Several novel allelic groups of tomato (Solanum lycopersicum L.) mutants with impaired photomorphogenesis have been identified after γ-ray mutagenesis of phyA phyB1 double-mutant seed. Recessive mutants in one allelic group are characterized by retarded hook opening, increased hypocotyl elongation and reduced hypocotyl chlorophyll content under white light (WL). These mutants showed a specific impairment in response to blue light (BL) resulting from lesions in the gene encoding the BL receptor cryptochrome 1 (cry1). Phytochrome A and cry1 are identified as the major photoreceptors mediating BL-induced de-etiolation in tomato, and act under low and high irradiances, respectively. Phytochromes B1 and B2 also contribute to BL sensing, and the relative contribution of each of these four photoreceptors differs according to the light conditions and the specific process examined. Development of the phyA phyB1 phyB2 cry1 quadruple mutant under WL is severely impaired, and seedlings die before flowering. The quadruple mutant is essentially blind to BL, but experiments employing simultaneous irradiation with BL and red light suggest that an additional non-phytochrome photoreceptor may be active under short daily BL exposures. In addition to effects on de-etiolation, cry1 is active in older, WL-grown plants, and influences stem elongation, apical dominance, and the chlorophyll content of leaves and fruit. These results provide the first mutant-based characterization of cry1 in a plant species other than Arabidopsis.
Most terrestrial plants remain in one place for their entire life. Reproductive success therefore depends heavily on the ability of the plant to modify its development in response to changes in its environment. This has resulted in the evolution of mechanisms for sensing and responding to environmental variables such as temperature, light and water availability. The nature of the light reaching a plant varies with degree of shading by soil or foliage, time of day, and time of year, and in response to this variation plants are able to regulate germination, de-etiolation, shoot architecture and flowering (Smith, 1994). In general, the regions of the spectrum with the strongest effects on development are the blue (BL), red (R) and far-red (FR) wavebands. Molecular genetic analyses of light sensing have shown that the R and FR effects are exclusively mediated by the well known phytochrome family of biliprotein photoreceptors (Neff et al., 2000). Both phyA and phyB also absorb BL (e.g. Abe et al., 1989), and it is thought that the remaining members of the phytochrome family will be similar in this respect. However, of the five known phytochromes in Arabidopsis, only two (phyA and phyB) have so far been shown to contribute to BL sensing (Neff and Chory, 1998; Poppe et al., 1998).
Two other types of photoreceptor, in addition to the phytochromes, are now known to regulate BL responses: phototropin, which has a specific role in the perception of unilateral light (Briggs and Huala, 1999); and the cryptochromes. The latter are photolyase-related flavoproteins which are active in the control of de-etiolation under BL, green and UV-A light (Ahmad and Cashmore, 1993; Koornneef et al., 1980; Lin et al., 1995, Lin et al., 1998). There are two cryptochromes in Arabidopsis (Ahmad and Cashmore, 1993; Lin et al., 1995), and therefore at least four photoreceptors which contribute to BL regulation of de-etiolation in this species. The role and potential interaction of these photoreceptors has been a subject of several recent studies (see Casal, 2000). While the results do provide clear evidence for photoreceptor interactions under some conditions, the conclusions are complex, on some points contradictory, and have yet to be tested in another plant species.
In an attempt to identify additional photoreceptor mutants in tomato, we conducted mutagenesis on a phyA phyB1 double mutant background. Screening under white light identified several allelic groups of mutants differently impaired in various aspects of de-etiolation (Kendrick et al., 1997). Mutants in one group were recently shown to carry lesions in the PHYB2 gene (Kerckhoffs et al., 1999), and it has been shown that phyA, phyB1 and phyB2 together account for all the detectable morphological de-etiolation responses to R (Weller et al., 2000). Here we present the characterization of a second allelic group. We show that mutants in this group have a specific defect in BL perception that results from mutation of the gene encoding the BL receptor cryptochrome 1 (cry1). We have used these mutants to explore the role of cry1 throughout development and its interactions with phyA, phyB1 and phyB2 in the control of seedling de-etiolation under BL.
