One Divinyl Reductase Reduces the 8-Vinyl Groups in Various Intermediates of Chlorophyll Biosynthesis in a Given Higher Plant Species, but the Isozyme Differs between Species 1,

Divinyl reductase (DVR) converts 8-vinyl groups on various chlorophyll intermediates to ethyl groups, which is indispensable for chlorophyll biosynthesis. Up to date, five DVR activities have been detected, but adequate evidence of enzymatic assays using purfied or recombinant DVR proteins have not been demonstrated, and it is unclear whether one or multiple enzymes catalyze these activities. In this study, we systematically carried out enzymatic assays using four recombinant DVR proteins and five divinyl substrates, and then investigated the in vivo accumulation of various Chl intermediates in rice ( Oryza sativa ), maize ( Zea mays ) and cucumber ( Cucumis sativus ). The results demonstrated that both rice and maize DVR proteins can convert all of the five divinyl-substrates to corresponding monovinyl-compounds, while both cucumber and Arabidopsis thaliana DVR proteins can convert three of them. Meanwhile, the OsDVR ( Os03g22780 )-inactivated 824ys mutant of rice exclusively accumulated divinyl chlorophylls in its various organs during different developmental stages. Collectively, we concluded that a single DVR with broad substrate specificity is responsible for reducing the 8-vinyl groups of various chlorophyll intermediates in higher plants, but DVR proteins from different species have diverse and differing substrate preferences although they are homologous.


INTRODUCTION
After being grown in the dark for 6 d, the etiolated rice and maize seedlings almost entirely accumulated MV-Pchlide a ( Figure 6A and 6B). Conversely, the etiolated cucumber cotyledons accumulated only about 68% of MV-Pchlide a, and near one third of DV-Pchlide a remained to be converted ( Figure 6C). These data suggested that rice, maize and cucumber can convert DV-Pchlide a and/or -MPE and/or DV Mg-Proto to corresponding MV-compounds in vivo, though the conversion efficacy in the three plants could be different. Next, we examined the re-accumulation of Pchlide.
Etiolated cucumber cotyledons, rice and maize seedlings were illuminated for 14 h, and then returned to the dark for 0 min, 10 min, 2 h and 10 h, respectively. As showed in figure 7, at the beginning of the dark period (0 min), rice, maize and cucumber only accumulated a very small amount of Pchlide a, and their Pchlide a pools consisted almost entirely of DV-Pchlide a ( Figures 7A1 to 7A3). After 10 min of dark incubation, Pchlide a pools consisted of about 26% MV-Pchlide a in rice and maize (Figures 7B1 and 7B2). After 2 h of dark incubation, the amount of MV-Pchlide a in rice and maize accumulated higher than that of DV-Pchlide a (Figures 7C1 and 7C2). By the end of 10 h dark incubation, the Pchlide pools consisted almost completely of MV-Pchlide a in the etiolated rice and maize seedlings (Figures 7D1 and 7D2). Contrarily to these observation, MV-Pchlide a was still undetectable in cucumber after 2 h of dark incubation (Figures 7B3 and 7C3). However, by the end of 10 h dark incubation, considerable amount (about 40%) of MV-Pchlide a was detected in the etiolated cucumber cotyledons ( Figure 7D3). These results once again indicated that DV-Pchlide a and/or -MPE and/or DV Mg-Proto can be converted to corresponding MV-compounds in rice, maize and cucumber in vivo, but the conversion velocities in rice and maize are much faster than that in cucumber. These observations were consistent with our in vitro data that showed the efficient conversions of DV-Pchlide a and/or -MPE and/or DV Mg-Proto to corresponding MV-compounds by DVRs from rice and maize, while extremely slow by DVR from cucumber.
In addition, we investigated the accumulation of protoporphyrin IX (Proto), Mg-Proto and MPE.
Picolinic acid (PA) was shown to induce accumulation of these intermediates in plants (Mayasich et al., DV-MPE and/or DV Mg-Proto can be converted to corresponding MV-compounds in these plants in vivo. Moreover, a certain amount of MV Mg-Proto was also detected in the etiolated rice (peak 1' in Figure 8A), confirming that DV Mg-Proto can be converted to MV Mg-Proto in rice in vivo.

