Cyanobacterial non-mevalonate pathway: (E)-4-hydroxy-3-methylbut-2-enyl diphosphate synthase interacts with ferredoxin in Thermosynechococcus elongatus BP-1.

(E)-4-Hydroxy-3-methylbut-2-enyl diphosphate synthase (GcpE), which catalyzes the conversion of 2-C-methyl-D-erythritol cyclodiphosphate (MEcPP) into (E)-4-hydroxy-3-methylbut-2-enyl diphosphate (HMBPP), is an essential enzyme of the non-mevalonate (2-C-methyl-D-erythritol-4-phosphate (MEP)) pathway for isoprenoid biosynthesis. The terminal steps of the MEP pathway are still not fully understood, although this pathway is necessary for survival in various organisms such as cyanobacteria, plastids of algae and higher plants, and the apicoplast of human malaria parasites. To determine the efficient redox partner for thermophilic cyanobacterial GcpE, We have expressed the gcpE and petF genes in Escherichia coli and studied the protein-protein interaction of GcpE protein with ferredoxin I (PetF) from the thermophilic cyanobacterium Thermosynechococcus elongatus BP-1. Recombinant GcpE protein was purified by an N-terminal His(6) tag and reconstituted as a [4Fe-4S](2+) metalloprotein. GcpE was shown to interact strongly with PetF via the bacterial two-hybrid system designed to detect protein-protein interactions. Moreover, a direct protein-protein interaction between PetF and GcpE was confirmed in an in vitro glutathione S-transferase (GST) pull-down assay. To investigate electron transfer activity from PetF to GcpE, we also constructed a NADPH-dependent reducing shuttle system with purified recombinant ferredoxin-NADP(+) oxidoreductase (PetH) and PetF. The result demonstrated that PetF has the ability to transfer electrons to GcpE. Thus, the combined data provide the first evidence that GcpE is a ferredoxin-dependent enzyme in T. elongatus BP-1.

The first five enzymatic steps of the MEP pathway have been well established, but the terminal steps are still not fully understood (9 -11). Recent data has shown that (E)-4-hydroxy-3methylbut-2-enyl diphosphate synthase (GcpE) and LytB proteins are iron-sulfur proteins containing a [4Fe-4S] 2ϩ cluster after reconstitution of the purified protein (12)(13)(14). GcpE catalyzes the reduction of MEcPP into HMBPP via two successive one-electron transfers (11,13). The last step of the MEP pathway is catalyzed by IspH (or LytB), which converts HMBPP into isopentenyl diphosphate or dimethylallyl diphosphate via two successive one-electron transfers (15). These reactions were followed using flavodoxin/flavodoxin reductase/NADPH or sodium dithionite as a reductant (11)(12)(13)15). In contrast to the bacterial GcpE enzyme, which utilizes flavodoxin/flavodoxin reductase/NADPH as a reducing shuttle system, the plant GcpE enzyme could not use this reduction system (14). Yet, there have been no reports of an efficient redox partner for GcpE or LytB protein in cyanobacteria, the plastid of higher plants, or the relict plastid of human malaria parasite.
Here we report the interaction between GcpE and ferredoxin I (PetF), enabling transfer of electrons from photosystem I to ferredoxin-dependent enzymes, in the thermophilic cyanobacterium T. elongatus BP-1. GcpE protein was shown to interact strongly with PetF via a bacterial two-hybrid system designed to detect protein-protein interactions. Moreover, a direct protein-protein interaction between PetF and GcpE was confirmed using an in vitro GST pull-down assay. We also constructed an NADPH-dependent reduction system with purified recombinant ferredoxin-NADP ϩ oxidoreductase (PetH), PetF, and GcpE and investigated that electron transfer activity of PetF to GcpE. From the reductive titration of PetF with reconstituted GcpE, the dissociation constant K d for the electron transfer of PetF to GcpE was estimated as ϳ10 M. Therefore, We propose that GcpE is a ferredoxindependent enzyme in T. elongatus BP-1.

