Incomplete Hydrogenation by Geranylgeranyl Reductase from a Proteobacterial Phototroph Halorhodospira halochloris, Resulting in the Production of Bacteriochlorophyll with a Tetrahydrogeranylgeranyl Tail

Chlorophyll (Chl) and bacteriochlorophyll (BChl) pigments are critical in photosynthetic organisms for harvesting light energy and transferring it to photochemical reaction center (RC) complexes, where the charge separation takes place. Chl occurs in all oxygenic phototrophs (including plants, algae, and cyanobacteria) and some species of anoxygenic phototrophic bacteria with type-I RCs (1–3). In the green sulfur bacterium Chlorobaculum tepidum, Chl is attached to type-I RCs and functions as the primary electron acceptor A₀, although BChls are major pigments in the bacterium (2). BChl pigments are detected in all species of anoxygenic phototrophic bacteria, regardless of whether they have type-I or type-II RCs, but not in oxygenic phototrophs.

Photosynthetic organisms biosynthesize pigments through a series of catalytic reactions by various enzymes (4, 5). Chl and BChl share early biosynthetic steps, from an initial substance 5-aminolevulinic acid to chlorophyllide a or divinyl chlorophyllide a (4–7). The committed biosynthetic step for BChl a is branched at the chlorophyllide a reduction stage (6), whereas the committed step for BChl b and BChl g is branched at the 8-vinyl chlorophyllide a (7–9). These two steps correspond to the conversion of a chlorin ring into a bacteriochlorin ring.

Although Chl and BChl species possess specific core π-skeletons and peripheral substituents, the penultimate and last biosynthetic steps for all the photosynthetic (B)Chl pigments except Chl c are common, where the order of the two steps can be switched. The esterification of geranylgeranyl diphosphate into the substituent at the carbon 17 (C17) position of (bacterio)chlorophyllide (10–12) is the penultimate step. The esterification is catalyzed by an enzyme designated as (B)Chl synthase encoded by the bchG/chlG gene. The last step is the hydrogenation of the C17 geranylgeranyl to the phytyl tail, which is catalyzed by bchP/chlP gene-encoded geranylgeranyl reductase (GGR) (10, 13). As depicted in Fig. 1, the geranylgeranyl tail is reduced three times to a phytyl group. The triple double-bond reduction of geranylgeranyl group by GGR occurred in the order of C10=C11, C6=C7, and C14=15 (Fig. 1, left column) (14).

Halorhodospira halochloris, a halophilic anoxygenic phototrophic bacterium belonging to the phylum Gammaproteobacteria, produces BChl b esterified with a unique isoprenoid tail at the C17 position, namely, a tetrahydrogeranylgeranyl (THGG) tail (Fig. 1, middle column) (15, 16). The THGG tail in H. halochloris is characterized as the C10=C11 unreduced double bond (thereby also called 2,10-phytadienyl) (15, 16), indicating that the first double-bond reduction of the triple hydrogenation occurring in other phototrophic bacteria is likely skipped or inhibited in vivo in H. halochloris (Fig. 1, middle). Therefore, the following two hypotheses emerge: (i) GGR in H. halochloris only reduces the isoprenoid tail twice, or (ii) GGR in H. halochloris potentially reduces three times, as observed in other phototrophic bacteria, but is prevented from reducing the C10=C11 double bond by an unidentified component(s). The latter case has been proposed for Chl biosynthesis in green sulfur bacteria (17, 18).