In vivo Accumulation of Chl and Its Intermediates in Rice 824ys Mutant
In the previous study, we demonstrated that the rice 824ys mutant accumulated only DV-Chls and had no MV-Chls because it had a defective DVR protein encoded by mutant gene of Os03g22780 (Wang et al., 2010). However, at that time we examined Chl compositions of the mutant only using leaf blade of four-week-old plants (at the five-leaf stage). We could not exclude the possible accumulation of MV-Chls generated by DVRs other than OsDVR in the mutant if we did not check various tissues during various developmental stages (Islam et al., 2008;Ito et al., 2008). To this end, we systematically investigated Chl compositions in leaf blade, leaf sheath, uppermost internode, panicle rachis, glume and caryopsis of the 824ys mutant at seedling, tillering, flowering, filling and maturation stages, respectively. As showed in figure 9, contrary to the wild type rice that contained MV-Chl a and MV-Chl b, the 824ys mutant exclusively accumulated DV-Chls, and indeed, did not accumulate any MV-Chls in any tissues during any developmental stages tested, including rapidly greening tissues and senescent tissues. In addition, we examined Pchlide accumulation of the etiolated 824ys mutant seedlings grown in the dark for 6 d, showing that this mutant exclusively accumulated DV-Pchlide a in the dark ( Figure 6D). Moreover, we investigated accumulation of Proto, Mg-Proto and MPE in the etiolated 824ys mutant leaves treated with PA and ALA, showing that this mutant accumulated only DV-intermediates ( Figure 8D). These observations demonstrated that the 824ys mutant can not convert any DV-intermediates of Chls to corresponding MV-compounds in vivo, strongly suggesting that the OsDVR encoded by Os03g22780 is the only protein that is responsible for reducing the 8-vinyl groups of various intermediates of chlorophyll biosynthesis in rice.
Collectively, the in vivo accumulation of Chl intermediates in rice, maize and cucumber were consistent with the results of the above in vitro enzymatic assays. These data provide robust supporting evidence that a single DVR protein with broad substrate specificity is responsible for reducing the 8-vinyl groups of various intermediate molecules of Chl biosynthesis in higher plants, but homologous DVR proteins from different species could have quite different reductive activities.
However, it is unclear whether these DVR activities are catalyzed by one enzyme with broad substrate specificity or by multiple enzymes with narrow specificity. Nagata et al. (2007) reported that AtDVR is able to convert DV-Chlide a to MV-Chlide a, but unable to convert DV-Pchlide a, -Chlide b, -Chl a and -Chl b to corresponding MV-compounds. Nevertheless, Chew and Bryant (2007) demonstrated that a recombinant DVR protein (BciA) of green sulfur bacterium Chlorobium tepidum successfully reduced the 8-vinyl group of DV-Pchlide a in vitro. In previous report, we confirmed that OsDVR is able to convert both DV-Chlide a and -Chl a to MV-Chlide a and -Chl a, respectively (Wang et al., 2010). In this study, we found that OsDVR and ZmDVR have broader substrate specificity than CsDVR and AtDVR. Both OsDVR and ZmDVR proteins can convert the five Chl biosynthetic intermediates, DV-Chl a, -Chlide a, -Pchlide a, -MPE, and DV Mg-Proto, into corresponding MV-compounds (Figures 1 to 5). Both CsDVR and AtDVR proteins can convert DV-Chlide a, -Pchlide a and -MPE, but not DV-Chl a and DV Mg-Proto into corresponding MV-compounds ( Figures 1 to 5). These data indicate that a single DVR protein has broad substrate specificity with different range.

Homologous DVR Proteins from Different Species Have Diverse and Differing Substrate
Preferences DVR proteins derived from different plant species have quite different reductive activities on the same or on different substrates. DV-Pchlide could be partially converted to MV-Pchlide in barley plastids, but not in cucumber plastids (Tripathy and Rebeiz, 1988). The velocity of converting DV-Chlide a in etioplast membranes from etiolated cucumber cotyledons, barley and maize leaves were 50 to 300-fold higher than that of converting DV-Pchlide a in barley etioplasts (Tripathy and Rebeiz, 1988;Parham and Rebeiz, 1995). Our data demonstrate that OsDVR and ZmDVR have much higher catalyzing in the dark for 6 d or 10 h (Figures 6 and 7;Carey and Rebeize, 1985). In addition, the same DVR protein also has quite different reductive activities on different substrates. Both OsDVR and ZmDVR have significantly higher efficacy in converting DV-Chlide a and -Chl a to MV-Chlide a and -Chl a, respectively, than that converting DV-Pchlide a, -MPE and DV Mg-Proto to corresponding MV-compounds (Figures 2 to 5). Both CsDVR and AtDVR also have substantially higher efficiency to reduce DV-Chlide a into MV-Chlide a than that to reduce DV-Pchlide a and -MPE into corresponding MV-compounds (Figures 2, 3A, 4 and 5A).