EXPERIMENTAL PROCEDURES
Materials-Enzymes for DNA manipulation were obtained from New England Biolabs. Agar and organic nutrients for LB were obtained from Difco, and other chemicals were from Sigma, BD Biosciences, Clontech, and Qbiogene.
Cloning of Relevant T. elongatus BP-1 Genes-Standard procedures were used for most DNA manipulations. Gene sequences were obtained from the Kazusa DNA Research Institute CyanoBase (16,17), and all accession numbers given below refer to that data base. Using primers described below, all genes were amplified from T. elongatus BP-1 genomic DNA by PCR. The fidelity of all PCR-generated fragments was verified by direct DNA sequencing.
Cloning of the Gene Encoding T. elongatus BP-1 GcpE-Primers used to amplify the gcpE gene (tlr0996) were forward primer TeGcpENdeI and reverse primer TeGcpESacI (Table 1, also other primers used below). The resulting 1.2-kb product was digested with restriction enzymes NdeI and SacI and cloned into NdeI-and SacI-digested pET28b (Novagen), giving the plasmid pET28b-TeGcpE. The GcpE construct expressed from pET28b consists of GcpE fused at the N terminus to a 20-amino acid sequence that includes a His 6 -tag. cDNA encoding the gcpE gene was PCR-amplified using the forward primer TeGcpEBamHI and reverse primer TeGcpESalI. The product was digested with restriction enzymes BamHI and SalI and cloned into BamHI and XhoI restriction sites of the pTRG-target vector (Stratagene) in-frame with the RNAP␣ gene, giving the plasmid pTRG-TeGcpE.
Cloning of the Gene Encoding T. elongatus BP-1 PetH-Primers used to amplify the petH gene (tlr1211) were forward primer TePetH-ldNdeI and reverse primer TePetHXhoI. The resulting 0.92-kb PCR fragment was digested with restriction enzymes NdeI and XhoI and cloned into NdeI-and XhoI-digested pET24b (Novagen), giving the plasmid pET24b-TePetH-ld. This construct lacked CpcD-like rod linker domain contained in PetH.
Cloning of the Gene Encoding T. elongatus BP-1 PetF-Primers used to amplify the petF gene (tsl1009) were forward primer TePetFNcoI and reverse primer TePetFXhoI. The resulting 0.29-kb PCR fragment was digested with restriction enzymes NcoI and XhoI and cloned into NcoI and XhoI restriction sites of the pET42b vector (Novagen) in-frame with the GST gene, giving the plasmid pET42b-TePetF. cDNA encoding the petF gene was PCR-amplified using the forward primer TePetFNotI and reverse primer TePetFXhoI. The product was digested with restriction enzymes NotI and XhoI and cloned into NotI and XhoI restriction sites of the pBT-bait vector (Stratagene) in-frame with the cl gene, giving the plasmid pBT-TePetF. cDNA encoding the petF gene was PCR-amplified using the forward primer TePetFNdeI and reverse primer TePetFXhoI. The product was digested with restriction enzymes NdeI and XhoI and cloned into NdeI-and XhoI-digested pET21b (Novagen), giving the plasmid pET21b-TePetF.
Cloning of the Gene Encoding T. elongatus BP-1 Heme Oxygenase 1 (HO1)-Primers used to amplify the ho1 gene (tll0365) were forward primer TeHO1NotI and reverse primer TeHO1XhoI. The resulting 0.72-kb product was digested with restriction enzymes NotI and XhoI and cloned into NotI and XhoI restriction sites of the pTRG-target vector in-frame with the RNAP␣ gene, giving the plasmid pTRG-TeHO1.