Isolation of allelic mutants showing reduced responsiveness to blue light
Tomato mutants deficient in both phyA and phyB1 retain a relatively normal phenotype when grown in white light (WL), despite a complete inability to respond to FR and a greatly reduced response to R (Kendrick et al., 1997; Kerckhoffs et al., 1997). To identify additional genes involved in light signalling, we screened M2 progeny from γ-ray-mutagenized phyA phyB1 double-mutant seed under glasshouse conditions for phenotypic characteristics suggestive of impaired photomorphogenesis, such as elongated hypocotyls, elongated internodes and reduced content of anthocyanin or chlorophyll. The mutants identified in these screens comprised at least four allelic groups (Kendrick et al., 1997), one of which consists of mutants carrying lesions in the PHYB2 gene (Kerckhoffs et al., 1999; Weller et al., 2000). Another allelic group comprising two recessive mutants (lines 80B and 68G) showed seedling phenotypes generally similar to that of the phyA phyB1 phyB2 triple mutant, with elongated hypocotyls containing less anthocyanin and chlorophyll, and smaller cotyledons (Figure 1a,b). However, these mutants could be distinguished from the phyA phyB1 phyB2 triple mutant by a delay in opening of the apical hook. In addition, the combination of a more severe reduction in chlorophyll content and a much less severe reduction in anthocyanin content (Figure 1b) gave the hypocotyls of 80B a pinkish appearance in contrast to the green colour of phyA phyB1 phyB2 triple-mutant hypocotyls. In general, we could observe no substantial difference in severity of the two mutant alleles, and subsequent data are therefore presented for the 80B allele only.
The phenotype of the 80B mutant in WL was strongly suggestive of a defect in photomorphogenesis. To test whether 80B showed impaired response to a specific waveband, we grew mutant seedlings under continuous irradiation with light of different wavelengths, together with the phyA phyB1 progenitor and the phyA phyB1 phyB2 triple mutant for comparison. Figure 2 shows that phyA phyB1 seedlings retain substantial responsiveness to both R and BL for inhibition of hypocotyl elongation and for cotyledon expansion. The loss of phyB2 in a phyA phyB1 background resulted in the loss of virtually all the residual response to R, but only a small reduction in the response to BL. In contrast, the 80B mutant did not differ from its phyA phyB1 progenitor in response to R, but showed a substantial reduction in response to BL (Figure 2). The 80B phenotype therefore appeared to result from a specific defect in BL perception or signalling.
In parallel with mutants from the phyA phyB1 mutagenesis, we were also working with several putative photomorphogenic mutants isolated previously in a wild-type (WT) background. Mutant Z3-78 was isolated as a slightly elongated mutant in glasshouse-grown M2 progeny of ethyl methanesulfonate (EMS) -mutagenized WT cv. Ailsa Craig plants. Apart from a slightly elongated hypocotyl, the monogenic Z3-78 mutant showed no major alterations in photomorphogenesis. However, when the Z3-78 mutation was combined with phyB1 or aurea, a mutation impairing synthesis of the phytochrome chromophore (Terry and Kendrick, 1996), it greatly enhanced the slightly elongated phenotype conferred by these mutations (Kendrick et al., 1997). We interpreted this as indicating a possible redundancy between the Z3-78 gene product and the phytochrome signalling pathway, and crossed Z3-78 into a phyA phyB1 background. Triple mutant phyA phyB1 Z3-78 seedlings had a phenotype very similar to that of the 80B triple mutant described above, and screening under monochromatic light showed clearly that the Z3-78 mutation also specifically impaired the response to BL (data not shown). Complementation tests confirmed that Z3-78 and 80B were indeed carrying mutant alleles at the same locus (data not shown).
In addition to the seedling phenotype described above, the 80B and 68G triple mutants also showed a strong phenotype later in development relative to the phyA phyB1 double mutant. We quantified several characters in older glasshouse-grown plants of the 80B triple mutant, in comparison to its phyA phyB1 parent and the phyA phyB1 phyB2 triple mutant (Table 1). Although the 80B and phyB2 mutations had a similarly severe effect on elongation of the hypocotyl (Figure 1), the effect of the 80B mutation on internode elongation was much less severe than that of phyB2 (Table 1). The 80B and phyB2 mutations caused a similar reduction in lateral branching, and in chlorophyll content of the leaves and pericarp. In contrast, the clear effects of phyB2 on stem diameter, inflorescence length and fruit shape were not seen in the 80B mutant (Table 1).
Table 1. Effects of the 80B mutation on white-light-grown tomato plants
phyA phyB1 phyB2
Values expressed as mean ± SE.
Four fruits each from six plants measured at breaker stage.
Fruits harvested at mature green stage.
