INTRODUCTION

In order to regulate their development, higher plants employ a variety of photoreceptors to gather information about the light environment around them. The most extensively characterized of these photoreceptors are the phytochromes, which regulate many responses throughout the life cycle of the plant, including seed germination, stem elongation, leaf and chloroplast development, and flowering (1). The phytochromes are photoreversible chromoproteins comprised of an apoprotein of approximately 124 kDa that is covalently bound to a linear tetrapyrrole chromophore (2, 3). Assembly of the functional photoreceptor therefore requires two biosynthetic pathways, one for the apoprotein and one for the unbound chromophore, phytochromobilin (PB).¹ The apoprotein is encoded by a multigene family. There are five phytochrome genes (designated PHYA-E) in Arabidopsis, whereas in tomato there may be as many as nine to thirteen (4). In contrast to the multiple apoprotein components, it is likely that all phytochromes utilize the same chromophore, which is synthesized by a single pathway. The precursor of the bound phytochrome chromophore, 3E-PB (5), is synthesized in the plastid from 5-aminolevulinic acid (ALA) via the heme branch of the tetrapyrrole biosynthetic pathway (2, 6, 7). As shown in Fig. 1, the first committed step is the synthesis of biliverdin (BV) IX from heme (2, 8). By analogy with the biosynthesis of phycobilins in red algae, it has been proposed that the enzyme that accomplishes this reaction is a heme oxygenase (2, 9). BV IX is subsequently reduced to 3Z-PB by the enzyme PB synthase (10, 11). The final step in the pathway is the isomerization of 3Z-PB to 3E-PB by a putative PB isomerase (11).

Since there are two pathways required for the synthesis of phytochrome it has been possible to obtain two classes of phytochrome-deficient mutants. The isolation of mutants deficient in specific phytochromes such as phytochrome A (phyA) and phytochrome B (phyB) has enabled discrete functions to be assigned to these phytochrome species (3, 12). The second class of mutants is defined by the loss of multiple phytochrome responses including those mediated by phyA and phyB and is characterized by an elongated, yellow-green phenotype. The Arabidopsis mutants hy1 and hy2 fall into this class (13, 14) and have been shown to be deficient in phytochrome chromophore synthesis by rescue of the wild type (WT) phenotype following growth on BV IX (15). Other examples of this class of mutant include the pew mutants of Nicotiana plumbaginifolia (16) and the pcd1 mutant of pea, which was recently shown to be deficient in the conversion of heme to BV IX (8).

The aurea (au) and yellow-green-2 (yg-2) mutants of tomato are both elongated, with reduced anthocyanin and a reduced chlorophyll level, which results in pale-green or yellow looking plants (17). These distinctive features have led to both au and yg-2 being widely used as genetic markers for chromosomes 1 (18) and 12 (19), respectively. Since it was established as a photomorphogenic mutant, a number of studies have focused on the nature of the lesions in au mutants (17, 20, 21, 22, 23). Koornneef et al. (17) initially demonstrated that etiolated au seedlings contained no spectrophotometrically detectable phytochrome. It was subsequently shown that etiolated seedlings had no (20) or little (22) immunodetectable PHYA apoprotein, and this led to the hypothesis that au was specifically deficient in phyA. This hypothesis was supported by the observations that light grown au plants contained substantial amounts of photoreversible phytochrome (24, 25), including phyB (22), and that many phytochrome responses including the phyB-mediated end-of-day far-red (FR) response (24, 25) and the shade-avoidance response (26, 27) were retained in mature au plants.