A Single DVR Protein Is Responsible for Reducing the 8-Vinyl Groups in Higher Plants
Chl biosynthetic heterogeneity is assumed to originate mainly in parallel DV-and MV-Chl biosynthetic routes interconnected by 8-vinyl reductases (Parham and Rebeiz, 1995;Rebeiz et al., 2003). In the past decades, 8-vinyl Mg-Proto, MPE, Pchlide a, Chlide a and Chl a reductases were all proposed as a kind of 8-vinyl reductases (Ellsworth and Hsing, 1974;Tripathy and Rebeiz, 1988;Parham andRebeiz, 1992, 1995;Kim and Rebeiz, 1996;Adra and Rebeiz, 1998;Kolossov and Rebeiz, 2010), but none of these enzymes had been successfully purified until 2001 when the enzyme was solubilized and partially purified from etiolated barley leaves (Kolossov and Rebeiz, 2001). In recent years, AtDVR (AT5G18660) and OsDVR (Os03g22780) genes were isolated from Arabidopsis and rice, respectively, and identified as homologues encoding an 8-vinyl reductase (Nagata et al., 2005(Nagata et al., , 2007Wang et al., 2010). However, it has been not fully clear whether one or multiple enzymes are responsible for the reduction of the 8-vinyl groups of various Chl intermediates. A maize mutant accumulated only DV-Chls (Bazzaz, 1981;Bazzaz and Brereton, 1982;Bazzaz et al., 1982), which implied that a single gene product might be responsible for the reduction of the 8-vinyl group of Chl intermediates, although the possibility that the gene encodes a regulator for the divinyl reduction could not be excluded. The AtDVR (AT5G18660)-inactivated pcb2 mutant of Arabidopsis also exclusively accumulated DV-Chls in 16-day-old whole plants, which suggested that Arabidopsis could contain only one functional DVR (Nakanishi et al., 2005;Beale, 2005), but other possibilities were also proposed (Islam et al., 2008;Ito et al., 2008). Previously, we reported that the OsDVR substrate specificity, though homologous DVR proteins from different species could have quite different DVR activities (Figures 1 to 5). In addition, the DVR proteins used in this study all are encoded by a single-copy gene, and no other homologous genes were found in the rice, maize, cucumber and Arabidopsis genomes. Collectively, we are confident that a single DVR protein is responsible for reducing the 8-vinyl groups of various intermediate molecules of Chl biosynthesis in higher plants, at least in angiosperms represented by rice and Arabidopsis.

Chl Biosynthetic Pathways Include Multi-branched Routes Resulted from a Single DVR Protein in Higher Plants
A multi-branched chlorophyll biosynthetic pathway was proposed due to the detection of MV-and DV-tetrapyrrole intermediates and their biosynthetic interconversion in extracts of different plant tissues (Rebeiz et al., 1983(Rebeiz et al., , 1986(Rebeiz et al., , 1999Whyte and Griffiths, 1993;Kolossov and Rebeiz, 2010). Nagata et al. (2007) suggested that the major route for chlorophyll synthesis includes DV-Chlide a reduction in Arabidopsis, but also observed that DV-Pchlide a can be converted to MV-Pchlide a in vivo although the conversion of DV-Pchlide a to MV-Pchlide a is much slower than that of DV-Chlide a to MV-Chlide a, which actually indicated the existence of multi-branched pathway in Arabidopsis. In this paper, three or two MV-Chl intermediates were simultaneously detected in the etiolated rice, maize and cucumber after ALA and PA treatment (peaks 1', 2' and 3' in Figure 8), also suggesting that multi-branched pathway do exist. In addition, the enzymatic assays demonstrated that both OsDVR and ZmDVR can convert five DV-substrates to corresponding MV-compounds, and both CsDVR and AtDVR can convert three of them (Figures 1 to 5), which match the branch points of the multi-pathways converting DV-to MV-Chl intermediates. Meanwhile, our data indicated that a single DVR is responsible for reducing 8-vinyl groups of various chlorophyll intermediates. These results strongly suggested that Chl biosynthetic pathways include multi-branched routes resulted from a single DVR in higher plants.
On the basis of MV-or DV-Pchlide accumulation during the dark and light phases of the photoperiod, green plants have been classified into three different greening groups, namely, dark divinyl -light divinyl (DDV-LDV), dark monovinyl-light divinyl (DMV-LDV) and dark monovinyl-light monovinyl (DMV-LMV) plants (Carey et al., 1985;Carey and Rebeiz, 1985;Shioi and Takamiya, 1992;Ioannides et al., 1994;Mageed et al., 1997). In the dark, the Chl synthesis pathway leads only to the formation of Pchlide in angiosperms. Once a critical level of Pchlide has reached, 5-Aminolevulinic acid (ALA) synthesis slows down (Griffiths, 1975(Griffiths, , 1978Masuda and Takamiya, 2004). Our data demonstrated that the conversions of DV-Pchlide a, -MPE and DV Mg-Proto to corresponding MV-compounds were much slower than that of DV-Chl(ide) a (Figures 1 to 7), which could be a kind of adaptation to the diurnal variation of chlorophyll biosynthesis. On the other hand, ALA-induced tetrapyrrole accumulation in green plants could cause extensive photodynamic damage to some plant species, while other plant species remained unaffected. The non-susceptible plants species maintained higher relative levels of MV-Pchlide than the susceptible ones under the subdued light levels (Carey and Rebeiz, 1985). So the DV to MV-Pchlide conversion may be considered as one means of eliminating the excess DV-Pchlide (Tripathy and Rebeiz, 1998). Therefore, the multi-branched Chl biosynthetic pathways could be physiologically important.
Natural selection has often produced multiple (bio)chemical and physical ways of conveying the same message. It is also possible that via natural selection, Chl biosynthetic heterogeneity has imparted an evolutionary advantage to higher plants. Ioannides et al. (1994) proposed that the DDV-LDV greening group is evolutionarily ancestral because so far all representative primitive plant species, including algae, bryophytes, ferns and gymnosperms, fall in this greening group. The DMV-LMV greening group comprises a small number of plant species, and evolutionary studies suggested that it is derived (Ioannides et al., 1994). The DMV-LDV greening group comprises by far the largest number of plant species so far surveyed, and plant species of major agronomic importance belong to this group, (5804R; Eppendorf) for 15 min, and the supernatants were dried under N 2 gas and redissolved with 100% acetone. DV/MV-Chl a were separated by HPLC on a C18 column (4.6 mm i.d.× 150 mm long; 5 μ m; Agilent, U.S.A.), and eluted with the solvent (methanol:acetonitrile:acetone = 1:3:1) at a flow rate of 1.0 mL min -1 at 40℃ (as described by Nakanishi et al. [2005]). Elution profiles were monitored by measuring absorbance at 660 nm.

