The plant photoreceptor phytochrome A utilizes three signal transduction pathways, dependent upon calcium and/or cGMP, to activate genes in the light. In this report, we have studied the phytochrome A regulation of a gene that is down‐regulated by light, asparagine synthetase (AS1). We show that AS1 is expressed in the dark and repressed in the light. Repression of AS1 in the light is likely controlled by the same calcium/cGMP‐dependent pathway that is used to activate other light responses. The use of the same signal transduction pathway for both activating and repressing different responses provides an interesting mechanism for phytochrome action. Using complementary loss‐ and gain‐of‐function experiments we have identified a 17 bp cis‐element within the AS1 promoter that is both necessary and sufficient for this regulation. This sequence is likely to be the target for a highly conserved phytochrome‐generated repressor whose activity is regulated by both calcium and cGMP.
Light is perceived in plants by three major classes of photoreceptors: the phytochromes, the blue/UVA receptors (cryptochromes) and the UVB receptors (Quail et al., 1995). Of these, the most intensively studied are the phytochromes, which exist in two photo‐reversible forms: the red light absorbing form, Pr, generally considered to be physiologically inactive, and the far‐red absorbing form, Pfr, known to mediate a broad range of plant responses to light (Quail et al., 1995; Smith, 1995; von Arnim and Deng, 1996). Some responses mediated by Pfr can be reversed by far‐red light, which converts Pfr back to Pr.
In higher plants, the phytochromes are encoded by multigene families (Quail et al., 1995). Each phytochrome is thought to have a different physiological role and the recent availability of mutants deficient in individual phytochromes is allowing further definition of these specificities (reviewed in Millar et al., 1994; Quail et al., 1995; von Arnim and Deng, 1996). Some responses have now been linked to particular phytochromes, although there nonetheless appears to be some overlap between the functions of individual phytochromes within any given plant species (Reed et al., 1994).
The different phytochromes make up two distinct classes, known as type I and type II (Quail et al., 1995; Smith, 1995). Type I phytochromes are the most abundant in dark‐grown plants, but they are light labile due to the rapid degradation and/or sequestration of the Pfr form in the light. In contrast, the type II phytochromes are present in much lower amounts, but their stability in the Pfr form ensures that they are predominant in light‐grown plants. Hence, type I phytochrome is thought to play a specific role during the initial de‐etiolation process, whereas type II may be more important for mediating phytochrome responses in mature plants. Phytochrome A (PHYA) is the only type I phytochrome to have been identified and it may in fact be the only molecular species within the type I pool (see Clack et al., 1994). Like the PHYA apoprotein, PHYA mRNA abundance also decreases in the light (see Sharrock and Quail, 1989, and references therein), particularly in monocotyledons, where down‐regulation of PHYA gene expression has been found to be mediated by an autoregulatory mechanism involving phytochrome itself (Lissemore and Quail, 1988).
In addition to PHYA, several other genes have been found to be down‐regulated by light. These include genes encoding NADPH protochlorophyllide oxidoreductase (Mösinger et al., 1985), β‐tubulin (Colbert et al., 1990; Tonoike et al., 1994; Leu et al., 1995), asparagine synthetase (AS1) (Tsai and Coruzzi, 1990, 1991), the homeodomain proteins Athb‐2 and Athb‐4 (Carabelli et al., 1993) and two genes denoted NPR1 and NPR2 in Lemna (Okubara et al., 1993). Phytochrome regulates these responses and two formal possibilities can be considered to account for how it does so (Bruce et al., 1991): (i) Pfr generates a repressor in the light; (ii) Pr generates an activator in the dark. Current knowledge of phytochrome function would tend to favour Pfr repression as the most likely mechanism, because much evidence implicates Pfr, and not Pr, in controlling many other responses. However, it has proved extremely difficult to design physiological experiments that could definitively distinguish between the two possibilities.
