Protochlorophyllide synthesis by recombinant cyclases from eukaryotic oxygenic phototrophs and the dependence on Ycf54

All chlorophototrophic organisms rely on the unique chemical properties of (bacterio)chlorophyll [(B)Chl] molecules for light harvesting and photochemical reactions, so the elucidation of the (B)Chl biosynthesis pathways is of great importance. In the common (B)Chl biosynthetic pathway, the least well-characterized step is the formation of the isocyclic E ring, via the oxidation and cyclization of the C13 methyl propionate group of Mg-protoporphyrin IX monomethyl ester (MgPME), producing 3,8-divinyl protochlorophyllide a (DV PChlide a). This step is catalyzed by two mechanistically unrelated cyclases: an O₂-sensitive, radical SAM enzyme containing [4Fe–4S] and cobalamin cofactors [1] is encoded by the bchE gene in most anoxygenic phototrophic bacteria and some cyanobacteria [2], whereas an O₂-dependent diiron monooxygenase is found in many purple bacteria, as well as cyanobacteria, algae and plants [3]. The catalytic subunit of the O₂-dependent cyclase was first identified from the purple betaproteobacterium Rubrivivax (Rvi.) gelatinosus and named AcsF (aerobic cyclization system Fe-containing subunit) [4]. Following the identification of two auxiliary subunits, Ycf54 [5,6] and BciE, the O₂-dependent cyclase has been delineated into three distinct classes: the Ycf54-dependent enzyme from oxygenic phototrophs, the AcsF-only enzyme in most anoxygenic phototrophs (e.g. in Rvi. gelatinosus) and the BciE-dependent alphaproteobacterial enzyme [3] (Figure 1B). The activity of the O₂-dependent cyclase has been demonstrated in Escherichia coli using heterologously expressed enzymes from Rvi. gelatinosus and the cyanobacterium Synechocystis sp. PCC 6803 (hereafter Synechocystis) [7], thus completing the identification of the O₂-dependent cyclase components in prokaryotic phototrophs.

Oxygenic phototrophs, in particular higher plants, rely solely on the O₂-dependent cyclase for Chl biosynthesis. Early research work was focused on biochemical characterization of the enzyme using fractions from cucumber developing chloroplasts, resolving the enzyme into soluble and membrane components [8,9]; a similar subcellular distribution was found subsequently for the Synechocystis enzyme [10]. The model plant Arabidopsis thaliana (hereafter Arabidopsis) contains a single AcsF protein, CHL27 [11], a homolog of AcsF from Rvi. gelatinosus [4], whereas the green alga Chlamydomonas reinhardtii (hereafter Chlamydomonas) has two AcsF isoforms, CRD1 and CTH1. Genes encoding these isoforms are expressed reciprocally; CRD1 has a basal level of expression that increases under copper- or O₂-deplete conditions, whereas CTH1 is produced under copper- and O₂-replete conditions [12,13]. Similarly, Synechocystis also has two AcsF homologs, CycI and CycII; CycI is constitutively active under both aerobic and micro-oxic conditions, whilst CycII is additionally required under micro-oxic conditions [14,15]. AcsF proteins have been shown to be membrane-bound in various organisms [5,11,16,17]. In the search for the ‘missing’ soluble component, Ycf54/LCAA was identified as an auxiliary cyclase subunit from Synechocystis by in vivo pulldown experiments using FLAG-tagged CycI and CycII as bait [5], and from tobacco by screening antisense plants with Chl deficiency [6]. The key role of Ycf54 proteins in cyclase function is exemplified by its ubiquity in oxygenic phototrophs and the severe phenotypes exhibited by knockout mutants including impaired AcsF protein levels, accumulation of MgPME, and lowered amounts of DV PChlide a and Chl [5,6,18–21].

Despite this recent progress with the O₂-dependent cyclase from purple bacteria and cyanobacteria, there has been no activity-based assignment of algal or plant cyclase components. We have shown previously that Rvi. gelatinosus and E. coli are suitable hosts for heterologous expression of cyclase components and demonstration of cyclase activity [3,7], and in this study we extend this approach to the O₂-dependent cyclase subunits from Synechocystis, Chlamydomonas and Arabidopsis, representative of cyanobacteria, green algae and higher plants, respectively. The detection of in vivo cyclase activity from recombinant Arabidopsis and Chlamydomonas enzymes shows that the eukaryotic cyclase consists of AcsF and Ycf54 subunits and Ycf54 is, therefore, an essential component of cyclases in oxygenic phototrophs.