Green sulfur bacteria produce BChl a with the usual phytyl tail, but they also produce Chl a esterified with a unique THGG group (Fig. 1, right column) (2, 19). The hydrophobic THGG tail of Chl a in C. tepidum displays the C6=C7 unreduced double bond (2), which is different from the C10=C11 unreduced double bond in H. halochloris (Fig. 1). In the case of C. tepidum, a mutant lacking GGR accumulated BChl a and Chl a with the geranylgeranyl tails at the C17 position (17, 18), and it was concluded that a single bchP gene, CT2256, is responsible for saturating the geranylgeranyl tails esterified with both BChl a and Chl a in C. tepidum. Therefore, GGR (gene product of CT2256) of C. tepidum potentially has a catalytic ability of three reductions of the geranylgeranyl moiety, yielding the phytyl tail; however, this ability is somehow inhibited or unachieved in Chl biosynthesis but not in BChl biosynthesis. This C. tepidum GGR model correlates with the aforementioned second hypothesis. Recently, through an analysis of whole-genome sequencing of H. halochloris, a bchP gene was identified in the photosynthetic gene cluster of its genome (20). In this study, we investigated the catalytic activities of the H. halochloris GGR by creating a series of complementation mutants.

We first constructed the ΔbchP mutant of R. sphaeroides lacking GGR (Fig. 2) (for details, see Materials and Methods). The ΔbchP mutant strain served as a host to construct complementation mutants containing wild-type and variant GGRs of H. halochloris and R. sphaeroides. We examined the pigment compositions of these mutant strains using high-performance liquid chromatography (HPLC) (Fig. 3). The HPLC elution profile of pigment extracts from the wild type of R. sphaeroides revealed an authentic phytylated BChl a, which was eluted at roughly 20.5 min (Fig. 3, profile 1). The ΔbchP mutant of R. sphaeroides lacking GGR did not exhibit the phytylated BChl a profile at 20.5 min; instead, the mutant accumulated geranylgeranylated BChl a, which was eluted at roughly 14.5 min (Fig. 3, profile 2). Profile 3 in Fig. 3 depicts the HPLC elution profile of pigment extracts from Rhodopseudomonas sp. strain Rits, which was discovered to accumulate BChl a molecules with unreduced and partially-reduced isoprenoid tails (14, 21). Rhodopseudomonas sp. strain Rits exhibited four elution peaks at 14.5, 16, 18, and 20.5 min, which were attributed to BChl a esterified with GG, dihydrogeranylgeranyl (DHGG), THGG, and phytyl tails, respectively, according to the previous study (21) (Fig. 3, profile 3). The HPLC elution profile of pigments extracted from the Hh_P_wt mutant, which has an intact GGR of H. halochloris in the background of the R. sphaeroides ΔbchP strain, revealed three peaks of BChl a esterified with GG, DHGG, and THGG tails (Fig. 3, profile 4), but no phytylated BChl a. The result indicates that the GGR of H. halochloris catalyzes only two double-bonds hydrogenation, and thereby the final product has the THGG tail, rather than the phytyl tail. Although both organisms produce (B)Chl pigments with THGG tails, the potential catalytic activity of H. halochloris GGR differs from that of C. tepidum GGR (17, 18). C. tepidum produces Chl a with the THGG tail, but Harada et al. showed that a mutant complemented with C. tepidum GGR produced phytylated BChl a, indicating that C. tepidum GGR can reduce a GG tail attached to a bacteriochlorin ring to a phytyl tail (17).

Figure 4 depicts the N-terminus of the amino acid sequence alignment of GGR from several phototrophic bacteria and nonphotosynthetic archaea. H. halochloris GGR has a characteristic insertion at the N-terminal side of its primary sequence (Fig. 4, colored in red). We constructed the ΔbchP mutant of R. sphaeroides complemented with the variant GGR of H. halochloris that lacks the insertion region, designated Hh_P_del, because the insertion region could be relevant to presumably inhibiting the unachieved double-bond reduction at the C10=C11 position. The HPLC elution profile of pigments from the Hh_P_del mutant exhibited the GG peak eluting at roughly 14.5 min and the DHGG peak at around 16 min; however, the mutant did not accumulate peaks derived from BChl a with THGG and phytyl tails (Fig. 3, profile 5). The result indicates that the loss of the insertion region of H. halochloris GGR caused the loss of either the first or last hydrogenation and that the variant GGR catalyzes only a single hydrogenation reaction, yielding the DHGG tail.

