circadian rhythmicity. Loss of phyB alters period length (Somers et al., 1998), and phyB is itself a target of circadian regulation (Kozma-Bognár et al., 1999). It would also be interesting to determine whether phyB1 and phyB2 mutations affect rhythmicity in tomato.

Phytochrome interactions in de-etiolation under red light

In contrast to the clear redundancy of phyB1 and phyB2, the contribution of phyA to most aspects of tomato seedling de-etiolation is unaffected by the presence or absence of phyB1 or phyB2. However, in the control of anthocyanin synthesis, phyA interacts with the other phytochromes in three distinct ways. In seedlings grown under long-term cR, there is a strong negative effect of phyA on phyB1-mediated anthocyanin accumulation. A similar antagonistic effect of phyA on a phyB-mediated response has also been reported for control of hypocotyl elongation in Arabidopsis (Mazzella et al., 1997). A contrasting, positive interaction of phyA with phyB1 is seen in the enhancement of phyB1-mediated anthocyanin synthesis by pre-treatments with cFR (J.L. Weller, unpublished data). A positive interaction between phyA acting in the FR-HIR mode and phyB is also seen in Arabidopsis (Hennig et al., 1999b).

The third interaction involving phyA seen in tomato is the dependence of phyA action under low-irradiance R on the presence of either phyB1 or phyB2. The response to low-irradiance R cannot be mediated by phyB1 or phyB2 alone, because even when amplified sixfold in the hp1 genetic background it is still completely eliminated by the loss of phyA (Kerckhoffs et al., 1997). It thus appears that phyB1 or phyB2 are required for the expression of a response in an irradiance range where neither appears to be active in its own right. Our data do not exclude the possibility that the reduction in response of the phyB1 phyB2 mutant at 0.01 µmol m−2 sec−1 is due to an increased sensitivity to phyA (causing a shift towards lower irradiances), rather than a reduction in phyA activity. The former explanation would imply that both phyB1 and phyB2 could negatively regulate phyA signalling, whereas the latter would require phyA to depend positively on phyB1/phyB2. In either case, it is clear that an interaction between phyA and phyB1/phyB2 is essential for normal anthocyanin biosynthesis under low-irradiance cR, and it would be of interest to examine whether this is also true for other aspects of tomato seedling de-etiolation. This type of interaction between phyA and phyB-type phytochromes has so far not been identified in other species. In both Arabidopsis and pea, the phyA-dependent inhibition of hypocotyl or epicotyl elongation under low-irradiance cR or R pulses is not affected by the loss of phyB (Mazzella et al., 1997; Weller et al., 1995; Weller et al., 1997). However, it remains to be seen whether in either species this phyA response is affected by the additional loss of other phyB-type phytochromes.

The action of tomato phyA under low-irradiance R is thus different in nature from its action in the FR-HIR, not only in terms of wavelength- and irradiance-dependence, but also in terms of dependence on a phyB-type phytochrome for expression. The results presented here in effect provide a genetic separation of these two different modes of phyA action. A genetic distinction between two modes of phyA action has previously been demonstrated in a different manner by Yanovsky et al. (1997). These authors observed that the Arabidopsis ecotype Columbia (Col) possessed a strongly attenuated response to low fluences of R or FR in terms of inhibition of hypocotyl elongation relative to the Ler ecotype, and they identified two quantitative trait loci contributing to this difference. Interestingly, Col retained a normal phyB-mediated response compared to Ler (Yanovsky et al., 1997), suggesting that the low-irradiance R response in Col is impaired downstream of the point at which it enhances phyB signalling. Several phyA signalling mutants have also been identified in Arabidopsis. The fhy1 mutants impair phyA responses to both low- and high-irradiance FR (Cerdán et al., 1999), whereas eid1 mutants enhance phyA action in both these modes (Büche et al., 2000). The same appears to be the case for the spa1 mutants, which show a phyA-dependent enhancement of responses to both cR and cFR (Hoecker et al., 1998). By contrast, mutations in the FAR1 gene have no consistent effect on R responses (Hudson et al., 1999), suggesting that they may affect the FR-HIR specifically.