The recent mapping of PHYA to chromosome 10 has demonstrated that neither au nor yg-2 are mutations in PHYA.² Moreover, a phyA-deficient tomato mutant, fri, which maps to chromosome 10 has a phenotype that is clearly distinguishable from that of both au and yg-2 (28), confirming that the au and yg-2 phenotypes are not simply a consequence of a deficiency in phyA alone. An alternative hypothesis to explain the au and yg-2 phenotypes is that they are deficient in phytochrome chromophore synthesis. This hypothesis is supported by a considerable amount of indirect evidence. The phenotype of au and yg-2 is similar to that of the chromophore-deficient, hy1 and hy2 mutants of Arabidopsis (13, 14) and pcd1 of pea (8). Etiolated seedlings of the phyA-deficient fri mutant contain a readily detectable level of stable, photoreversible phytochrome that is not present in au (17, 23, 28), suggesting that at the seedling stage au is also deficient in phytochromes other than phyA. Finally, it has recently been demonstrated that oat PHYA3 overexpressed in an au background accumulates as apoprotein and does not receive a chromophore (23). However, attempts to recover a WT phenotype by growing au on the chromophore precursor BV IX have failed (discussed in 23) despite the success of this procedure in Arabidopsis (15).

The au mutant has been widely used as a tool to investigate phytochrome-mediated responses at both the physiological (24, 25, 26, 27, 29, 30, 31) and biochemical (32, 33, 34, 35) level. Most recently, it has been used to great advantage to study the signal transduction of phytochrome responses using microinjection techniques (36, 37). It is therefore clearly of great importance that questions concerning the biochemical basis of this mutation are resolved as this knowledge will greatly aid the interpretation of such data. We have addressed this problem by measuring the committed steps of phytochrome chromophore synthesis directly in the au and yg-2 mutants of tomato. Such an approach has been made possible by the recent availability of HPLC-based assays for these steps (8, 11). We report here that au and yg-2 are specifically deficient in the conversion of BV IX to 3Z-PB and heme to BV IX, respectively.

EXPERIMENTAL PROCEDURES

The aurea (au) mutant of tomato (Lycopersicon esculentum Mill.) used in this study was isolated by Lesley and Lesley (38) and later bred into cv. Ailsa Craig to make a near isogenic line in this genetic background (39). The yellow-green-2 (yg-2) mutant was originally isolated as an irradiation-induced mutation in Lycopersicon pimpinellifolium (17, 40). It was subsequently transferred into L. esculentum, but the precise genetic background is unknown. The WT plants used for these experiments were Ailsa Craig.

Seeds of both WT and mutants were pre-treated with 1% (v/v) bleach and sown on 0.6% (w/v) agar containing 0.46 g/liter of Murashige-Skoog salts (Life Technologies, Inc.) in plant tissue culture containers (Flow laboratories, McLean, VA). Seedlings were grown for 5 days at 25 °C in complete darkness. For all experiments, the upper half of the hypocotyl and the cotyledons were harvested following the removal of any remaining seed coats. These procedures were performed under dim green safelight.

BV IX was obtained from Porphyrin Products, Inc. (Logan, UT) and further purified by C18 reverse phase HPLC as described previously (8). Heme was prepared by dissolving hemin chloride (Sigma) in 0.1  NaOH and adjusting to pH 7.7 with 1  HCl. 3E-PB and 3Z-PB were purified by HPLC following conversion of BV IX to PB by isolated pea etioplasts (8) and stored dry at 80 °C. Aliquots were dissolved in HPLC solvent immediately prior to injection. The following molar absorption coefficients were used for bilin quantitation: 66,200 ¹ cm¹ at 377 nm for BV IX (41) and 64,570 ¹ cm¹ at 386 nm and 38,020 ¹ cm¹ at 382 nm for 3E-PB and 3Z-PB, respectively (42).

Etioplasts were isolated from tomato seedlings by differential centrifugation using the method of Terry and Lagarias (10) with minor modifications. The homogenization medium used was 20 m TES/10 m Hepes/NaOH buffer, pH 7.9, containing 500 m sorbitol, 0.5% (w/v) PVP, 1 m MgCl₂, 1 m EDTA (free acid), 1 m EDTA (sodium salt), and 142 m 2-mercaptoethanol, 5 m cysteine and 0.2% (w/v) bovine serum albumin, which were added immediately prior to use. Following purification, the final plastid pellet was washed once with assay buffer stock (20 m TES/10 m Hepes/NaOH buffer, pH 7.7, containing 500 m sorbitol) before use. For the experiments measuring protoporphyrin IX synthesis from ALA, 2-mercaptoethanol was omitted from the homogenization medium, and the pH was 7.7.