We constructed a positive-control complementation mutation by introducing the intact bchP gene for R. sphaeroides GGR in the ΔbchP mutant of R. sphaeroides (designated as Rs_P_wt). The positive-control mutant restored the production of phytylated BChl a, as the HPLC elution peak of the pigment was observed at 20.5 min (Fig. 3, profile 6).

In a converse experiment to the Hh_P_del mutant, we constructed a mutant with a modified bchP gene, in which the characteristic insertion sequences encoding PAPGVALPPDAKDG (Fig. 4, colored in red) derived from H. halochloris were inserted into the corresponding region of R. sphaeroides bchP gene. The mutant designated Rs_P_ins accumulated phytylated BChl a (Fig. 3, profile 7), which was consistent with the result observed for the Rs_P_wt strain.

It has been unknown whether a single GGR enzyme (BchP) in H. halochloris is responsible for the unusual THGG tail formation, or whether H. halochloris GGR can catalyze phytyl formation as observed in most phototrophs, but unknown protein(s) are involved in inhibiting the C10=C11 reduction in this bacterium. Here, we demonstrated that the heterologous expression of H. halochloris GGR in the R. sphaeroides mutant lacking its original GGR resulted in the accumulation of BChl a esterified with a THGG group. The results indicate that the GGR of H. halochloris itself is responsible for the production of the THGG moiety and that the enzymatic activity of GGR in the bacterium is distinct from that in most other phototrophic bacteria producing phytylated BChl a.

C. tepidum, a green sulfur bacterium, produces Chl a esterified with the THGG moiety, although it also biosynthesized BChl a with a regular phytyl tail. The THGG moiety detected in C. tepidum is 2,6-phytadienyl and therefore differs from 2,10-phytadienyl in H. halochloris (for structural comparison, see Fig. 1). Before recent studies, it had been assumed that C. tepidum has two GGRs: one catalyzes phytyl formation in BChl a biosynthesis, and another catalyzes THGG formation in Chl a biosynthesis. However, Harada et al. constructed a C. tepidum mutant lacking the single bchP gene (CT2256) and showed that the mutant accumulated BChl a and Chl a both esterified with the GG group (17). This indicated that there is only one bchP gene responsible for GG reduction in the pigment biosynthesis of C. tepidum (17, 18). Harada et al. also made complementation experiments and introduced CT2256 into R. capsulatus strain lacking its authentic GGR. The R. capsulatus mutant was verified to produce phytylated BChl a (17). These indicated that GGR (gene product of CT2256) of C. tepidum exhibits potentially catalytic activities to reduce GG to phytyl moiety. By contrast, when it reacts with Chl a_GG, the reduction of the C6=C7 double-bond of the isoprenoid tail is somehow inhibited or at least unachieved. In this study, we demonstrated that the GGR of H. halochloris itself is responsible for the THGG formation and that it has partial, unusual hydrogenation activities lacking the function of the C10=C11 double-bond reduction. Therefore, we conclude that the model proposed for C. tepidum GGR differs from that of H. halochloris GGR. In terms of catalyzing hydrogenation of the GG moiety only twice, the GGR of H. halochloris is likely to exhibit a novel catalytic mechanism and will provide insights into protein engineering. Additionally, the GGR variant, in which the N-terminal insertion peptides specific for H. halochloris were omitted (Hh_P_del), showed another-type partial activity catalyzing only a single hydrogenation reaction (Fig. 3, profile 5). The variant somehow acquired a novel reaction mode, which differs from the original GGR of H. halochloris.