Gene fragments encoding Arabidopsis CHL27 (AT3G56940.1), YCF54 (AT5G58250.1), Chlamydomonas CRD1(CHLREDRAFT_183476), CTH1 (CHLREDRAFT_205856), and CGL78 (CHLREDRAFT_162021) proteins, lacking the N-terminal chloroplast transit peptides predicted by the ChloroP 1.1 Server [22], were synthesized (Integrated DNA Technologies) with codons optimized for expression in E. coli. It is worth noting that, based on the sequence alignments of Ycf54 proteins and experimental tests, Arabidopsis YCF54 was assumed to contain a 72 aa chloroplast transit peptide instead of the predicted 80 aa. The nucleotide sequences of synthesized genes are listed in Supplementary Table S1.

The pBB[gene] plasmids were made by cloning the indicated gene fragment into the BglII/NotI or BglII/XhoI sites of pBBRBB-Ppuf_843–1200 [23]. The Synechocystis cycI and cycI-ycf54 genes were amplified using the pK18[cycI-ycf54] plasmid [3] as a template. Overlap extension PCR was used to generate a CHL27-YCF54 gene fragment with a ribosome binding site (TATAGGAGCTTGGATT) placed between the two genes. To apply the link and lock method [24], genes were first cloned individually into the NdeI/SpeI sites of a modified pET3a vector (containing an added SpeI site immediately upstream of the BamHI site). Then the genes were cut from the pET3a plasmids and adjoined consecutively in the described order using the procedures described previously [7]. Primers used in this study are listed in Supplementary Table S2. Plasmids were sequenced by Eurofins Genomics and are listed in Supplementary Table S3. E. coli strains were grown in LB medium and, if required, antibiotics were supplemented at 30 µg ml⁻¹ for kanamycin and 100 µg ml⁻¹ for ampicillin. Rvi. gelatinosus strains were grown in polypeptone-yeast extract-sodium succinate (PYS) medium [25] at 30°C and, when required, antibiotics were added at 40 µg ml⁻¹ for rifampicin and 50 µg ml⁻¹ for kanamycin. A spontaneous rifampicin-resistant mutant, ΔbchEΔacsF Rif^R, was isolated from the Rvi. gelatinosus ΔbchEΔacsF [3] mutant and served as a platform strain to test foreign cyclases from Synechocystis and Arabidopsis. Cyclase genes were cloned into the expression vector pBBRBB-Ppuf_843–1200 to get the pBB[gene] plasmids, which were conjugated into the ΔbchEΔacsF Rif^R strain via the E. coli S17-1 strain. E. coli S17-1 cells containing the plasmid were grown in LB medium with 30 µg ml⁻¹ kanamycin at 37°C for 24 h and 30 µl of the resulting culture were mixed with Rvi. gelatinosus cells harvested from 30 ml culture and resuspended in 100 µl of LB medium. The mating mixture was placed onto a well-dried LB agar medium and incubated at 30°C overnight before streaking out onto PYS agar medium with 50 µg ml⁻¹ kanamycin and 40 µg ml⁻¹ rifampicin to select for transconjugants. Bacterial strains used in this study are listed in Supplementary Table S3.

Rvi. gelatinosus strains were grown in 10 ml PYS medium, supplemented with 50 µg ml⁻¹ kanamycin if the strain contains a pBBRBB-puf_843–1200-based plasmid, in 50 ml Falcon tubes at 30°C with shaking at 175 rpm for 2 days. Harvested cells were either used for recording whole-cell absorption spectra or subjected to pigment extraction. Absorption spectra were recorded on a Cary 60 UV–Vis spectrophotometer with cells resuspended in 60% (w/v) sucrose, and normalized by the absorbance at 950 nm, followed by subtraction of the ΔbchEΔacsF Rif^R spectrum to correct for light scattering. Pigment extraction and subsequent HPLC analysis were performed as described previously [26].