Assay for PB Synthesis

PB synthesis from heme and BV IX was assayed in isolated tomato etioplasts as described previously (8, 11), but with the following minor modifications. In the PB synthesis assays, dithiothreitol was replaced by 5 m sodium ascorbate, and HPLC analysis of bilins was performed using the mobile phase ethanol/acetone/100 m formic acid (25:65:10, v/v/v) with a flow rate of 1.5 ml/min. The concentration of plastid protein used was between 0.91 and 1.59 mg/ml for assays with BV IX and 0.44-0.55 mg/ml for assays with heme.

Etioplasts were resuspended in assay buffer (20 m TES/10 m Hepes/NaOH buffer, pH 7.7, containing 500 m sorbitol, 1 m dithiothreitol, and 0.1% (w/v) bovine serum albumin) and subjected to 10 photographic flashes separated by 15-s intervals to deplete endogenous protochlorophyllide. The reaction mixture contained 1 m EDTA (sodium salt) in a final volume of 1 ml, and the reaction was initiated by the addition of 10 m ALA. The reaction mixtures were incubated at 28 °C with shaking (50/min) and kept under dim green safelight. Reactions were stopped by the addition of 3 ml of cold acetone/0.1 NH₄OH (90:10, v/v). Following a 10-min incubation, the mixture was centrifuged at 30,000 × g for 10 min. The supernatant was then washed with an equal volume of hexane to remove esterified tetrapyrroles and analyzed using a Hitachi fluorescence spectrophotometer F-3010 (Hitachi Ltd., Tokyo, Japan). Fluorescence from protoporphyrin IX was measured by the emission peak at 632 nm following excitation at 410 nm. The amount of protoporphyrin IX synthesized was determined by subtracting the fluorescence from the 0 h sample and calculating the concentration of protoporphyrin IX in the sample from a standard curve constructed using protoporphyrin IX (Sigma) in the same solvent. The protoporphyrin IX concentration in stock solutions was calculated using a molar absorption coefficient of 158,000 ¹ cm¹ at 404 nm for the dimethyl ester in ether (43). No magnesium porphyrins were synthesized under these assay conditions.

Total noncovalently bound heme was quantitated from 3 g of fresh weight etiolated seedlings as described previously (8).

Protein was determined by the method of Bradford (44) using Bio-Rad protein assay reagent and bovine serum albumin as a standard.

RESULTS

The immediate precursor of the committed steps of phytochrome chromophore synthesis is heme (see Fig. 1). Heme is a vital component of the electron transport chains required for both photosynthesis and respiration. Because au and yg-2 plants are relatively healthy and are not impaired in photosynthetic efficiency (25, 35, 45), they are unlikely to be deficient in heme. In order to test this hypothesis directly we measured total noncovalently bound heme in 5-day-old etiolated seedlings as an estimate of the total heme level in these mutants. WT seedlings contained 6.11 ± 0.23 nmol/g fresh weight (mean ± S.E., n = 5) while au and yg-2 contained 6.10 ± 0.44 and 6.67 ± 0.36 nmol/g fresh weight (n = 3), respectively. These results demonstrate that there is no significant difference in the level of noncovalently bound heme in WT, au, and yg-2 seedlings. Therefore a deficiency in phytochrome chromophore synthesis in the au and yg-2 mutants is predicted to be the result of a block between heme and PB, and we examined these steps individually.