In this report, we present the results of experiments that can discriminate between Pfr repression and Pr activation as possible mechanisms controlling the down‐regulation of gene expression in the light. Specifically, we have studied the signal transduction events stimulated by PHYA to regulate expression of one of these negatively light regulated genes, AS1, using microinjection to deliver individual molecules into the cells of wild‐type and aurea mutant tomato seedlings, as previously described (Neuhaus et al., 1993). PHYA is present in etiolated seedlings of the aurea mutant at 20% wild‐type levels and is spectrally inactive, whereas PHYB (a type II phytochrome) is present and active at normal levels (Sharma et al., 1993). In contrast to the behaviour of wild‐type seedlings, chloroplasts and anthocyanin pigments fail to develop within the hypocotyl cells of etiolated aurea seedlings in response to light. However, a wild‐type phenotype can be restored to aurea hypocotyl cells by injection of exogenous PHYA (Neuhaus et al., 1993). This system therefore allows the manipulation and subsequent dissection of the signal transduction pathways used by PHYA by identifying agonists or antagonists of these responses. In this way, we have previously reported that the Pfr form of PHYA (PfrA) acts through heterotrimeric G proteins to stimulate gene expression that results in chloroplast development and anthocyanin biosynthesis (Neuhaus et al., 1993). Three different signal transduction pathways downstream of the G protein were subsequently identified that require cGMP and calcium (Bowler and Chua, 1994; Bowler et al., 1994a). cGMP can stimulate genes such as chalcone synthase (CHS) that are required for anthocyanin biosynthesis, whereas calcium and calcium‐activated calmodulin (CaM) can stimulate other genes (e.g. chlorophyll a,b binding protein genes, CAB) necessary for partial chloroplast development. A third pathway, that requires both calcium and cGMP, is utilized to stimulate genes encoding the photosystem I (PSI) and cytochrome b6f (cyt. b6f) complexes (e.g. the gene encoding ferredoxin NADP+ oxidoreductase, FNR). The combination of these three pathways therefore leads to full chloroplast development and anthocyanin biosynthesis.
Using similar experiments we wanted, specifically: (i) to address whether PfrA, PrA or both control AS1 regulation, (ii) to determine whether AS1 regulation requires calcium and/or cGMP or whether other signalling molecules are utilized and (iii) to identify specific cis‐elements within the AS1 promoter which are targets of PHYA regulation. Our results show that PfrA represses AS1 expression in the light and that it does so via the calcium/cGMP‐dependent pathway used to activate other responses, such as FNR gene expression. Hence, probably the same signal transduction pathway is used to simultaneously ‘turn on’ and ‘turn off’ different events. One cis‐element within the AS1 promoter, which in our assay system displays all the properties of the intact promoter, is highly homologous to the RE1 element within the oat PHYA gene, previously proposed to be a target for phytochrome autoregulation (Bruce et al., 1991).
AS1–GUS is negatively regulated by PfrA
To examine the regulation of AS1 by phytochrome, a plasmid containing 559 bp of the pea AS1 promoter (Tsai, 1991) fused upstream of the gene encoding the reporter β–glucuronidase (AS1–GUS) was injected into subepidermal hypocotyl cells of 7‐ to 10‐day dark‐grown wild‐type and aurea mutant tomato seedlings. For comparison, equivalent experiments were also performed with a CAB–GUS reporter gene (Neuhaus et al., 1993). Following injection (under green safelight conditions where necessary) the seedlings were exposed to different light irradiations. As we would predict from expression of the endogenous AS1 and CAB genes, AS1–GUS was expressed in injected cells of wild‐type seedlings maintained in the dark but not in the light, whereas CAB–GUS was only expressed in the light (Table I). Furthermore, expression of AS1–GUS in the dark could be down‐regulated by a pulse of red light, but reactivated by 10 min of far‐red irradiation subsequent to the red light pulse. In contrast, CAB–GUS expression could be stimulated in the dark by a pulse of red light and could be down‐regulated by a far‐red light pulse given immediately after the red light irradiation (Table I). These results thus demonstrate that in wild‐type seedlings both AS1–GUS and CAB–GUS expression are regulated by phytochrome, but that this regulation acts in opposite ways, in one case down‐regulating and in the other case up‐regulating expression. Furthermore, the behaviour of the AS1–GUS gene in these injection experiments clearly reflects endogenous AS1 expression in pea, which has been previously shown to be down‐regulated at the level of transcription by white and red light (Tsai and Coruzzi, 1990, 1991).
In injected cells of aurea seedlings both reporter genes were insensitive to the light conditions: AS1–GUS was expressed both in the light and in the dark, whereas CAB–GUS was never expressed (Figure 1 and Table I). The lack of expression in aurea of CAB–GUS, even in the light or after a red light pulse, is consistent with its known requirement for Pfr, because, unlike in the wild‐type, etiolated aurea seedlings are largely deficient in phytochrome (Sharma et al., 1993). Furthermore, the fact that in aurea AS1–GUS is expressed under all conditions implies that Pfr normally represses AS1–GUS expression in the light but that in phytochrome‐deficient cells it is expressed constitutively.