E. coli C43(DE3) [27] was transformed with the pET3a-based plasmids and selected on LB agar supplemented with 100 µg ml⁻¹ ampicillin. A single colony was inoculated to 10 ml LB medium with 100 µg ml⁻¹ ampicillin and grown overnight at 37°C with shaking at 220 rpm. The next day, 50 µl of the resulting culture were used to inoculate 10 ml of LB medium with 100 µg ml⁻¹ ampicillin and grown at 37°C for 3 h. Then isopropyl-β-D-thiogalactopyranoside (IPTG) was added at 0.5 mM to induce gene expression. δ-Aminolevulinic acid (ALA) and Mg²⁺ (equimolar mixture of MgCl₂ and MgSO₄) were also added at 10 mM. The culture was incubated at 30°C with shaking at 175 rpm for 24 h before pigment extraction. Pigment extraction and high-performance liquid chromatography (HPLC) analysis were conducted as described previously [7] with slight modifications. The column was a Phenomenex Luna C18(2) reversed-phase column (particle size, 5 µm; pore size, 100 Å; 250 × 4.6 mm). The wash with 100% solvent B was shortened from 15 to 5 min. Elution was additionally monitored by fluorescence at 640 nm excited at 440 nm.

E. coli strains containing the pET3a-based plasmids were incubated and induced as for the in vivo cyclase assay. After 5 h induction cells were harvested and resuspended in 50 mM Tris–HCl, pH 8.0, 5 mM EDTA, supplemented with Proteinase Inhibitor Cocktail (Sigma–Aldrich), and lysed by sonication on ice (3 ×  20 s with 20 s resting between cycles). The cell lysate was clarified by centrifugation at 3000×g at 4°C for 5 min to remove unbroken cells and cell debris, followed by SDS–PAGE analysis using NuPAGE 12% Bis-Tris Protein Gels (Thermo Fisher Scientific). Protein bands were transferred to a polyvinylidene difluoride (PVDF) membrane, which was probed with specific primary antibodies raised against Arabidopsis CHL27 (Agrisera), YCF54 (Agrisera) and Synechocystis Ycf54 [5], and then with a secondary anti-rabbit antibody conjugated with horseradish peroxidase (Sigma–Aldrich). Chemiluminescent signal was developed using the WESTAR SUN enhanced chemiluminescence substrate (Cyanagen) and detected by an Amersham Imager 600 (GE Healthcare).

The catalytic subunit of O₂-dependent cyclase, AcsF, containing a signature diiron binding motif (E-x_n-E-x-x-H-x_n-E-x_n-E-x-x-H), is classified as a membrane-bound carboxylate diiron protein [28] (Figure 2). We used a ΔbchEΔacsF Rif^R mutant of the facultative phototroph Rvi. gelatinosus, which lacks both the O₂-sensitive and -dependent cyclase enzymes [3], as a background strain for assaying plasmid-borne Synechocystis cycI and ycf54 genes, or genes encoding the Arabidopsis CHL27 and YCF54 proteins. Whole-cell absorption spectra of transconjugant strains showed that neither CycI nor CHL27 was functional but when partnered with Ycf54 or YCF54, respectively, each was able to replace the missing native cyclase and restore the assembly of photosynthetic complexes to variable levels (Figure 3). It is striking that the activity of the Synechocystis CycI-Ycf54 pair matched that of the native AcsF protein (Figure 3). We further analyzed the accumulation of BChl a in these strains by HPLC and the detected level is in good agreement with the abundance of photosynthetic complexes revealed by the corresponding whole-cell absorption spectra (Figure 3).