Length between main stem nodes 0 and 2 (cm)
5.9 ± 0.2
18.4 ± 0.5
7.9 ± 0.4
Stem diameter between nodes 4 and 5 (mm)
8.2 ± 0.3
5.9 ± 0.2
8.0 ± 0.2
Chlorophyll content of leaf 6 (mg g−1 FW)
3.3 ± 0.3
1.2 ± 0.1
1.8 ± 0.2
Total lateral shoot length at 8 weeks (cm)
73 ± 7
20 ± 4
9 ± 3
Leaf number at first inflorescence
7.8 ± 0.2
7.5 ± 0.3
7.5 ± 0.2
Number of flowers in first inflorescence
7.6 ± 0.3
7.8 ± 0.3
7.8 ± 0.5
Length between nodes 1 and 4 of first inflorescence (cm)
The Z3-78 and 80B mutants carry lesions in the gene encoding cryptochrome 1
In Arabidopsis, only two allelic groups of mutants are known to specifically impair de-etiolation responses to BL. These are the cry1 (hy4) and cry2 (fha) mutants which, respectively, carry lesions in genes encoding the BL photoreceptors cry1 and cry2 (Ahmad and Cashmore, 1993; Lin et al., 1998). Genes coding for these two cryptochromes are also present and expressed in tomato (Ninu et al., 1999; Perrotta et al., 2000). This suggested a high probability that the 80B, 68G and Z3-78 mutants would be mutated in one or other of these genes. Genomic sequences of CRY1 and CRY2 showed polymorphisms between Solanum lycopersicum and Solanum pennellii which were used to devise species-specific PCR-based markers for both genes (Perrotta et al., 2000). We used these markers to examine whether the photomorphogenic phenotype co-segregated with either gene in a mapping population of 54 plants derived from a cross between Z3-78 and S. pennellii. The Z3-78 phenotype segregated independently of the CRY2 gene on chromosome 9, but showed tight linkage with the CRY1 gene on chromosome 4 (data not shown).
We next checked whether the level or size of the CRY1 mRNA was altered in the mutants. RT–PCR analysis showed that CRY1 message was present at a normal level in both the Z3-78 and 80B mutants, but was absent in the 68G mutant (Figure 3a). Further analyses indicated that this mutant contains a large rearrangement at the CRY1 locus (data not shown). To localize the lesion in the other two mutants, we completely sequenced the CRY1 cDNA in Z3-78 and 80B.
We found that Z3-78 contained a G to A transition at position 1680, introducing a premature stop codon predicted to truncate the CRY1 protein at Pro444. The 80B mutant contained a single base deletion, at residue A921, with the resultant frameshift predicted to follow Glu192 with 24 nonsense amino acids before termination of transcription (Figure 3b). The predicted CRY1 product in the 80B mutant has lost all the putative sites for interaction with the FAD chromophore (Park et al., 1995), and 80B is therefore expected to be functionally null for CRY1. In the case of Z3-78, the predicted truncation occurs just before the C-terminal extension, leaving the putative chromophore-binding residues intact. However, this allele is also likely to be functionally null, as nonsense and mis-sense mutations in the carboxy terminus of Arabidopsis CRY1 abolish the functioning of that protein (Ahmad et al., 1995).
To check that the 80B mutant phenotype did in fact co-segregate with the identified lesion in the CRY1 gene, we used the elimination of an MboII site in 80B as a CAPS marker in a cross between the original 80B mutant and its phyA phyB1 progenitor line. We observed perfect co-segregation of phenotype and mutation in a population of 32 F2 plants. In this cross, seedlings heterozygous for the mutation could be easily identified as a distinct class (Figure 3c), indicating incomplete (26%) dominance of the WT allele. A similar haplo-insufficiency has also been observed for CRY1 in Arabidopsis (Koornneef et al., 1980) and for various phytochromes in both Arabidopsis and tomato (Koornneef et al., 1980; van Tuinen et al., 1995a; van Tuinen et al., 1995b; Whitelam et al., 1993).
Taken together, these data provide clear evidence that the 80B mutant phenotype results from mutation of the CRY1 gene. We therefore assigned the allele numbers cry1-1, cry1-2 and cry1-3 to the CRY1 alleles in Z3-78, 80B and 68G, respectively, according to the order of mutant isolation.
Selection of new photoreceptor mutant combinations
Of the three cry1 alleles, two were isolated as triple mutants in the cv. Moneymaker (MM) background. Using a combination of phenotypic screening and dCAPS markers for the phyA-1 and phyB1-1 mutations (Weller et al., 2000), we identified monogenic cry1 mutants and the cry1 phyA and cry1 phyB1 double mutants in progeny of a cross of the 80B triple mutant to WT cv. MM. Figure 4(a) shows that the monogenic cry1-2 mutant has a relatively mild phenotype when grown under continuous BL, demonstrating a partial redundancy between cry1 and phyA and/or phyB1.