Cloning of the Gene Encoding T. elongatus BP-1 Phycocyanobilin: Ferredoxin Oxidoreductase (PcyA)-Primers used to amplify the pcyA gene (tll2308) were forward primer TePcyANotI and reverse primer TePcyAXhoI. The resulting 0.711-kb product was digested with restriction enzymes NotI and XhoI and cloned into the NotI and XhoI restric-tion sites of the pTRG-target vector in-frame with the RNAP␣ gene, giving the plasmid pTRG-TePcyA.
Expression and Purification of Recombinant Proteins-The plasmid pET28b-TeGcpE was transformed into Escherichia coli strain BLR (DE3) (Novagen). A fresh single colony of E. coli BLR (DE3) was transformed with the plasmid expressing the His-TeGcpE fusion protein was cultured overnight at 37°C in 100 ml of Luria-Bertani medium containing 1% glucose. 80 ml of this culture was incubated overnight and used to inoculate 8 liters of Luria-Bertani medium. The cells were grown at 37°C to mid-log phase, and then His-TeGcpE was induced by adding 0.2 mM isopropyl ␤-D-thiogalactopyranoside at 20°C. Cells were harvested after overnight induction and lysed by sonication in binding buffer (50 mM sodium phosphate, pH 7.4) containing 10 mM ␤-mercaptoethanol, 300 mM NaCl, and 5 mM imidazole for 30 s on ice. The lysate was centrifuged at 50,000 ϫ g for 30 min, and the supernatant was applied to a BD TALON superflow metal affinity column (1.5 cm ϫ 5 cm BD Biosciences Clontech). The His-TeGcpE fusion protein was purified, according to the manufacturer's instructions (BD Biosciences Clontech). The peak fraction was concentrated to 3.7 ml using an Amicon Ultra-15 unit with a 30-kDa cut-off (Millipore). Final purification was carried out by gel filtration using an XK26/100 Sephacryl S-200HR column (Amersham Biosciences) equilibrated with buffer A (50 mM HEPES-NaOH, pH 7.5) containing 1 M NaCl and 5 mM dithiothreitol (DTT). The main fraction was concentrated to 18 mg ml Ϫ1 and rebuffered in buffer A containing 100 mM NaCl and 5 mM DTT, using prepacked Sephadex G-25 gel filtration columns NAP-10 (Amersham Biosciences). The plasmid pET24b-TePetH-ld containing the petH gene and the plasmid pET21b-TePetF containing the petF gene were expressed in E. coli BL21 (DE3) (Novagen) and purified essentially as described previously (18,19). The plasmid pET42b-TePetF containing the petF gene fused at the 5Ј-end to the gene coding for Schistosoma japonicum GST was constructed and transformed into E. coli strain HMS174 (DE3) (Novagen). A fresh single colony of E. coli HMS174 (DE3) was transformed with the plasmid expressing the GST-TePetF fusion protein and cultured overnight at 37°C in 50 ml of Luria-Bertani medium containing 1% glucose, according to the manufacturer's instructions (Novagen). 10 ml of the overnight culture was used to inoculate 1 liter of Luria-Bertani medium. The cells were grown at 37°C to mid-log phase, and then GST-TePetF was induced by adding 1 mM isopropyl ␤-D-thiogalactopyranoside at 25°C. Cells were harvested after overnight induction and lysed by sonication in the binding buffer (phosphate-buffered saline, pH 7.3) containing 1 mM DTT, 140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 and 1.8 mM KH 2 PO 4 for 30 s on ice. The lysate was centrifuged at 50,000 ϫ g for 30 min, and the supernatant was applied to a glutathione-Sepharose high performance column (1.5 cm ϫ 5 cm; Amersham Biosciences). The GST fusion protein was purified according to the manufacturer's instructions (Amersham Biosciences).