Etioplasts from au Seedlings Are Unable to Convert BV IX to 3Z-PB

Firstly we addressed the question of whether au and yg-2 are deficient in PB synthase activity. Isolated etioplasts were incubated with BV IX and an NADPH regenerating system, and the products were analyzed by reverse phase HPLC (Fig. 2). WT etioplasts converted BV IX to both 3Z-PB and 3E-PB (trace c, Fig. 2), which were identified by co-injection with purified PB samples (data not shown). These products are identical to those detected previously following the incubation of BV IX with etioplasts from both oat (11) and pea (8) seedlings. No PB was synthesized following incubation of BV IX in the absence of etioplasts or with WT, au, or yg-2 etioplasts in the absence of BV IX (traces a and b, Fig. 2; data not shown). Incubation of etioplasts isolated from yg-2 seedlings also resulted in the synthesis of 3Z-PB and 3E-PB (trace d, Fig. 2). The identity of both products was again confirmed by co-injection with purified PB samples (data not shown). When BV IX was incubated with au etioplasts no PB products were observed (trace e, Fig. 2). To verify that the peak with a retention time of approximately 12 min was not 3Z-PB, co-injection experiments with authentic, purified 3Z-PB were performed (Fig. 3). The peak at 12 min elutes as a shoulder of the 3Z-PB peak, demonstrating clearly that it is not 3Z-PB and that no 3Z-PB accumulates following incubation of au etioplasts with BV IX (Fig. 3). Experiments using an additional au allele, au^w (17, 46), resulted in identical results (data not shown). The inability of au to synthesize 3Z-PB from BV IX is therefore most likely to be a direct consequence of the au mutation.

HPLC analysis of PB synthesis from BV IX by isolated plastids from WT,

-PB with the products of incubating

plastids with BV IX.

Although the data presented suggest that the au mutation results in the loss of the enzyme activity that converts BV IX to 3Z-PB, there is a formal possibility that the mutation leads to the degradation of PB, thereby preventing its accumulation. The au mutation cannot reside in a putative catabolic enzyme itself because the mutation is recessive. However, a mutation in a repressor of such an activity would be expected to result in the increased degradation of PB or bilins in general. To test this possibility we analyzed PB synthesis by WT etioplasts in the presence of au etioplasts. Co-incubation of au etioplasts with WT etioplasts did not prevent the accumulation of either 3Z-PB or 3E-PB (data not shown). In addition, analysis of BV IX recoveries during the course of these experiments demonstrated that au etioplasts did not result in increased BV IX degradation compared with WT or yg-2 etioplasts (data not shown), indicating there is no increase in nonspecific bilin catabolism in au. These results confirm that the absence of 3Z-PB and 3E-PB following the incubation of BV IX with au etioplasts is a consequence of reduced PB synthesis and not increased degradation.

Etioplasts from yg-2 Seedlings Are Unable to Convert Heme to BV IX

To test whether au and yg-2 could convert heme to BV IX, isolated etioplasts were incubated with heme, and the bilin products were analyzed by reverse phase HPLC. Incubation of etioplasts isolated from WT seedlings with heme in the presence of an NADPH regenerating system resulted in a primary product with a retention time identical to BV IX on this HPLC system (trace b, Fig. 4). Since the major product following incubation of pea etioplasts with heme is 3Z-PB (8), we examined the identity of this product further by co-injection of purified 3Z-PB and BV IX. As shown in Fig. 5, co-injection of 3Z-PB resulted in two major peaks (trace b, Fig. 5), whereas co-injection of BV IX resulted in a single major peak (trace c, Fig. 5). Therefore, in contrast to the situation in pea, the primary product from incubating tomato etioplasts with heme is BV IX. Such a result may reflect the relative activities of the heme oxygenase and PB synthase enzymes in the two species. Alternatively, it may be due to differences in PB stability in extracts of pea and tomato.

Identification as BV IX as the major product following incubation of heme with WT plastids.