We have previously found that injection of PHYA into hypocotyl cells of etiolated aurea seedlings in the light can restore chloroplast development and anthocyanin biosynthesis and can activate expression of CAB–GUS, CHS–GUS and FNR–GUS reporter genes (Figure 1 and Table II; Neuhaus et al., 1993; Bowler et al., 1994a). To determine whether PHYA could also regulate AS1–GUS expression, we co‐injected AS1–GUS together with PHYA into aurea hypocotyl cells. We found that injection of PfrA (i.e. injection of PHYA in white light conditions) was able to down‐regulate AS1–GUS expression in aurea, whereas injection of the Pr form (PrA) (i.e. injection of PHYA in green safelight conditions) could not (Figure 1 and Table II). Furthermore injection of PrA, followed by its conversion in situ to PfrA by a red light pulse could also inhibit expression. This down‐regulation by red light could, however, be relieved by subsequent irradiation with far‐red light (Table II). These results thus demonstrate that PHYA can control AS1–GUS expression and that it does so in an opposite way compared with CAB–GUS, CHS–GUS and FNR–GUS (Neuhaus et al., 1993; Bowler et al., 1994a).
Down‐regulation of AS1–GUS by PfrA requires calcium and cGMP
Previous microinjection experiments in aurea, together with pharmacological studies in soybean SB‐P cells, have led to the identification of three major signal transduction pathways used by PfrA to control chloroplast development and anthocyanin biosynthesis (Neuhaus et al., 1993; Bowler and Chua, 1994; Bowler et al., 1994a,b). It was, therefore, of interest to determine whether these pathways are not only used for activation of these responses but also for down‐regulation of other responses, e.g. negative regulation of AS1 expression. To test this, we co‐injected a range of previously characterized molecules known to stimulate various PfrA responses. Activation of heterotrimeric G proteins, by injection of GTPγS and cholera toxin (CTX), has been shown to stimulate full chloroplast development and anthocyanin biosynthesis in aurea hypocotyl cells (Neuhaus et al., 1993) and to activate the reporter genes CAB–GUS, FNR–GUS and CHS–GUS (Figure 1 and Table II; Bowler et al., 1994a). In contrast, co‐injection of GTPγS and CTX with AS1‐GUS in aurea led to down‐regulation of AS1–GUS and, unlike with PfrA, this response was now unaffected by the light conditions (Figure 1 and Table II). Hence, the response was now light‐independent, i.e. it had been uncoupled from the normal stimulus. These data therefore demonstrate that, as for CAB–GUS, FNR–GUS and CHS–GUS activation (Neuhaus et al., 1993; Bowler et al., 1994a), the PfrA‐mediated down‐regulation of AS1–GUS requires G protein activation and also reveal that there are no light‐requiring steps downstream of G protein activation for AS1–GUS down‐regulation. This has also been shown for CAB–GUS activation, indicating that the only light‐dependent step between PfrA and nuclear gene regulation is likely to be photoreceptor activation (Neuhaus et al., 1993).
Injection of calcium and activated calmodulin (CaM) have been found to stimulate CAB–GUS expression and partial chloroplast development in etiolated aurea hypocotyl cells (Figure 1 and Table II; Neuhaus et al., 1993). Conversely, injection of cGMP can stimulate CHS–GUS expression and anthocyanin biosynthesis (Figure 1 and Table II; Bowler et al., 1994a). These molecules therefore control distinct subsets of PfrA responses and act downstream of G protein activation (Neuhaus et al., 1993; Bowler et al., 1994a). To determine if these previously characterized PfrA signalling intermediates also regulate AS1 expression, they were co‐injected with AS1–GUS into aurea. Interestingly, neither calcium, activated CaM nor cGMP alone (at concentrations previously found to be effective, 2 μM, 10 000 molecules, and 50 μM, respectively, estimated final intracellular concentrations; Neuhaus et al., 1993; Bowler et al., 1994a,b) could down‐regulate AS1–GUS expression in the light in aurea cells (Figure 1 and Table II). However, a combination of calcium or activated CaM together with cGMP was able to effectively block AS1–GUS expression (Figure 1 and Table II), suggesting that the down‐regulation of AS1–GUS by PfrA is controlled by the same signalling molecules that it uses to activate other responses. Specifically, it appeared that AS1–GUS down‐regulation may be controlled via the same calcium/cGMP‐dependent pathway we have found to activate expression of genes encoding PSI and cyt. b6f components, such as FNR (Figure 1 and Table II; Bowler et al., 1994a).