Reconstitution of the Iron-Sulfur Cluster in GcpE-Reconstitution of as-isolated GcpE with iron and sulfide was carried out inside an anaerobic chamber with argon-saturated buffers and solutions that were prepared with deoxygenated water. A typical reconstitution reaction contained 200 M TeGcpE and a 10-fold molar excesses of FeCl 3 and Na 2 S in a final volume of 1 ml. The protein was initially treated with a 50-fold molar excess of DTT for 10 min on ice. The FeCl 3 was then added, and a solution of Na 2 S was added dropwise over 10 min. After 4 h, the reaction mixture was desalted on a NAP-10 column (Amersham Biosciences) equilibrated with 50 mM HEPES-NaOH buffer (pH 7.5). To record the UV-visible absorption spectrum, a fraction of the reconsti-  CaCl 2 method, cotransformed with relevant constructs, and incubated overnight at 37°C. Interactions were determined to be positive as measured by growth on "selective screening medium" consisting of minimal medium plus 5 mM 3-amino-1,2,4-triazole (3-AT), 25 g ml Ϫ1 chloramphenicol, and 12.5 g ml Ϫ1 tetracycline and validated by growth on "dual selective screening medium" consisting of minimal medium plus 5 mM 3-AT and 12.5 g ml Ϫ1 streptomycin, 25 g ml Ϫ1 chloramphenicol, and 12.5 g ml Ϫ1 tetracycline, with all media prepared as outlined by the B2H instruction manual. The pBT-LGF2 and pTRG-GAL11 p constructs provided with the B2H kit served as a positive control (22).
GST Pull-down Assays-Purified GST-TePetF fusion proteins were incubated with glutathione-Sepharose high performance beads in phosphate-buffered saline buffer (pH 7.4) at 4°C with rotation and washed repeatedly with phosphate-buffered saline buffer (pH 7.4). When appropriate, the beads were incubated with reconstituted His-TeGcpE at 4°C for 15 min, followed by a wash step in phosphate-buffered saline buffer (pH 7.4). The beads were then washed repeatedly, digested at 2 and 12 h with factor Xa (Novagen) and separated from GST according to the manufacturer's instructions (Novagen). The reaction mixtures were centrifuged at 21,500 ϫ g for 5 min, and the supernatant was boiled for 3 min in 2ϫ SDS sample buffer, separated by SDS-PAGE, and stained with Coomassie Brilliant Blue.
Spectrometric Assay of Electron Transfer Activity-An electron transfer pathway from NADPH to GcpE was reconstituted using PetH and PetF. The assay mixture contained in a final volume of 500 l: 50 mM HEPES-NaOH (pH 7.5), 100 mM NaCl, 5 mM DTT, 50 M NADPH, 1 mM glucose 6-phosphate, 0.5 units of glucose-6-phosphate dehydrogenase, 7.2 nM PetH-ld, and 75.5 M reconstituted GcpE, and 0, 5, 10, 20, and 40 M PetF, respectively. The reaction was initiated by adding PetF at 25°C. In the assay system, a reduction of the [4Fe-4S] cluster in GcpE was directly measured spectrophotometrically as the decrease in A 412 .

Identification of gcpE Gene in T. elongatus BP-1-gcpE
represents a highly conserved gene identified in a variety of organisms including eubacteria, higher plants (1,23), and the human malaria parasite-Plasmodium falciparum and other protozoan parasites of the phylum Apicomplexa, all of them known to possess the MEP pathway ( Fig. 1) (2, 3, 24). Recently, we identified a GcpE homologue in the T. elongatus BP-1 genome data base (Kazusa DNA Research Institute CyanoBase) that encoded a whole putative TeGcpE sequence, evidenced by its high sequence identity to other GcpEs and the existence of three conserved cysteine residues (Fig. 1). The common binding motif for [4Fe-4S] clusters, the CXXC motif (25), is present in TeGcpE.