Incubation of heme with etioplasts isolated from au seedlings also resulted in the synthesis of BV IX (trace c, Fig. 4), which was again identified by co-injection studies (data not shown). Identical experiments using au^w demonstrated that conversion of heme to BV IX was also unaffected by this allele (data not shown). When etioplasts isolated from yg-2 seedlings were incubated with heme, a small quantity of BV IX (identified by co-injection of purified BV IX; data not shown) was detected (see peak at approximately 12 min; trace d, Fig. 4). The yield of BV IX from the yg-2 sample was similar to that obtained from the coupled oxidation of heme in the control assay (trace a, Fig. 4). The yg-2 mutant is therefore deficient in the synthesis of BV IX from heme and is analogous to the pcd1 mutant of pea (8).

Etioplasts from au and yg-2 Seedlings Are Specifically Impaired in PB Synthesis

Although the data shown here clearly demonstrate that au and yg-2 are deficient in PB synthesis, it has been proposed that such deficiencies might arise as a consequence of impaired etioplast development. Hypocotyl cells from dark grown au seedlings do not have fully developed etioplasts but instead contain proplastid-like organelles (36). To test whether defects in plastid development might result in a general inhibition of tetrapyrrole biosynthesis in isolated etioplasts, we measured the synthesis of protoporphyrin IX from ALA. Six enzymes are required for the conversion of ALA to protoporphyrin IX, and these are thought to be membrane-associated (47). It might therefore be expected that pleiotropic effects of the mutations on plastid structure would lead to a reduction in the ability of au and yg-2 to synthesize protoporphyrin IX from ALA. However, as shown in Fig. 6, there was no effect of the au and yg-2 mutations on protoporphyrin IX synthesis in isolated etioplasts. This result, together with the demonstration that au and yg-2 are each deficient in a single step between heme and 3Z-PB, shows that these mutants are specifically impaired in the conversion of BV IX to 3Z-PB and heme to BV IX, respectively.

DISCUSSION

In this paper we have examined the hypothesis that au and yg-2 are mutations that affect the biosynthetic pathway of the phytochrome chromophore by measuring the steps in this pathway directly. Since au and yg-2 were not heme-deficient, we focused on the steps committed to the synthesis of PB. Analysis of these biosynthetic steps demonstrated that the yg-2 mutant was unable to convert heme to BV IX (Fig. 4), while the synthesis of 3E- and 3Z-PB from BV IX was normal (Fig. 2). The yg-2 mutant is therefore blocked at the step that has been proposed to be accomplished by heme oxygenase and is analogous to the pcd1 mutant of pea (8). Conversely, au is able to synthesize BV IX from heme (Fig. 4) but cannot synthesize 3Z-PB from BV IX, the reaction accomplished by PB synthase (Figs. 2 and 3). Two independent alleles were shown to lack the ability to convert BV IX to 3Z-PB, confirming that this deficiency is the biochemical basis of the au phenotype. Whether au is also deficient in the proposed PB isomerase activity is unknown. In addition, the additive phenotype of the au,yg-2 double mutant (23) is consistent with the identification of au and yg-2 as mutations affecting sequential steps in this pathway.

We believe that the deficiencies in phytochrome chromophore synthesis in au and yg-2 are most likely to be the result of mutations in the biosynthetic enzymes themselves and are not due to pleiotropic effects on plastid structure and function. Such a conclusion is based on the observation that au and yg-2 are only deficient in a single step of the pathway committed to PB synthesis and that etioplasts from the mutants are equally capable of synthesizing protoporphyrin IX from ALA as WT etioplasts (Fig. 6). Six enzyme steps are required for protoporphyrin IX synthesis from ALA and such an assay represents a good measure of the integrity of the isolated etioplasts.