PfrA signal transduction pathways have been found to be subject to cross‐talk regulation, which has been termed reciprocal control (Bowler et al., 1994b). For example, activity of the calcium/cGMP‐dependent pathway has been found to be inhibited by high concentrations of cGMP, but not activated CaM, and to be able to function with significantly lower amounts of cGMP (at least 6‐fold) than does the cGMP‐dependent pathway. To examine whether regulation of AS1–GUS expression was also modulated by these phenomena, we co‐injected different concentrations of activated CaM and cGMP. Indeed, high concentrations of cGMP (110 μM) injected with activated CaM (3000 molecules) were no longer effective in down‐regulating AS1–GUS, whereas, in the presence of high concentrations of activated CaM (100 000 molecules) and normal amounts of cGMP (50 μM), down‐regulation was still observed, as it was when co‐injecting low levels of cGMP (3.5 μM) with activated CaM (10 000 molecules) (Table II). Again, these results indicated that AS1–GUS down‐regulation by PHYA was likely mediated by the same calcium/cGMP‐dependent pathway that has been previously characterized as activating other responses (Bowler et al., 1994b).
It was interesting to observe that in these experiments with PfrA signalling intermediates, phenotypes characteristic of both dark‐ and light‐exposed material were manifested concurrently in the same cell, e.g. although injection of calcium or activated CaM alone in the light resulted in CAB–GUS activation and biogenesis of partially developed chloroplasts and injection of cGMP alone resulted in CHS–GUS activation and anthocyanin pigment biosynthesis, in both cases these cells could not down‐regulate AS1–GUS (Figure 1 and Table II).
As further evidence that AS1–GUS down‐regulation was mediated by the previously characterized calcium/cGMP‐dependent pathway, we tested the effect on AS1–GUS expression of previously characterized pharmacological agents. Genistein (an inhibitor of tyrosine and histidine protein kinases; Huang et al., 1992) is known to inhibit the cGMP‐dependent pathway, whereas trifluoperazine (a calmodulin antagonist; Massom et al., 1990) and staurosporine (a non‐specific protein kinase inhibitor; Rüegg and Burgess, 1989) both inhibit the two calcium‐dependent pathways (Bowler et al., 1994b). For these experiments, we injected dark‐grown wild‐type seedlings and then incubated them in the light in the presence of these different compounds. For comparison, we also examined the expression of CAB–GUS, CHS–GUS and FNR–GUS under the same conditions. As predicted from previous experiments in aurea (Bowler et al., 1994b), CAB–GUS, CHS–GUS and FNR–GUS were expressed in the light in these wild‐type seedlings (Table III). Furthermore, as already observed in aurea, CHS–GUS expression was inhibited by genistein, whereas CAB–GUS and FNR–GUS expression were inhibited by trifluoperazine and staurosporine (Table III). These results reveal the consistency of data obtained from aurea and wild‐type seedlings.
AS1–GUS, however, was not expressed in the light, as previously observed (Table I), and this down‐regulation by light was found to be sensitive to trifluoperazine and staurosporine, but not to genistein (Table III). Based on these data, together with that presented in Table II, it is therefore highly likely that the same signal transduction pathway (i.e. the calcium/cGMP‐dependent pathway) is used by PfrA to control both up‐regulation of some genes (e.g. FNR–GUS) and down‐regulation of others (e.g. AS1–GUS).
The target of calcium and cGMP regulation within the AS1 promoter
The above data imply that there is a target(s) within the AS1 promoter for PfrA‐mediated down‐regulation by calcium and cGMP. Most simply, PfrA may act via calcium and cGMP to activate a repressor that binds to such a sequence. To date, the best characterized cis‐acting element found to be important for phytochrome‐mediated down‐regulation is RE1, an 11 bp GC‐rich sequence centered at −75 bp within the oat PHYA promoter (Bruce et al., 1991). When the RE1 sequence is mutated by linker scanning mutagenesis, this promoter retains maximal expression following a far‐red light pulse but is no longer down‐regulated by a red light pulse (Bruce et al., 1991). Interestingly, the RE1 core sequence, TGGG, is present within other PHYA promoters and can also be found in the promoters of all genes so far characterized as being down‐regulated by light (Figure 2). Examination of the AS1 promoter sequence revealed the presence of two such sequences, albeit on the opposite DNA strand with respect to monocotyledon PHYA promoters, showing significant homology with the RE1 core sequence, one centered at −43 and the other centered at −160 (Figure 2). Thus, it appeared possible that these elements may be the targets for PfrA‐mediated repression within the AS1 promoter. To determine whether these sequences were required for down‐regulation of AS1–GUS by light, we performed competition experiments using a tetramer of the most proximal RE1‐related element within the AS1 promoter (denoted RE3, centered at −43) (Figure 2). Similar competition experiments have recently been performed in tobacco cotyledon cells to study regulation of the cauliflower mosaic virus (CaMV) −90 35S promoter (Neuhaus et al., 1994).