Generally, GGR catalyzes the hydrogenation of carbon-carbon double bonds of unsaturated hydrocarbons to produce the corresponding single bond and works in various biosynthetic pathways for isoprenoid products, including α-tocopherols, phylloquinone, and archaeal cell membranes. One of the big enigmas on the catalytic mechanism of GGR is whether multiple hydrogenation reactions are successively conducted without releasing the substrate intermediates, or whether the intermediates (DHGG/THGG tails) dissociate from GGR before the next hydrogenation. The crystal structures of archaeal GGR from Sulfolobus acidocaldarius, including GGR bound to the substrate geranylgeranyl pyrophosphate, have been determined (22). In the 3D structure, the substrates were detected at three positions within GGR, although there is an active site in the vicinity of a single FAD. Therefore, it seems that the two observed substrates other than the one closest to the FAD were caught at binding pockets before or after hydrogenation reactions. In addition to site-directed mutation studies, the structure study of archaeal GGR has proposed the catalytic mechanism that the first and second hydrogenation might be processive and that the last third hydrogenation is probably not processive (22). The characteristic insertion of H. halochloris GGR (Fig. 4) is probably located close to the binding site that is relevant to the first and second hydrogenation in the archaeal GGR, according to the alignment of primary structures. Asparagine 90 and glycine 91 of S. acidocaldarius GGR (Fig. 4, colored in green) located in the vicinity of the pyrophosphate moiety of the substrate geranylgeranyl pyrophosphate (22) are at a similar position to the insertion region of H. halochloris GGR in the alignment (Fig. 4). We also performed protein structure prediction with AlphaFold for GGR of H. halochloris (Fig. 5). In the predicted structure, the characteristic insertion region constitutes a loop structure in the vicinity of the substrate binding site closest to FAD (Fig. 5B). These results could support the phenomenon that H. halochloris lacks hydrogenation at the C10=C11 position, which is probably the first hydrogenation event in other phototrophs, and that the variant GGR of Hh_P_del mutant lacks two hydrogenations, which are probably the first and second ones. However, we could not determine whether the DHGG moiety in the Hh_P_del mutant was a 2,6,10- or 2,10,14-phytatrienyl group. To determine the position of the unreduced double bond, a large number of pigment materials for analysis and the chemical standards of those intermediates are required. Further analysis on this will be reported elsewhere.

Using Rhodobacter species, Rhodopseudomonas species, barley, and radish sprouts, the order of the three hydrogenations by GGR was previously determined to be C10=C11, C6=C7, and then C14=15 (14) (see Fig. 1). Therefore, in the case of H. halochloris, the unachieved hydrogenation at the C10=C11 double bond corresponds to the first reaction (Fig. 1). In this study, the elution peaks of BChl a_DHGG and BChl a_THGG from mutants having H. halochloris GGRs were eluted slightly earlier than those prepared from Rhodopseudomonas sp. Rits strain (Fig. 3, profiles 3, 4, and 5). Mizoguchi et al. have reported HPLC elution profiles of BChl b esterified with two types of THGG moieties, namely, 2,10- and 2,14-phytadienyl tails, where BChl b_THGG with a 2,10-phytadienyl tail eluted slightly earlier than BChl b_THGG with a 2,14-phytadienyl tail (16). Similarly, BChl b_DHGG esterified with a 2,10,14-phytatrienyl tail eluted slightly earlier than BChl b_DHGG with a 2,6,14-phytatrienyl tail (16). These suggest that the isoprenoid moieties of BChl a_THGG and BChl a_DHGG from mutants with H. halochloris GGRs would be 2,10-phytadienyl and 2,10,14-phytatrienyl tails, respectively.

We thank Yasuhiro Shimane for his help on performing AlphaFold analysis. This work was partially supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI grant numbers 19H02018 and 18H03743 (to Y.T.) and JP17H06436 in the Scientific Research on Innovative Areas “Innovation for Light-Energy Conversion (I⁴LEC)” (to H.T.), and by the Astrobiology Center of National Institutes of Natural Sciences (grant number AB021015 to Y.T.).

Y.T. and H.T. designed the research and developed the concept. Y.T., J.H., and H.T. wrote the manuscript. Y.T., J.H., K.K., and K.T. performed the experiments. All authors analyzed the data and discussed the results for completing the manuscript.