The preliminary characterization of the triple mutant 80B (phyA phyB1 cry1-2) showed that this line retains a substantial residual response to BL (Figure 2). Phytochrome B2 clearly makes a contribution to BL sensing in the absence of phyA and phyB1 (Figure 2), and we therefore selected phyA phyB1 phyB2 cry1 quadruple mutants in order to determine whether any other photoreceptors, in addition to phyB2, might also contribute. Quadruple mutants were readily identified under white light in F2 progenies of a cross between phyA phyB1 phyB2 and phyA phyB1 cry1 triple mutants, on the basis of their extremely long, chlorophyll- and anthocyanin-deficient hypocotyls, and very small, unexpanded cotyledons relative to triple-mutant sister plants in the first 2 weeks after emergence (Figure 4c). However, after several weeks in WL, most quadruple mutant segregants managed eventually to achieve a relatively normal degree of cotyledon expansion (Figure 4b). Despite this, many failed to develop normally beyond the hypocotyl stage, with most seedlings developing into white plantlets with rudimentary leaves and very short internodes (Figure 4b,e). Plants lingered in this state for many weeks, and despite manipulation of temperature and light conditions, were never seen to support the development of normal flowers. The phyA phyB1 phyB2 cry1 mutant was therefore effectively lethal. In order to examine its light responses under monochromatic light, it was necessary to maintain it both as a CRY1/cry1 heterozygote in a phyA phyB1 phyB2 family for selection under BL, and as a PHYB2/phyB2 heterozygote in a phyA phyB1 cry1 family for selection under R.
Cryptochrome 1 and phytochromes A, B1 and B2 all contribute to control of de-etiolation under monochromatic blue light
To make an initial assessment of the relative contributions of cry1, phyA, phyB1 and phyB2 to BL sensing in tomato, we compared the effect of the loss of each of these photoreceptors on several aspects of de-etiolation under BL in WT and various mutant backgrounds. Figure 5 compares seedlings of various genotypes grown under BL with WT seedlings grown in complete darkness. Dark-grown seedlings of other genotypes did not differ significantly from WT for all three characters. It was not possible to include dark-grown quadruple mutants due to their lethality, and the fact that it was not possible to identify them in segregating progenies grown in darkness. The results in Figure 5 show firstly that there is no significant difference in hypocotyl length, cotyledon mass or hypocotyl anthocyanin content between the phyA phyB1 phyB2 cry1 quadruple mutant grown in BL and etiolated WT seedlings. This indicates that the residual response seen in the phyA phyB1 cry1 triple mutant is almost entirely accounted for by the action of phyB2. However, small differences in three other characteristics were consistently observed. After several days under BL, the apical hook in quadruple mutant seedlings showed partial opening relative to dark-grown seedlings of other genotypes (data not shown), and the cotyledons showed some visible chlorophyll accumulation and a relatively normal degree of unfolding (Figure 4d).
In general, of all the photoreceptors only phyA and cry1 had a significant effect on de-etiolation under BL in a WT background (Figure 5). An effect of phyB1 was revealed only in the absence of phyA and/or phyB2, and an effect of phyB2 only in the absence of phyB1. The loss of both phyB1 and phyB2 had an effect on hypocotyl length similar to the loss of phyA alone, but had only a very minor effect on anthocyanin accumulation (Figure 5).
We next examined irradiance dependence for each of the four photoreceptors, by measuring anthocyanin synthesis in response to 24 h BL irradiation of 4-day-old dark-grown seedlings. Figure 6 shows that the threshold irradiance for phyA activity in this response (deduced by comparing responses in WT and the phyA mutant) is ≈0.01 µmol m−2 sec−1, whereas that of cry1 is more than two orders of magnitude higher (≈3 µmol m−2 sec−1). In contrast to the strong effects of these photoreceptors, phyB1 had only a small effect at the highest irradiance (≈80 µmol m−2 sec−1), and phyB2 did not have a detectable effect across the entire irradiance range (Figure 6). However, a comparison of the phyA phyB1 and phyA phyB1 phyB2 mutants revealed a small effect of phyB2 at the highest irradiance, which was considerably amplified by the hp1 mutation (Figure 6). The hp1 mutation also amplified a residual response in the phyA phyB1 phyB2 triple mutant (Figure 6), suggesting that hp1 may also amplify responses to cry1. In this system there was no detectable anthocyanin synthesis in the phyA phyB1 cry1 triple mutant (data not shown).
Photoreceptor interaction in amplification of responsiveness to phytochrome by blue light
Physiological studies of photoreceptor interactions have often focused on the fact that, in many species, the phytochrome-mediated response of etiolated seedlings to a pulse of R can be enhanced by prior irradiation, given either as a pulse or as a period of continuous irradiation. The R, FR, BL and UV-A wave-bands are all known to be effective for this so-called ‘responsiveness amplification’ (Mohr, 1994). In general, BL and UV-A are the most effective, and this has been interpreted to suggest that a non-phytochrome BL receptor acts to amplify the phytochrome signal (Mohr, 1994). In tomato, single R pulses given to etiolated seedlings are ineffective for the induction of anthocyanin synthesis during the subsequent 24 h period. However, pretreatment for several hours with R or BL results in a substantial increase in effectiveness of the R pulse (Adamse et al., 1989; Peters et al., 1992). This response to the R pulse is partially reversible by FR, showing that it is a phytochrome response which is amplified. Loss of phyB1 resulted in the loss of virtually all the amplification response to a R pretreatment, but had little effect on the response to a BL pretreatment (Kerckhoffs et al., 1997). This showed that photoreceptors other than phyB1 are necessary for the BL enhancement of phytochrome responsiveness. We examined the possibility that cry1 might be one of these photoreceptors, by testing the effect of BL pretreatment on phytochrome-regulated anthocyanin synthesis in CRY1 and cry1 seedlings on different genetic backgrounds.