The deficiency in phytochrome chromophore biosynthesis can account for the phenotype of au and yg-2 seedlings, since at this developmental stage they would be expected to be reduced in all holophytochrome species. Consistent with this conclusion, both au and yg-2 lack responses that are attributable to multiple phytochrome species (28, 48). For example, hypocotyls of au and yg-2 are elongated in both R and FR (17), responses that are specifically deficient in mutants lacking phyA (28) and a phyB (48), respectively. Similarly, anthocyanin synthesis in response to R and FR is absent in au (17, 49). There is also a well characterized role for phytochrome in the synthesis of both chlorophyll (33) and chlorophyll-binding proteins (21, 32, 34) in tomato, and phytochrome-deficient tomato seedlings would be expected to display the pale, yellow-green phenotype of au seedlings. Interestingly, both phyA- and phyB-deficient plants are not chlorophyll-deficient at the seedling stage (28, 48). This suggests that additional phytochromes or other photoreceptors are involved in these responses, although redundancy between phytochromes, as noted for chlorophyll synthesis in Arabidopsis (50), cannot be ruled out.

The phenotype of au and yg-2 seedlings is also consistent with the phenotypes observed in seedlings of other phytochrome chromophore-deficient mutants such as the hy1 and hy2 mutants of Arabidopsis (13, 14) and the pcd1 mutant of pea (8). One exception is that in both au and yg-2, PHYA accumulates at best to 20% of the WT level (22, 23), whereas PHYA levels are close to WT in pcd1 (8) and hy1 and hy2 (14, 51). This observation was one reason that au was considered to be specifically deficient in phyA. However, this PHYA instability under conditions of chromophore deficiency appears to be specific for the tomato protein since oat PHYA overexpressed in an au or yg-2 background accumulates to levels equivalent to those detected in WT seedlings overexpressing this protein (23). A second misleading observation, the reported inability to rescue a WT phenotype following growth of au seedlings on BV IX (23), is not surprising in the light of the data presented here. However, attempts to recover au and yg-2 on phycocyanobilin, an analog of PB, and yg-2 on BV IX were also unsuccessful.³ A similar observation was noted with the pcd1 mutant of pea (8), and Arabidopsis appears to be the only species tested in which chromophore-deficient mutants can be fully rescued (in terms of elongation growth) by chromophore precursors (15).

The phenotype of mature au plants with respect to elongation growth is considerably less exaggerated than the seedling phenotype. The observation that the end-of-day FR response (25) and shade-avoidance responses are normal (26, 27) or nearly normal (31) in mature au plants is consistent with the presence of more than half the holophytochrome of WT plants (25) including photoactive phyB (22). The au (and yg-2) mutants are therefore leaky and can synthesize phytochrome chromophore in more mature tissues. In this respect they are similar to the pcd1 mutant of pea, which responds progressively to end-of-day FR with increasing developmental age (8). The ability of these mutants to synthesize some chromophore may be the result of the mutant enzymes retaining some biosynthetic activity. Alternatively, the increased biosynthetic activity may be the result of additional genes encoding these enzymes, which may be more highly expressed at later stages of plant development. One phenotypic characteristic that remains pronounced in mature au and yg-2 plants is the yellow-green foliage. Why such a chlorophyll deficiency remains so pronounced when other phytochrome responses are present is not clear. One possibility is that the lesions in the chromophore pathway lead to an accumulation of free heme, which has been proposed to act as an inhibitor of tetrapyrrole biosynthesis by inhibiting the rate-limiting step of ALA synthesis (52). This hypothesis is currently under investigation but would help explain not only the pale phenotypes of mature au and yg-2 plants but also the observation that photoreversible protochlorophyllide is reduced in etiolated au seedlings (33) and the altered development of au etioplasts (36).

The results in this paper have a number of implications for both the extensive published data using au and also for future studies with au and yg-2. While its use as a phytochrome-deficient mutant is completely justified, it should be borne in mind that the deficiency will not be restricted to the phyA pool but will most likely affect all phytochrome species to some extent. Conversely, such mutants are unlikely to be null for phyA or any other phytochrome. In addition, in experiments where chlorophyll synthesis or chloroplast function are being assessed (32, 33, 34, 36, 37) consideration should be given to the possibility that the accumulation of intermediates in the phytochrome chromophore pathway, such as heme, may affect the parameters being studied.