GcpE Is an Iron-Sulfur Protein-GcpE proteins are reported to be unstable, losing activity quickly during purification and, to some extent, even in the cell (11,14). His-GcpE protein is derived from GcpE by addition of a tag of six histidines at the N terminus. The protein was obtained by E. coli overexpression and purified aerobically using a BD TALON superflow column that specifically retains proteins containing a cluster of histidines. The enzyme was found by SDS-PAGE to be 99% pure ( Fig. 2A). The purified protein has a reddish-brown color in agreement with the light absorption spectrum (Fig. 2B, lower  spectrum), and the analysis for labile iron and sulfide suggested the presence of a protein-bound [4Fe-4S] center. However, iron content was substoichiometric with regard to GcpE, and the protein contained sulfide in slight excess with regard to iron, probably as a consequence of loss of the cluster during purification (Fig. 2B, lower spectrum). The as-isolated His- GcpE protein was therefore reconstituted with a 10-fold excess of ferrous iron and sodium sulfide under anaerobic conditions as described under "Experimental Procedures." After anaerobic desalting on a Sephadex G-25 column, the protein was intensely brown. The UV-visible spectrum of the reconstituted protein is also shown in Fig. 2B (upper spectrum). The electronic absorption spectrum of the as reconstituted His-GcpE displays absorption bands, including a shoulder at 305 and 585 nm and a hump at around 395 nm, more consistent with a [4Fe-4S] cluster (Fig. 2B, upper spectrum). During anaerobic reduction of reconstituted His-GcpE with 0.5 and 5 mM sodium dithionite, bleaching of the solution and a loss of the visible absorption bands were observed (Fig. 3).
Bacterial Two-hybrid Assay-To examine whether the interactions between PetF and GcpE occurred in T. elongatus BP-1, constructs were made to test the direct protein-protein interactions between PetF and ferredoxin-dependent enzymes, HO1, PcyA, and GcpE, via the B2H system. The region encoding the petF gene was ligated into the bait vector of the B2H system to produce the construct pBT-TePetF. The entire coding region of the ho1 and pcyA genes was ligated into the target vector to produce the construct pTRG-TeHO1 and pTRG-PcyA, and the entire coding region of gcpE gene was ligated into the target vector to produce the pTRG-TeGcpE construct. When the reporter strain was cotransformed with hybrid bait and target proteins. If the proteins interact, the RNA polymerase is recruited to the promoter, activating the detectable transcription of HIS3. Growth of the reporter strain on medium lacking histidine and containing 5 mM 3-AT occurs when transcriptional activation increases expression of the HIS3 gene product FIG. 4. Detection and comparison of protein-protein interactions between PetF and GcpE and ferredoxin-dependent enzymes (HO1 and PcyA) using a bacterial two-hybrid system. Cotransformed E. coli cells (BacterioMatch II validation reporter strain) were spotted onto non-selective screening medium (A) and selective screening medium (B) plates. Reporter strains contained pBT-empty (vector alone) and pTRG-empty (vector alone), pBT-LGF2 and pTRG-Gall11 P , pBT-TePetF and pTRG-empty, pBT-empty and pTRG-TeHO1, pBT-empty and pTRG-TePcyA, pBT-empty and pTRG-TeGcpE, pBT-TePetF and pTRG-TeHO1, pBT-TePetF and pTRG-TePcyA, or pBT-TePetF and pTRG-TeGcpE on the bait (pBT) and target (pTRG) plasmid. The known interaction between LGF2 and Gal11 P was used as a positive control (22), whereas the lack of interaction between pBT-empty (vector alone) and pTRG-empty (vector alone) serves as a negative control. B shows that the bacterial transformant (pBT-TePetF/pTRG-TeGcpE) was grown on selective screening medium plates (5 mM 3-AT). The expression of both proteins is not lethal to the reporter strain as evidenced by their growth on non-selective screening medium plates (A).
to levels that are sufficient to overcome competitive inhibition by 3-AT. This allows for positive selection for plasmids encoding interacting proteins on media containing 5 mM 3-AT. Interaction of the HO1 and PcyA proteins with PetF was not detected on medium lacking histidine (Fig. 4). The GcpE protein was shown to interact strongly with PetF, as indicated by the strong growth on medium lacking histidine (HIS3 activation) (Fig. 4) and validated by a resistance to streptomycin (aadA activation) (Fig. 5).