Roles for phyE and phyF in tomato

Although the phyA phyB1 phyB2 mutant is indistinguishable from dark-grown wild-type tomato seedlings grown under cR, loss of HP1 function or the activation of BL receptors reveals a residual R response for all characteristics. It therefore appears that at least one other phytochrome (phyE or phyF) is active in controlling de-etiolation in tomato, but is functionally dependent on cryptochrome activity, at least in the absence of phyA, phyB1 and phyB2. This may be similar in nature to the conditional dependence of Arabidopsis phyB and phyD on cryptochrome 1 (Casal and Mazzella, 1998; Hennig et al., 1999a) and of tomato phyB2 on cryptochrome 1 (J.L. Weller, unpublished data). The strong shade-avoidance response of the phyA phyB1 phyB2 triple mutant is comparable to that shown by the phyA phyB phyD mutant of Arabidopsis (Devlin et al., 1999), and also identifies an important role for phyE and/or phyF in older plants. Expression of tomato PHYE is highest in the hypocotyl and stem (Hauser et al., 1997), and in contrast to the other tomato PHY, is increased by light and continues to increase for 2–3 weeks following imbibition (Hauser et al., 1998). This is similar to the expression pattern of Arabidopsis PHYE which overlaps spatially with that of PHYB and PHYD, but is increased rather than reduced by light (Goosey et al., 1997). PHYE therefore appears to have acquired (or retained) patterns of expression that are distinct from those of PHYB-type PHY genes, and this suggests a more important role for this phytochrome in older, de-etiolated plants than in seedlings (Hauser et al., 1998). Consistent with these observations, the loss of phyE in Arabidopsis has no detectable effect at the hypocotyl stage, but confers greatly elongated rosette internodes on a phyA phyB1 background (Devlin et al., 1998). Phytochrome E clearly contributes to R:FR-sensing and thus acts in a similar mode to phyB and phyD, but at a different developmental stage. With respect to redundancy, phyE therefore appears to be an intermediate case, as might be expected given its earlier divergence from the phyB lineage (Clack et al., 1994; Hauser et al., 1995). It seems probable that phyE mutants of tomato would be readily identified in a phyA phyB1 phyB2 background, on the basis of an elongated phenotype in daylight.

Experimental procedures

Plant material

We have adopted a revised nomenclature for several of the mutants used in this study. Mutant phyA-1 corresponds to the previously described fri1 mutant isolated in the tomato (Solanum lycopersicum L.) wild-type cv. Moneymaker (MM) background, which behaves as a physiologically null mutant (Lazarova et al., 1998a; van Tuinen et al., 1995a). The phyB1-1 null mutant corresponds to the previously described tri1 mutant isolated in the wild-type cv. GT background (Lazarova et al., 1998b; van Tuinen et al., 1995b). Finally, the hp1-2 allele corresponds to the hp-1w allele described by Peters et al. (1989).

The phyB1-1 material used in this study was obtained by back-crossing the original mutant (in the GT background, van Tuinen et al., 1995b) once into the wild-type cv. MM background. Selection of the phyA-1 phyB1-1 double mutant has been described previously (Kerckhoffs et al., 1997). The phyB2-1 (70F) and phyB2-2 (55H) mutant alleles were identified in WL screens of γ-ray-mutagenized phyA-1 phyB1-1 double mutant seedlings and are both predicted to be null for phyB2 (Kendrick et al., 1997; Kerckhoffs et al., 1999). The hp1-2 lines used in this study were derived from crosses of the hp1-2 phyB1-1 double mutant (Kerckhoffs et al., 1997) with the phyA-1 phyB1-1 double mutant or the phyA-1 phyB1-1 phyB2-1 triple mutant.