The Synechocystis CycI-Ycf54 and Arabidopsis CHL27-YCF54 pairs were clearly active in the purple phototrophic bacterium Rvi. gelatinosus. Next, we adopted a previously reported strategy [7] to demonstrate their activity in the non-photosynthetic bacterium E. coli. Cyclase encoding genes were cloned into the pET3a-based plasmid IM [7], which contains genes encoding the first two enzymes of the Chl pathway that convert native, endogenous protoporphyrin IX (PPIX) to Mg-protoporphyrin IX (MgP), then to the cyclase substrate, MgPME (Figure 1A). These constructs were transformed into the E. coli C43(DE3) strain, followed by in vivo cyclase assays. Pigments from the harvested E. coli cells were extracted and analyzed by HPLC with elution profiles monitored by absorbance, and also fluorescence for a higher sensitivity for detecting Chl biosynthesis intermediates. In the absence of a functional cyclase, only MgPME (Soret band at 416 nm; emission maximum at 593 nm) was detected, which eluted at ∼29 min (Figure 4). The E. coli strains expressing Rvi. gelatinosus AcsF (plasmid IA), Synechocystis CycI-Ycf54 (IM-cycI-ycf54) and Arabidopsis CHL27-YCF54 (IM-CHL27-YCF54), accumulated DV PChlide a (Soret/Q_y bands at 440 and 631 nm, respectively; emission maximum at 639 nm) as represented by a peak at ∼26 min, which is accompanied by a decreased level of MgPME (Figure 4). Arabidopsis CHL27-YCF54 was less active than the Rvi. gelatinosus and Synechocystis enzymes, leaving a considerable amount of MgPME substrate not converted to product (Figure 4).

Unlike the single AcsF found in Arabidopsis, the green alga Chlamydomonas contains two AcsF isoforms, CRD1 and CTH1 [13]. Genes encoding CRD1, CTH1 and the Chlamydomonas Ycf54 protein, CGL78, were synthesized and cloned into the IM plasmid [7] for in vivo cyclase assay in E. coli. HPLC analysis of pigments extracted from the assays reveals that CRD1 was functional only when CGL78 was co-expressed (Figure 5). In the current test system, no activity was detected for CTH1 even in the presence of CGL78 (Figure 5).

Our in vivo cyclase assays using enzymes from representative species of cyanobacteria, green algae and plants, reveal the universal requirement of Ycf54 for cyclase activity among oxygenic phototrophs. Sequence alignments of Ycf54 proteins clearly identify a conserved Ycf54 domain (designated PF10674/DUF2488 in the Pfam database; Figure 6A). The alignments also reveal C-terminal extensions present only in green algal and plant Ycf54 proteins (Figure 6A), containing nine conserved residues (F180, W186, A188, P189, Y190, Y193, W197, W198 and W201; numbering in Arabidopsis YCF54, NP_200633) among these species (Figure 6B). To investigate the potential role of the C-terminal extension, we truncated the Arabidopsis YCF54 gene to generate a YCF54* mutant, with the C-terminal 37 aa sequence removed. We also swapped Ycf54 proteins between Arabidopsis, Chlamydomonas and Synechocystis cyclases to check whether Ycf54 is functional with a foreign AcsF protein. CHL27-YCF54* behaved essentially the same as CHL27 alone in the in vivo cyclase assays, indicating the inactive nature of YCF54* (Figure 7). The Synechocystis Ycf54 protein was unable to replace the Arabidopsis or Chlamydomonas counterpart whereas Arabidopsis YCF54 and Chlamydomonas CGL78 supported some cyclase activity (detectable by fluorescence) when partnered with Synechocystis CycI (Figure 7). These results suggest green algal and plant Ycf54 proteins differ from the proteins found in other organisms, probably by their C-terminal extensions.

Our in vivo cyclase assays demonstrate the cyclase from oxygenic phototrophs is dependent on the Ycf54 protein for activity. To check whether Ycf54 affects the expression/stability of AcsF when expressed in E. coli, we detected the level of AcsF and Ycf54 proteins in clarified cell lysates using specific antibodies. The commercially available anti-CHL27 antibody also reacts with Synechocystis CycI but does not react with Chlamydomonas CRD1 or CTH1 so the immunodetection was limited to Arabidopsis and Synechocystis cyclases. In contrast with the indistinguishable CycI levels with and without Ycf54, the CHL27 level was clearly dependent on its native interacting partner YCF54 (Figure 8). The absence or C-terminal truncation of YCF54 significantly decreased the abundance of CHL27, which could not be rectified by the presence of Synechocystis Ycf54 (Figure 8). Additionally, the C-terminal truncation seems to destabilize YCF54, and/or affect the interaction between YCF54 and its antibody as YCF54* produced a much weaker signal than YCF54 (Figure 8).