Figure 7 presents both the absolute level of anthocyanin after different treatments (top panel), and the derived values for the amount of R/FR reversible anthocyanin synthesis (bottom panel). These data show that the loss of phyB1 or cry1 alone had no effect on the level of anthocyanin accumulated in response to the BL pretreatment, or to the subsequent R and FR pulses. In contrast, the loss of phyA caused a clear reduction in both responses. In the absence of phyA, effects of both phyB1 and cry1 became apparent (Figure 7). Cry1 is therefore not the only photoreceptor active during the BL pretreatment, and both phyA and phyB1 can also contribute. Nevertheless it is clear that BL, acting through cry1, can enhance phytochrome responsiveness. It is of note that the small R/FR-reversible response remaining in the phyA phyB1 double mutant was eliminated by the further loss of either cry1 or phyB2. This implies specifically that cry1 can act to amplify responsiveness to phyB2. Finally, the phyB1 phyB2 double mutant retained a relatively large R/FR reversible response, attributable to phyA, phyE or phyF. This response was lost in the phyA phyB1 phyB2 triple mutant, but it is not clear whether this reflects a contribution of phyA to the amplification effect or to the inductive response itself.
Photoreceptor interaction under dual-wavelength irradiation
Interactions between phytochrome and cryptochrome photoreceptors have also been explored using dual-wavelength irradiations (Mancinelli, 1989; Mohr, 1994). Addition of R to background BL allows phytochrome action to be assessed under conditions where cryptochromes are also active. Conversely, addition of BL to background R allows selective activation of cryptochromes in the presence of a saturating phytochrome activity. Experiments of this nature have identified a conditional interaction between phyB and cry1 in the control of hypocotyl elongation in Arabidopsis (Casal and Mazzella, 1998). We used a similar experimental regime to examine the possibility that cry1 and phytochromes might interact in a similar way to control hypocotyl elongation and anthocyanin accumulation in tomato. We grew seedlings of various mutant combinations under continuous 7 µmol m−2 sec−1 R supplemented with either 3 or 24 h of 7 µmol m−2 sec−1 BL per day for 7 days post-emergence. Figure 8(b) presents derived data expressing the contribution of phyB1 and cry1 to the inhibition of hypocotyl elongation by this added BL, and the way in which this contribution depends on the presence or absence of other photoreceptors. Figure 8(b) shows that both phyB1 and cry1 contribute to the inhibition of hypocotyl elongation in response to 3 h daily BL exposure in WT seedlings grown under R, and that the effect of each photoreceptor is increased in the absence of the other. A similar result was seen for seedlings given 24 h daily BL, and for anthocyanin accumulation under both BL supplementation regimes (data not shown). This partially redundant action of phyB1 and cry1 contrasts with the apparent complete dependence of cry1 on phyB under limiting irradiances of BL in Arabidopsis (Casal and Mazzella, 1998).
These experiments revealed another interesting result. Figure 8(c) shows that WT seedlings and all phy mutant combinations (with the exception of the phyA phyB1 phyB2 triple mutant) responded to 3 h daily BL by an increase in anthocyanin synthesis. This was true even for lines carrying the cry1 mutation, suggesting that another BL photoreceptor, in addition to cry1, is active in mediating responses to short daily treatments with BL. It is unlikely that this response is due to the action of a phytochrome, because plants were grown under continuous R, and the added BL would not be expected to cause a substantial change in the phytochrome photoequilibrium (Jabben et al., 1982). Accumulation of additional anthocyanin in response to a further 21 h of daily BL irradiation was strongly reduced by the cry1 mutation in all backgrounds, suggesting that cry1 becomes the predominant BL-sensing photoreceptor under longer daily exposures. These results also hint at a complex interaction between the phytochromes and cry1, because the loss of phyA and/or phyB1 clearly enhances the effect of cry1, whereas the additional loss of phyB2 clearly reduces its effect (Figure 8c).