PCR markers for identification of mutant alleles

Derived cleaved amplified polymorphic sequence markers (dCAPS; Michaels and Amasino, 1998; Neff et al., 1998) were designed for the phyA-1 and phyB1-1 mutations (Lazarova et al., 1998a; Lazarova et al., 1998b) using the dCAPS Finder program kindly supplied by Michael Neff. The three-residue phyB2-1 lesion (Kerckhoffs et al., 1999) created a CAPS marker by removal of a Fok1 site, but we found it simpler to detect it by PCR alone, using two primers corresponding to the wild-type forward strand sequence and the phyB2-1 reverse strand sequence together with a non-specific primer on each strand. The phyB2-2 mutation introduced a novel Mnl1 site and was detected as a simple CAPS marker. Details of other primers are as follows, listed 5′ to 3′ with mismatched residues in lower case: phyA-1, forward TAACTGAATACACCATTCCcTTAACC, reverse ATAATCGCTCTATAGTCACC, mutant product cut by EcoN1; phyB1-1, forward CTAAAATTCAAAGAGGAGGTCAgATT, reverse GAAGGGGTAAAAAGGGTCCTAA, wild-type product cut by Hinf1; phyB2-1 (70F), forward non-specific CCCTTTTTCCTTTTCTGACC, forward wild-type-specific GACAATATTGAGGATGGGTA, reverse mutant-specific GTCTTGATTTCGTCTGGA, reverse non-specific GTCTTGATTTCGTCTGGA; phyB2-2 (55H), forward TTATTGAGGGAAAGAGGAGT, reverse GATATAGCTGGGAAAAGTGA.

Growth conditions and light sources

Seeds were sown directly in trays of peat-based compost, or on moistened filter paper in plastic boxes (Peters et al., 1992). Monochromatic light sources and growth cabinets have been described previously (van Tuinen et al., 1995a), as has the threshold-box apparatus used for the irradiance response experiments (Peters et al., 1992). Narrow band irradiation was obtained using interference filters (Figure 3, peak transmission 660 nm, Schott, Mainz, Germany; Figure 4, peak transmissions 657, 687, 708 and 728 nm, Baird-Atomic, Bedford, MA, USA). For general-purpose genetic experiments, plants were grown in a heated glasshouse under natural light and day-length conditions. The supplementary FR experiments were conducted in custom-built growth cabinets (Fisons, Loughborough, UK) which have been described previously (McCormac et al., 1991). In these experiments, plants were maintained at 25°C under an 18 h light/6 h dark cycle of white light (photo-synthetically active radiation (PAR) 150 µmol m−2 sec−1), either with or without supplementary FR, giving R:FR ratios of 0.08 and 6.28, respectively.

Seedling measurements and pigment assays

Hypocotyl lengths were measured from soil surface to apical bud using a ruler. Cotyledon mass, anthocyanin and hypocotyl chlorophyll measurements were taken from pools of five seedlings (or 10 in irradiance response experiments). Anthocyanin was extracted from seedling hypocotyls by incubating overnight in acidified methanol. Extracts were diluted with a 3/4 volume of water and partitioned against a 7/6 volume of chloroform. After centrifugation (2000 g, 30 min) anthocyanin content was measured as the absorbance of the aqueous phase at 535 nm normalized to an original extracting volume of 1.2 ml. Chlorophyll was extracted by incubating plant material in DMSO overnight at 60°C. Chlorophyll content was determined from the absorbance at 649 and 665 nm according to Hiscox and Israelstam (1979).

Acknowledgements

We thank Tanja Borst-Vrenssen for assistance with irradiance response experiments, Chris Kendrick for assistance with seedling measurements and seed administration, Marie-Michèle Cordonnier-Pratt for initial help with primer design, Michael Neff for supplying the dCAPS Finder program, Jan Laurens and other staff of Unifarm for general care of the plants and for help with sowing and harvesting of seed, and Matthew Terry for his comments on the manuscript. We also thank members of the Department of Biology at Leicester University for help with the shade-avoidance experiment, and in particular Graeme Benskin at the Botanic Gardens facility for his meticulous care of the plants. This work was supported by grants to R.E.K., M.K. and H.A.S. under the EC Framework 4 project CT-972124 ‘PHOTARCH’.

Ancillary