Co‐injection of AS1–GUS (5000 molecules) with a plasmid containing the RE3 tetramer sequence into wild‐type cells indeed resulted in inhibition of the down‐regulation of AS1–GUS normally observed in the light (Table IV). A 4‐fold molar excess of the competitor (i.e. 5000 molecules) was sufficient to cause this effect, although higher concentrations were more effective (Table IV). In contrast, a tetramer of an RE3 element containing a mutated core sequence (RE3m) (Figure 2) was not able to inhibit down‐regulation by light, even when injected at an 80‐fold molar excess (i.e. 100 000 molecules) per cell (Table IV). The sensitivity of AS1–GUS down‐regulation to competition specifically by the RE3 tetramer therefore strongly implies that a light‐activated repressor indeed interacts with the RE3 element and that its removal (by competition) results in release of repression of the AS1 promoter in light. Conversely, we have found that introduction of a large excess of AS1–GUS molecules by microprojectile bombardment also results in de‐regulated expression (data not shown), suggesting again that a repressor is being titrated out. When co‐injected into aurea cells, neither RE3 nor RE3m had any effect on AS1–GUS expression, i.e. the reporter gene was always expressed (Table IV). This is consistent with the notion that in aurea this repressor is either not present or not active, due to the phytochrome deficiency in mutant seedlings.
To relate the activity of the repressor to PfrA and to the PfrA signalling intermediates, we performed experiments in aurea co‐injecting AS1–GUS with RE3 or RE3m, together with various signalling intermediates. Normal repression of AS1–GUS by PfrA co‐injection into aurea cells in light could indeed be inhibited in the presence of sufficient amounts of the RE3 tetramer (an 8‐fold molar excess), although RE3m was not able to inhibit AS1–GUS repression by PfrA (even at an 80‐fold molar excess) (Table V). Hence, we can conclude that down‐regulation of AS1–GUS via RE3 in light is due to PfrA repression rather than PrA activation. Furthermore, repression of AS1–GUS expression by co‐injection of calcium and cGMP could be similarly competed by the RE3 tetramer, but not by RE3m (Table V). These data thus indicate that the RE3 element within the AS1 promoter is necessary for PfrA‐mediated repression of AS1–GUS and requires either calcium, cGMP or both.
The RE3 repressor is a target for calcium and cGMP
The experiments described above do not demonstrate unequivocally that RE3 is a negative element regulated by both calcium and cGMP. It is possible, for example, that RE3 is a target for only one of the PfrA signalling intermediates and that repression of AS1–GUS by calcium and cGMP is mediated by interactions between the RE3 binding factor and other DNA binding proteins recognizing different cis‐elements within the AS1 promoter. To test the role of the RE3 element more precisely, we inserted the RE3 tetramer between the 35S B domain (−343 to −90) and the minimal −46 35S TATA box, which is normally constitutively expressed in both light and dark (Lam and Chua, 1990). The artificial promoter was placed upstream of GUS (35S–RE3–GUS). When injected into wild‐type cells, we found that expression of this reporter gene was now repressed in light, in the same way as was AS1–GUS (Table VI). Furthermore, consistent with the behaviour of AS1–GUS, 35S–RE3–GUS was not repressed in injected aurea cells by light (Table VI). The RE3m tetramer, however, could not confer light repression on the reporter gene in either wild‐type or aurea cells. Clearly then, the RE3 element is both necessary and sufficient to mediate light repression and can function in a heterologous context.
We tested whether RE3 itself was a target for PfrA, and for calcium and cGMP, by injecting 35S–RE3–GUS into aurea cells together with these signalling intermediates. Consistent with the above data, co‐injection with PfrA in light resulted in repression of GUS expression from 35S–RE3–GUS but not from 35S–RE3m–GUS, whereas injection of PrA (performed under a green safelight) and subsequent incubation of seedlings in darkness did not result in 35S–RE3–GUS or 35S–RE3m–GUS repression (Table VI). Furthermore, although injection of neither calcium, activated CaM nor cGMP alone had any repressive effect, as with AS1–GUS, a combination of activated CaM with cGMP resulted in repression of 35S–RE3–GUS. This was not observed with 35S–RE3m–GUS (Table VI). These experiments therefore demonstrate that RE3 is a target for both calcium and cGMP and that, at least in these experiments, it can mediate light repression in an identical manner to the intact AS1 promoter.