In this study, we have isolated cry1 mutants in a second higher plant species, tomato, and have used these mutants to explore the roles and interactions of cry1 in the control of tomato development. We have found that cry1 makes a strong contribution to several aspects of BL-induced de-etiolation, and in addition is clearly active in mature WL-grown plants. We have also used the cry1 mutants to explore the interactions of cry1 with phytochromes A, B1 and B2 in mediating responses to BL. Our results show that all four photoreceptors are capable of mediating responses to BL under some circumstances. However, the extent and relative importance of their individual contributions differ depending on irradiance, on which other photoreceptors are present, and on which response is examined.
Photoreceptor contributions to blue light sensing in tomato
We monitored BL effects on de-etiolation using three different traits: hypocotyl elongation, cotyledon mass, and anthocyanin accumulation in the hypocotyl. However, for more detailed photophysiological characterizations we focused on anthocyanin accumulation. Phytochrome A and cry1 appear to act additively, and together are responsible for ≈95% of the anthocyanin synthesis in the hypocotyls of WT seedlings grown under 10 µmol m−2 sec−1 BL. They also act additively to control hypocotyl elongation and cotyledon expansion, but are somewhat less dominant, accounting for only 65% of the WT response. Irradiance dependence of anthocyanin accumulation in cry1 seedlings shows that cry1 is active only at irradiances above ≈3 µmol m−2 sec−1. This value is consistent with previous reports of the threshold irradiance for cry1 in the inhibition of hypocotyl elongation in Arabidopsis (5 µmol m−2 sec−1, Poppe et al., 1998; 8 µmol m−2 sec−1, Lin et al., 1998). Tomato cry1 seedlings retain a normal anthocyanin accumulation response to lower irradiances of BL, which is mediated by phyA, as is also the case in Arabidopsis (Poppe et al., 1998). The threshold irradiance for cry1 action corresponds very closely to the saturating irradiance for phyA action, illustrating the way in which phyA and cry1 neatly combine to provide sensitivity to BL irradiance over a range of nearly four orders of magnitude.
Both phyB1 and phyB2 contribute to the high irradiance component of the response, and appear to act redundantly with cry1 as they have a significant effect on anthocyanin synthesis only in a phyA or phyA cry1 background. A stronger effect of phyB1 and phyB2 is seen in the control of hypocotyl elongation and cotyledon expansion, where the combined action of phyA and cry1 is somewhat weaker. Effects of phyB2 are seen only on a phyB1 background, indicating a redundancy between these phytochromes similar to that observed previously under R (Weller et al., 2000). For all three characters, the phyA phyB1 phyB2 cry1 quadruple mutant shows no measurable response to BL, indicating that all the de-etiolation under BL can be explained in terms of these four photoreceptors. We recently characterized several modes of interaction between phyA- and phyB-type phytochromes under R in tomato, including a phyB-dependent, phyA-mediated response to low-irradiance R, and phyA antagonism of a phyB1 response to high-irradiance R (Weller et al., 2000). There is no indication from the results presented here that phyA action under BL depends on either phyB1 or phyB2. This therefore suggests that phyA is not acting in the same way under low-irradiance BL and R.
Evidence for other blue light photoreceptors in tomato
In Arabidopsis a second cryptochrome, cry2, has been shown to contribute to BL inhibition of hypocotyl elongation in a narrow range of irradiance between 0.6 and 10 µmol m−2 sec−1 (Lin et al., 1998). However, we could not detect any anthocyanin in phyA;phyA cry1; or phyA phyB1 hp1 tomato seedlings exposed to 24 h of 1 µmol m−2 sec−1 BL. This suggests that in the absence of phyA, there is no other photoreceptor capable of inducing anthocyanin synthesis under these conditions. This could mean (i) that no other photoreceptors are active in mediating responses to BL in tomato; (ii) that they are present but functionally dependent on phyA; or (iii) that the light conditions used were simply not sufficient for demonstrating their activity.
We also could not identify any residual BL responsiveness for hypocotyl growth, cotyledon expansion or anthocyanin synthesis in the phyA phyB1 phyB2 cry1 quadruple mutant. However, the unfolding of quadruple mutant cotyledons under BL and their eventual expansion under WL suggests that there may be additional photoreceptors still active at a low level under these conditions. There is clear evidence that either phyE or phyF plays a significant role in tomato seedlings grown under high irradiances of WL (Weller et al., 2000), and it may be that the cotyledon expansion in the quadruple mutant is due to the action of one or both of these phytochromes. However, the fact that cry1 seedlings retain the ability to respond to BL when given on a background of R suggests that an additional non-phytochrome BL receptor may also be active in tomato. In addition, results for anthocyanin accumulation suggest that this photoreceptor may act primarily under short daily exposures to BL (Figure 8). Some precedent for this observation comes from studies of cry2 in Arabidopsis. The level of the cry2 protein is rapidly reduced in response to BL (Lin et al., 1998), suggesting a role for this photoreceptor under low irradiance or short daily exposures to BL (Lin et al., 1998). A homologue of Arabidopsis CRY2 is present and expressed in tomato (Perrotta et al., 2000), and transgenic analysis of its function is currently in progress.