The enigma of the subunit composition of the O₂-dependent cyclase has been clarified by the discovery of three distinct classes of the enzyme among photosynthetic organisms [3]. Genetic complementation experiments conducted in the Rvi. gelatinosus cyclase mutant revealed the equivalence between the three classes of the enzyme, as indicated in Figure 1B. The next advance involved demonstration of heterologous cyclase activity in E. coli, with two of the three classes of cyclase, from Rvi. gelatinosus and Synechocystis [7], essentially completing the long endeavour to find the ‘missing’ cyclase components. It is widely accepted that cyanobacteria are evolutionarily related to the chloroplasts of algae and plants, endorsed by the conservation of photosynthetic components and processes between cyanobacteria and eukaryotic phototrophs. This conservation also exists in the O₂-dependent cyclase as indicated by the shared subunits, AcsF and Ycf54, so the likelihood that eukaryotic cyclase would require extra subunits was low. In this study, we show that, indeed, no extra cyclase-specific subunits are required, and heterologously expressed Arabidopsis CHL27 and YCF54 function as an active cyclase in Rvi. gelatinosus (Figure 3) and E. coli (Figure 4). We did not detect any cyclase activity of Chlamydomonas CTH1 co-expressed with CGL78 in E. coli (Figure 5), possibly because CTH1 was not expressed or it was expressed but unstable in E. coli. Alternatively, the activity of CTH1 may depend on specific copper and/or oxygen levels that differ from those in the E. coli cellular context since it has been documented that CTH1 is produced under copper- and oxygen-replete conditions in Chlamydomonas [12,13]. Nevertheless, we were able to demonstrate the in vivo cyclase activity of Chlamydomonas CRD1 and CGL78 produced in E. coli (Figure 5). A reductase is likely required, but E. coli can provide this function for a variety of bacterial, algal and plant cyclases, and Rvi. gelatinosus can also supply reductase activity for a cyclase from a cyanobacterium [3] and a plant (Figure 3). Thus, this study has defined the cyclase-specific subunits of the eukaryotic cyclase.

At this point, it is necessary to mention the cyclase studies conducted with barley xantha-l and viridis-k mutants, which suggested that the enzyme in this organism contains at least one soluble and two membrane-bound components, of which one is encoded by the Xantha-l gene, an acsF orthologue, and the other could be defective in the viridis-k mutants [16]. The suggestion that barley Ycf54 is a membrane-associated cyclase component increased the potential number of components to at least four [29]. Analysis of barely mutants is a powerful tool, but it is likely to detect lesions in serially linked processes because Chl biosynthesis is co-ordinated with photosystem biogenesis through multiple regulation checkpoints such as the initiating magnesium chelatase and the light-dependent PChlide oxidoreductase steps, and potentially the O₂-dependent cyclase step [30–32]. Thus, cyclase-deficient phenotypes are not necessarily due to lack of a cyclase subunit but can be a product of perturbed regulation or unstable cyclase components or even shortage of co-substrates/cofactors. Moreover, various factors can affect cyclase activity as indicated by the studies showing that the NADPH-dependent thioredoxin reductase and 2-Cys peroxiredoxins [33,34], and the thylakoid plastoquinone pool, are linked with cyclase activity [33,35].