The work presented in this manuscript is a continuation of our use of the tomato aurea mutant for dissection of PHYA signal transduction. We have previously used this mutant to identify positively acting signalling intermediates controlling PfrA‐activated chloroplast development and anthocyanin biosynthesis (Neuhaus et al., 1993; Bowler et al., 1994a). Unfortunately, several inadvertant mistakes in data presentation were made in these articles [see Erratum, Cell, 1994, 79(4)]. Although these mistakes did not affect the conclusions of our experiments, we have nonetheless repeated key experiments relevant to the essential features of the scheme. Our new experiments confirming the identity of the three signalling pathways controlling CAB, CHS and FNR gene expression are shown in Figure 1 and Table II. Additionally, a combination of microinjection experiments in aurea and physiological analyses in SB‐P cultures has allowed us to begin to understand cross‐talk phenomena acting between different PfrA signal transduction pathways (e.g. reciprocal control) (Bowler et al., 1994b) and other experiments have revealed that distinct phytochrome‐responsive cis‐elements are controlled by calcium and cGMP (Wu et al., 1996).
The phenotype of the aurea mutant is rather complex. Although likely to be a mutation affecting chromophore biosynthesis (Terry and Kendrick, 1996), biochemical and physiological experiments have clearly indicated that the mutation affects primarily PHYA, i.e. functional PHYA is absent in dark‐grown seedlings and the mutant displays type II phytochrome‐regulated end of day far‐red responses (van Tuinen et al., 1996). However, the aurea phenotype is not wholly consistent with that of Arabidopsis phyA null mutants (Whitelam and Harberd, 1994) nor with recently isolated tomato phyA mutants (van Tuinen et al., 1995). We nonetheless believe that the signalling pathways elucidated in this and previous articles are controlled by PHYA, because injection of PHYA should rescue only PHYA‐mediated events. This is supported by recent observations that injection into aurea of recombinant reconstituted PHYA produces identical responses, whereas equivalent concentrations of PHYB do not (Kunkel et al., 1996).
In the present report, we have performed a series of microinjection experiments to elucidate how PHYA down‐regulates expression of certain nuclear genes in light. In particular, we have used the promoter of the pea AS1 gene as a target. Tsai and Coruzzi (1990) have previously shown that this gene is highly expressed in the dark but rapidly down‐regulated in light. Moreover, down‐regulation is mediated by phytochrome primarily at the transcriptional level (Tsai and Coruzzi, 1991). In agreement with this data, our current results show that an AS1–GUS chimeric gene is down‐regulated by phytochrome in wild‐type tomato cells, whereas in the aurea mutant it is expressed constitutively, regardless of the irradiation conditions (Figure 1 and Table I). Furthermore, by restoring negative light regulation of the AS1–GUS gene in aurea by co‐injection with PfrA, we have been able to show that PHYA can specifically mediate this expression pattern (Figure 1 and Table II).
The down‐regulation of AS1–GUS by PfrA requires G proteins, calcium and cGMP (Figure 1 and Table II), previously characterized as signalling intermediates for PfrA‐mediated activation of anthocyanin biosynthesis and chloroplast development (Bowler et al., 1994a). By all known criteria (Bowler et al., 1994b), this down‐regulation appears to be controlled by the same signal transduction pathway that is used to activate the FNR promoter: down‐regulation is blocked by high concentrations of cGMP but not CaM, it can be inhibited by trifluoperazine and staurosporine but not by genistein and it requires only low concentrations of cGMP (Tables II and III). Hence, PfrA appears to use the same signal transduction pathway to both activate (e.g. FNR) and down‐regulate (e.g. AS1) different genes. This implies that there are different oppositely acting targets for the same PfrA signal transduction pathway, an efficient and mechanistically simple means for concurrently activating and repressing different responses. The identification of other phytochrome responses that are oppositely regulated by a single PfrA signalling pathway will allow a better assessment of the physiological importance of this novel regulatory mechanism.
It has been proposed that expression of genes that are down‐regulated by light, such as AS1, may be modulated by phytochrome either by Pfr repression in light or by Pr activation in the dark (Bruce et al., 1991). Although current knowledge would tend to favour the former mechanism, the lack of experimental tools has made it impossible to distinguish definitively between these two possibilities. Concerning the regulation of AS1 by phytochrome, our current experiments have demonstrated: (i) that in wild‐type cells AS1–GUS can be down‐regulated by red light whereas in aurea it cannot (Table I); (ii) that co‐injection of PfrA in aurea can prevent AS1–GUS expression in light (Table II). This information would suggest that PfrA is the mediator of AS1 down‐regulation in light and that in its absence AS1–GUS is expressed regardless of whether PrA is present or not. However, as with previous data, these experiments do not prove that PfrA, and not PrA, is the active molecule. More definitive experiments, however, have shown: (i) that it is possible to prevent PfrA‐mediated down‐regulation of AS1 by co‐injection of a specific tetramer sequence corresponding to a putative cis‐element within the AS1 promoter (Table IV); (ii) that this sequence by itself is sufficient to confer PfrA‐mediated down‐regulation on a heterologous constitutively active 35S promoter (Table VI). These observations therefore provide compelling evidence that PfrA is the mediator of AS1 down‐regulation in light and that it functions by activating a putative repressor that binds to this cis‐element.