Physiological interactions between specific photoreceptors in tomato
In tomato, cry1 can clearly act independently of phyA, phyB1 and phyB2, as the phyA phyB1 phyB2 triple mutant still shows cry1-dependent responsiveness both to BL, and to BL supplementation of R. In fact, for hypocotyl elongation in BL the relative effect of cry1 is greater in the absence of all three phytochromes than on any other background. However, several results indicate that phytochromes and cryptochromes can also interact in tomato. Expression of the de-etiolation response to an additional phytochrome (phyE or phyF) in a phyA phyB1 phyB2 background is seen only if BL photoreceptors are also activated (Weller et al., 2000). In the present study, pretreatment experiments show that cry1 activation in response to BL can modify subsequent responsiveness to phytochrome. More specifically, a dependence of phyB2-mediated anthocyanin biosynthesis on cry1 can be seen on a phyA phyB1 background in continuous BL and BL pretreatment experiments. The converse dependence of cry1 on phyB2 is also evident in BL pretreatment and R + BL experiments.
There has been considerable discussion on the question of whether effects of cry1 are dependent on phytochrome. It was initially suggested that the loss of phyA and phyB resulted in the loss of cry1 activity in Arabidopsis (Ahmad and Cashmore, 1997), but several groups have subsequently shown that the phyA phyB mutant does retain cry1-mediated responsiveness to BL (Neff and Chory, 1998; Poppe et al., 1998). However, physiological studies of photoreceptor interactions in Arabidopsis have identified conditions under which cry1 does interact with phyB-type phytochromes in an apparently specific manner. For example, WL-induced amplification for phytochrome control of hypocotyl elongation is absent in both cry1 and phyD mutants on a phyB background (Hennig et al., 1999). Also, under certain conditions the effects of cry1 are reduced in a phyB background, and vice versa (Casal and Mazzella, 1998). The main difference between tomato and Arabidopsis appears to be that in Arabidopsis these interactions are apparent on a phyB single mutant background (Casal and Mazzella, 1998; Hennig et al., 1999), whereas in tomato they are revealed only in the joint absence of phyA and phyB1.
Significance of physiologically defined photoreceptor interactions
At the physiological level, it is clear that cryptochrome activity can both modify and be modified by the action of phytochromes. It is also clear that specific interactions depend on irradiance and the presence or absence of other photoreceptors. The full potential for photoreceptor interaction may therefore be difficult to appreciate from the analysis of only a limited number of mutant combinations under a limited range of experimental conditions, and care must be taken in interpreting mutant phenotypes with respect to interactions and residual responses. For example, the apparent lack of response to BL in the phyA phyB double mutant of Arabidopsis led to claims that cry1 action is dependent on phyA or phyB1 (Ahmad and Cashmore, 1997). Similarly, the lack of an R-induced anthocyanin biosynthesis in the phyA phyB1 tomato mutant led to suggestions that additional phytochromes might not be active in R (Kerckhoffs et al., 1997), and the lack of a flowering response to R : FR in the phyB mutant of Arabidopsis gave the initial impression that phyB might be the sole photoreceptor for this response (Halliday et al., 1994). However, more detailed analyses have shown that what was initially interpreted as a lack of response was simply a threshold effect for expression of the response. In each case, residual responses were seen on examination at the molecular level (Batschauer et al., 1996), at higher irradiance (Weller et al., 2000), or on a more sensitive genetic background (Halliday et al., 1994).
Although a number of recent studies have provided evidence for photoreceptor reaction partners and signalling intermediates, there is, so far, no clear picture of how phytochrome and cryptochrome may interact at the molecular and biochemical level. One possibility is that cry1 and phyA may physically interact (Ahmad et al., 1998b). However, such an interaction is obviously not essential for the function of either photoreceptor, as each can act in the absence of the other (this study; Neff and Chory, 1998). Another possibility is that signalling via one pathway generates a component or components necessary for the other pathway. This has been most convincingly suggested in studies of BL-induced protoplast shrinking in Arabidopsis, where there is a 15 min lag phase between formation of Pfr and establishment of cry1 responsiveness (Wang and Iino, 1998). A third possibility is that photoreceptor signals converge at the level of transcriptional regulation, by differential activation of transcription factors recognizing the same cis element. For example, cry1, phyA and phyB all contribute to the induction of expression of the anthocyanin biosynthesis gene chalcone synthase (CHS) in Arabidopsis (Batschauer et al., 1996; Jackson and Jenkins, 1995). Pharmacological manipulations of signalling intermediates in protoplast systems suggest that phytochrome and cryptochrome effects occur via distinct pathways involving cGMP and Ca2+, respectively (Bowler et al., 1994; Christie and Jenkins, 1996). Nevertheless, a small light-responsive element (LRE) in the CHS promoter is sufficient to drive both phytochrome-and BL-induced expression in Arabidopsis (Batschauer et al., 1996; Hartmann et al., 1998). The Arabidopsis bZIP transcription factor HY5 binds a consensus sequence present in this LRE and is essential for normal CHS induction by both WL and R (Ang et al., 1998; Martínez-García et al., 2000), while other bZIP transcription factors binding to the same sequence are necessary for induction under BL/UV-A in parsley (Sprenger-Haussels and Weisshaar, 2000).