Analogous to other diiron enzymes, the diiron center of cyclase needs to be reduced from +3 to +2 during the catalytic cycle by a reductant or an electron donor, which is believed to be NADPH as suggested by in vitro cyclase assays conducted with chloroplast/plastid fractions from plants and Chlamydomonas [8,10,16] and with Synechocystis cell extracts [10]. However, both AcsF and Ycf54 lack an apparent NAD(P)H-binding motif, implying that the direct electron donor is yet to be determined. As a cellular electron currency, NADPH is utilized to reduce many redox-active components, and the biochemical fractions used for in vitro cyclase assays probably contained some of these redox-active components, one of which could be the direct electron donor to the diiron center of cyclase. Herbst et al. [20] recently proposed that the plastidal ferredoxin-NADPH reductase (FNR) could be involved in reducing the cyclase based on the interaction between FNR1 and YCF54, and the disturbed cyclase activity that resulted from FNR1 deficiency. In oxygenic phototrophs FNR utilizes the electrons from photosynthetically reduced ferredoxin (Fd) to produce NADPH, and when there is a shortage of reduced Fd, such as in the dark, the enzyme can also catalyze the reverse reaction. Reduced Fd serves as the electron donor to a few redox enzymes including some involved in (B)Chl biosynthesis: Fd-dependent 8-vinyl reductase [36], dark-operative PChlide oxidoreductase [37], chlorophyllide oxidoreductase [38] and Chl a oxygenase [39]. In particular, some diiron enzymes use reduced Fd as the reductant as in the case for plant stearoyl-acyl carrier protein Δ⁹ desaturase [40] and p-aminobenzoate N-oxygenase [41]. To identify the direct electron donor to cyclase, future work should test reduced Fd as well as other types of reductants using purified cyclase subunits rather than undefined complex cellular fractions. Considering the heterologous activity of cyclase in non-pigmented E. coli [7] (Figures 4 and 5), it is predictable that the as-yet unidentified reductant and its potentially existing oxidoreductase are generic and shared with other metabolic processes rather than being specific to the cyclase reaction.

Our work further confirms that Ycf54 is an authentic cyclase subunit in oxygenic phototrophs. In their native hosts, Ycf54 is required for the accumulation of AcsF as demonstrated in Synechocystis [5,18,19], tobacco [6] and Arabidopsis [20] and its interaction with AcsF, mediated by the conserved R82 residue that forms part of a positively charged patch on Ycf54, is required for optimal PChlide formation as shown in Synechocystis [19]. The dependency of AcsF accumulation on Ycf54 was also found for the Arabidopsis enzyme when heterologously expressed in E. coli, but not for the Synechocystis cyclase as indicated by the unaffected level of CycI in the absence of Ycf54 (Figure 8). These results cannot rule out the possibility that Ycf54 plays a role in the assembly of the diiron center of CycI, which requires future characterization of purified CycI from a Ycf54-minus background. It is also possible that Ycf54 plays a role in substrate delivery/channelling or, alternatively, electron transfer during the catalytic cycle; in vitro biochemical assays using purified AcsF and Ycf54 proteins will be required to investigate a catalytic role for Ycf54. In addition, swapping Ycf54 proteins between different cyclase enzymes revealed the Synechocystis protein differs from its Arabidopsis and Chlamydomonas homologs as it was unable to stabilize CHL27 (Figure 8) and its co-expression with CHL27 or CRD1 did not result in a functional cyclase (Figure 7). This difference may be explained by the C-terminal extension, which is present only in green algal and plant Ycf54 proteins (Figure 6A), and which is highly conserved in eukaryotic phototrophs (Figure 6B). Despite its small size, ∼40 aa, this C-terminal extension is significant in the context of the ∼15 kDa Ycf54 protein and its removal abolishes cyclase activity (Figure 7), the basis for which appears to be the absence of any YCF54 protein, and also a significantly lowered level of CHL27 (Figure 8). We propose that the C-terminal extension plays a role in the normal function of green algal and plant YCF54 and CHL27 proteins. It has been reported that Arabidopsis and barley YCF54 proteins form oligomers [6,29] whereas the 1.3 Å structure of the Synechocystis protein indicates that it is monomeric [19]. The C-terminal extension may stimulate and/or maintain the oligomerisation of Ycf54 through hydrophobic interactions between its highly conserved residues, of which 7 out of 9 are aromatic (Figure 6B). Alternatively, the C-terminal extension may be involved in the interaction between green algal and plant Ycf54 and AcsF proteins. Our work also suggests that future studies involving the heterologous activity of CHL27, and possibly CRD1, in E. coli will benefit from co-expression of their cognate Ycf54 proteins.

The authors declare that there are no competing interests associated with the manuscript.

G.E.C. designed and performed research, analyzed data and wrote the paper. C.N.H. designed research and wrote the paper.

G.E.C. and C.N.H. gratefully acknowledge financial support from the Biotechnology and Biological Sciences Research Council [BB/M000265/1], and European Research Council Synergy Award 854126.

MgP

Mg-protoporphyrin IX

PPIX

protoporphyrin IX

PVDF

polyvinylidene difluoride

PYS

polypeptone-yeast extract-sodium succinate