The 17 bp cis‐element, denoted RE3, that we have identified as the binding site of the putative repressor, is centered at −43 and contains the TGGG core motif that is present within the promoters of all other genes so far characterized as being down‐regulated by light (Figure 2). Another similar sequence is centered at −160 bp (although its activity has not currently been tested). The importance of cis‐elements containing the TGGG core motif was initially inferred from studies with the oat PHYA promoter. In this latter case, linker scanning mutagenesis indicated that an 11 bp sequence containing a TGGG motif, denoted RE1, was a target for Pfr‐mediated negative regulation (Bruce et al., 1991). In loss‐of‐function experiments, we have corroborated this data by showing that a co‐injected RE3 tetramer can prevent down‐regulation of AS1–GUS mediated by light in wild‐type cells (Table IV) and by PfrA and CaM and cGMP in aurea cells (Table V), although a tetramer containing a mutated core sequence (RE3m) is ineffective. The simplest interpretation of these results is that the RE3 tetramer is able to compete away a repressor that binds this sequence within the AS1 promoter. If this is the case, a constitutively active activator/enhancer must also interact with the AS1 promoter. Consequently, the AS1 promoter would be constitutively active in the dark, due to the absence of active repressor, whereas activation of the repressor by PfrA in light would block activity of this positive element and hence inhibit expression. Such a mechanism has also been proposed for autoregulation of oat PHYA (Bruce et al., 1991).
The most convincing evidence that this sequence binds a PfrA‐generated repressor was derived from gain‐of‐function experiments: when placed within the constitutive 35S promoter (between the B domain and the −46 TATA box) the RE3 tetramer was sufficient to confer light repression in wild‐type cells and PfrA‐mediated repression in aurea cells (Table VI). Again, the RE3m tetramer was ineffective. Furthermore, both calcium (or CaM) and cGMP were required to reproduce the repression mediated by PfrA in aurea (Table VI), indicating that the putative repressor that binds to RE3 requires both signalling molecules for activation.
The homology between RE1 and RE3, at both the structural and functional levels, would strongly suggest that they are binding sites for the same (or at least a highly related) repressor, even though the RE1 and RE3 cis‐elements are present on opposite DNA strands within their respective promoters. It has been proposed that RE1 binds a critical repressor that acts as the molecular switch controlling expression of oat PHYA (Bruce et al., 1991). Our current results support this view and indicate, in addition, that the activity of this repressor is not limited to PHYA regulation but is also utilized in inactivating other genes that are down‐regulated by light. Indeed, we have found that the RE3 sequence is both necessary and sufficient to mediate this expression pattern. This information, together with the fact that RE sequences are present within the promoters of all light down‐regulated genes, infers that the repressor is well conserved and that it may be critical for inactivating expression of such genes in light. The isolation and characterization of this factor or complex will clearly be important for elucidation of the light‐mediated repression mechanism.
Although we have found that the RE3 repressor requires both calcium and cGMP for activation (Table VI), it is not known whether RE elements within the promoters of other light down‐regulated genes are regulated in the same manner. The presence of a family of repressor proteins each with a particular requirement for calcium and/or cGMP and with different binding affinities for the RE sequence, controlled by sequences around the TGGG core, would allow fine tuning of individual responses in spite of the utilization of common signalling molecules. How these signalling molecules may activate the RE repressor is currently open to speculation, although they may not modify DNA binding directly, because no differences in binding of nuclear factors in response to changing Pfr levels have been detected in footprint analyses of the oat PHYA promoter (Bruce et al., 1991).
In summary, the results presented here provide a good view of a plant signal transduction pathway. We have identified both the most upstream component (PfrA) and the most downstream component (a 17 bp cis‐element) and have information about some of the signal transduction intermediates and their effective concentrations. Most significantly, it is becoming clear that different responses can be controlled via the same signalling network. In the future, as other specific gene targets are linked to specific pathways, as their activation/repression thresholds in response to calcium and/or cGMP become defined and as the influence of reciprocal control on their expression is investigated, it may become possible to interpret complex physiological responses to light in terms of the functioning of this rather simple signal transduction circuitry.