These studies suggest that interactions between phytochrome and cryptochrome may occur in several different ways. Further physiological studies focusing on the specificity and temporal requirements of these interactions will be of great benefit, and will help in designing specific new screens for additional light-signalling components. More detailed pharmacological dissection of phytochrome and cryptochrome signalling pathways will also be of interest, and the possibility of conducting these experiments in tomato will be greatly enhanced by the new range of photoreceptor-deficient mutant combinations characterized here.
Plant material, growth conditions and light sources
Seedling measurements and pigment assays were performed as described elsewhere (Weller et al., 2000). Apical hook angles were measured with a protractor after photocopying seedlings fixed to a sheet of paper with adhesive tape. Chlorophyll content in fruits was measured in pericarp sections obtained by taking a core section 5 mm in diameter through the equator of the fruit.
Molecular analysis of tomato CRY1
Primers used for detection of S. lycopersicum/S. pennellii polymorphisms in the CRY1 and CRY2 genes have been described elsewhere (Perrotta et al., 2000). These polymorphisms were tested for co-segregation with the BL-insensitive mutant phenotype in the F1 generation of the second test-cross of the monogenic Z3-78 mutant to S. pennellii. Semi-quantitative RT–PCR amplification of the CRY1 transcript was carried out by reverse-transcribing 1 µg total RNA using the Superscript II Reverse Transcriptase (Life Technologies, Rockville, MD, USA) and oligo d(T) primer according to the manufacturer's specifications, in a volume of 50 µl. Total RNA was extracted from 8-day-old etiolated tomato seedlings as previously described by Giuliano et al. (1993). To test the level of gene expression, PCR reactions were carried for 30 cycles in a volume of 20 µl, using 2 µl of cDNA undiluted, and diluted 1 : 3 and 1 : 10. The primers were TCCTCTTCCTATAGTTGA (forward) and CTCTGTATTAGC CACTTG (reverse) for the CRY1 gene, and oligonucleotides ATTGTGGTCATTGGYCAYGT (forward) and CCAATCTTGTAV ACATCCTG (reverse) for the EF-1α control (Mahe et al., 1992). The different dilutions were loaded on a 1.5% (w/v) agarose/ethidium bromide gel at 15 min intervals, and quantified using a digital camera and NIH image software on a Macintosh G3 computer. For detection of mutations, the full-length CRY1 cDNA was amplified using the oligonucleotides ATACTTCTGTCTAAG AAAGAC and CTCTGTATTAGCCACTTG. RT–PCR products were directly sequenced using a Perkin-Elmer 373A automated sequencer (Perkin-Elmer, Foster City, CA, USA) and dye terminator chemistry.
PCR markers for identification of mutant alleles
PCR-based markers for the phyA-1, phyB1-1 and phyB2-1 alleles have been described elsewhere (Weller et al., 2000). The cry1-1 and cry1-2 mutations, respectively, eliminated NcoI and MboII restriction sites, and were detected as CAPS markers using the following primers: cry1-1 forward GGGCCTCAATCTTCTTAC, reverse ATGCTATTGGGGAGTCAG; cry1-2 forward ACCGCC GCTTGTAGGATT, reverse CCGACGGTTCTGCGAGTA. All PCR reactions were performed in 25 µl, with a stepped program consisting of 45 sec denaturation at 94°, 45 sec annealing at 50°C, and 45 sec elongation at 72°C for 30 cycles.
We thank Corrie Hanhart and Chris Kendrick for help with seed administration and early stages of mutant selection; Tanja Borst-Vrenssen for conducting the irradiance response experiments; Valerie Hecht and Vera Quecini for assistance with seedling measurements; Jan Laurens and other Unifarm staff for care of plants and processing of seed; and Elena Nebuloso for sequencing the cry1 alleles. Supported by the European Community – BIOTECH B104-CT97-2124 ‘PHOTARCH’ to G.G., M.K. and R.E.K.