Extraction of Pchlide a
DV-Pchlide a was prepared from six-day-old etiolated seedlings of the rice 824ys mutant, and MV-Pchlide a from that of wild-type rice 824B. MV-Pchlide a prepared from 6-day-old etiolated seedlings of wheat was used as control (Carey and Rebeiz, 1985).
Etiolated cucumber cotyledons, rice, maize and wheat seedlings were homogenized in acetone containing 0.1M NH 4 OH under a green safelight. Tris-HCl (pH 7.6) was added to the acetone solution to a final concentration of 200 mM in order to protect modification of Pchlide. Then, an equal volume of hexane was applied to the acetone solution to remove chlorohylls and carotenoid. Finally, Pchlide in in darkness at 28℃ for 14 h (Kim and Rebeize, 1996). Subsequently, the incubated tissues were homogenized in acetone : 0.1 N NH4OH (9:1, v/v) under a green safelight. The homogenate was centrifuged at 7, 830 × g (5430R; Eppendorf) at 1℃ for 30 min, and the resulting supernatant stored at -80℃ until tetrapyrrole extraction. Then, Chl and other fully esterified tetrapyrroeles were transferred from acetone to hexane by extraction with an equal volume of hexane, followed by a seconed extraction with 1/3 volume of hexane (Kim and Rebeiz, 1996). The remaining hexane-extracted acetone residue were analyzed by HPLC on a C8 column (4.6 mm i.d. × 150 mm long; 3.5 µm; Waters) according to the method of Zapata et al. (2000). Elution profiles were monitored by measuring absorbance at 410 nm.

Enzymatic Activity Assays
Full-length OsDVR, ZmDVR, CsDVR and AtDVR genomic DNA were amplified by PCR from rice, maize, cucumber and Arabidopsis genomic DNA, using primer OS (F: 5'-CAGGATCCATGGCTGCCCTCCTCCTCT-3', R:    coli, and the pigments extracted from reaction mixtures were subjected to HPLC, respectively. The left column is the chromatogram that was detected at 660 nm by HPLC, and the right column depicts the spectra of each peak. (A1) Product synthesized after incubation with E. coli lysates expressing the empty vector, which was used as negative control. Rice (wild-type) (A), maize (B), cucumber (C) and rice 824ys mutant (D) were grown in the dark at 28℃ for 5 d. Subsequently, the etiolated cucumber cotyledons, rice and maize leaves were incubated with 40 mM ALA and 30 mM PA in darkness at 28℃ for 14 h. Then, tetrapyrroles were extracted and subjected to HPLC. The left column is the chromatogram that was detected at 410 nm by HPLC, and the right column depicts the spectra of each peak.