Materials and methods
The AS1 promoter was cloned using the polymerase chain reaction (PCR) from pea genomic DNA prepared as described (Pruitt and Meyerowitz, 1986). Based on the original published sequence, primers were designed for PCR in order to generate an XbaI–ScaI fragment of 559 bp, which has been shown to be sufficient to mediate light down‐regulation in transgenic tobacco (Tsai, 1991). This fragment contains 559 bp of promoter sequence upstream of the transcription start site. The fragment was cloned as a transcriptional fusion to a GUS reporter gene (Jefferson et al., 1987) containing a downstream poly(A) addition sequence from the pea RBCS3C gene (Fluhr et al., 1986) in plasmid pBluescript IISK. Other reporter gene constructs, CAB–GUS, FNR–GUS and CHS–GUS, have been previously described (Bowler et al., 1994a).
RE3 was made by annealing the following two sets of oligonucleotides: 5′‐GATCTGGTGGGAGCTAG‐3′ and 5′‐GATCCTAGCTCCCACCA‐3′. RE3m was made with 5′‐GATCTGGACCGAGCTAG‐3′ and 5′‐GATCCTAGCTCGGTCCA‐3′. Each of the two sets of oligonucleotides were ligated and the 68 bp fragments (tetramers) were cloned into the BamHI site of pBluescript IISK. To construct 35S–RE3–GUS and 35S–RE3m–GUS, the RE3 and RE3m tetramer fragments respectively were inserted between the CaMV 35S B domain (−343 to −90) and the −46 minimal promoter. The synthetic promoters were then fused upstream of a GUS reporter gene and the pea RBCS3C poly(A) sequence (Benfey and Chua, 1990).
Seven‐ to 10‐day‐old etiolated seedlings of wild‐type tomato (Lycopersicon esculentum cv. Moneymaker) and the long hypocotyl mutant aurea (W616 genotype au/au) were used in all experiments. Techniques for the microinjection and subsequent analysis of aurea subepidermal hypocotyl cells have been described (Neuhaus et al., 1993; Bowler et al., 1994a) and essentially the same protocols were followed for injection and analysis of wild‐type seedlings. Preparation and handling of injection solutions were performed as described previously (Neuhaus et al., 1993; Bowler et al., 1994a), as was the treatment of injected seedlings with pharmacological inhibitors (Bowler et al., 1994b). Purified oat PHYA was stored in the dark as the PrA form. Hence, PHYA injections in green safelight conditions introduced PrA into the cells, whereas injections under normal (i.e. white light) conditions introduced PfrA.
Plasmids for microinjection were prepared using Qiagen and were stored in injection buffer at concentrations between 0.2 and 1 μg/μl (Neuhaus et al., 1993, 1994). Both reporter gene and competitor plasmids were injected in the circular form. For competition experiments, the target and competitor DNA were mixed immediately prior to injection.
Due to a technical refinement we now routinely use micropipettes with an aperture diameter of 0.3–0.5 μm (calculated as described; Schnorf et al., 1994). The estimated volume delivered during each injection is 1 pl and the estimated cell volume is 160 pl (calculated as described previously; Neuhaus et al., 1993). Hence, pipette concentrations of reagents are 160 times higher than those shown in the tables. This differs from our initial microinjection protocol (Neuhaus et al., 1993), in which we estimated an injection volume of 5 pl.
Phytochrome photo‐reversibility experiments
For red/far‐red experiments, all procedures for the preparation and subsequent injection of etiolated seedlings were performed under green safelight conditions (0.25 μE/m2/s) [type TL40W/17 (Phillips) with plexiglass filters PG303/3 mm and PG627/3 mm]. Injected seedlings were then either irradiated with red light (1.18 μE/m2/s) [type TL40W/15 (Phillips) with a red plexiglass filter (PG501/3 mm)] for 1 min or were irradiated for 1 min with red light (1.18 μE/m2/s) followed by 10 min of far‐red light (0.2 μE/m2/s) [120W Linestra lamp (Osram) with a plexiglass combination of one layer red (PG501/3 mm) and two layers blue (PG627/3 mm)]. Subsequently, the seedlings were returned to darkness for 48 h before analysis.
We thank Yan Wu and Alessandro Galli for their invaluable help with microinjection experiments and Klaus Apel for use of his photobiology facilities. G.N. was partially supported by the Huber Kudlich Stiftung (Switzerland). C.B. was supported in New York by a Science and Engineering Research Council (North Atlantic Treaty Organization) Postdoctoral Fellowship and later by a Fellowship from the Norman and Rosita Winston Foundation. This work was supported by a National Institutes of Health grant (GM 44640) to N.‐H.C and a Human Frontier Research Program award (RG0362/1995‐M) to G.N., C.B. and N.‐H.C.
- Copyright © 1997 European Molecular Biology Organization