GST Pull-down Assay (Interaction between PetF and GcpE in Vitro)-In vitro interaction between PetF and GcpE was verified using a GST pull-down assay. First, the cDNA coding sequence of PetF was subcloned into the pET42b vector to generate a GST-TePetF, and this fusion protein was expressed in the HMS174 (DE3) bacterial strain. The purified His-TeGcpE fusion protein was incubated with affinity-purified GST-TePetF fusion protein immobilized on glutathione-Sepharose high performance beads. The GST-TePetF fusion protein was digested with factor Xa and separated from GST. After the treatment for 2 and 4 h, the TePetF and bound GcpE were then separated by 12.5% SDS-PAGE, and the proteins were detected with Coomassie Brilliant Blue stain. Fig. 6 shows that purified GST-TePetF efficiently pulled down GcpE protein (lanes 4 and 5), but no protein was bound by glutathione-Sepharose high performance beads (lanes 6 and 7). Lane 5 shows that GcpE was isolated with PetF. This result indicates that GcpE interacts with PetF directly.
Spectrophotometric Assay for Electron Transfer from NADPH to Holo-GcpE Protein-In photosynthetic organisms, a major function of PetF is to transfer electrons from photosystem I to ferredoxin-dependent enzymes. An NADPH-de-  pendent reduction system using TePetH-ld was also able to support PetF reduction (data not shown). To investigate electron transfer activity from PetF to GcpE, a spectrophotometric assay for an NADPH-dependent reduction system was used as shown in Fig. 7. Electron transfer activity to GcpE was observed reflecting a reduction of the [4Fe-4S] cluster in GcpE by PetF. The data show that all components, NADPH, PetH, and PetF, were required for reduction of GcpE (Fig. 7A). Time-dependent absorbance changes associated with the reductive titration of PetF to reconstituted GcpE are shown in Fig. 7B. From this reductive titration of PetF to GcpE, the dissociation constant value (K d ) for the electron transfer of PetF with GcpE was determined to be around 10 M (Fig. 7B). DISCUSSION In this study we described the biochemical characterization of T. elongatus BP-1 GcpE. When overexpressed in E. coli, the isolated GcpE protein, containing an N-terminal His 6 tag, contained small amounts of iron and sulfide and displayed a weak UV-visible spectrum in the 300 -700-nm region consistent with the presence of a [4Fe-4S] cluster. Anaerobic treatment of the protein with FeCl 3 and Na 2 S in the presence of DTT resulted in the uptake of substantial amounts of iron and sulfide, as well as a dramatic increase in the activity of the protein (13). Hypothetical mechanisms for the GcpE-mediated reaction suggest that enzymatic conversion of MEcPP to HMBPP by GcpE is dependent on a [4Fe-4S] cluster as a cofactor, which is sensitive to dioxygen, and can be reduced by 5-deazaflavin or sodium dithionite as an artificial one-electron donor (11,13,14,26).
To our knowledge, this is the first report characterizing GcpE in detail in terms of protein-protein interactions with PetF. Bacterial two-hybrid analysis of GcpE and ferredoxin-dependent enzymes (HO1 and PcyA) with PetF from T. elongatus BP-1 indicated that GcpE and PetF interact strongly and that PetF and the ferredoxin-dependent enzymes (HO1 and PcyA) interact, but less strongly, in this system. Moreover, a direct protein-protein interaction between PetF and GcpE was con-firmed in an in vitro GST pull-down assay. This raises the interesting possibility of the formation of a functional complex between PetF and GcpE. Such a complex might serve as a system for electron donations to GcpE in vivo. We also constructed an NADPH-dependent reduction system with purified recombinant PetH, PetF, and GcpE and demonstrated that PetF had the ability to transfer electrons to GcpE (Fig. 7). From the reductive titration of PetF with reconstituted GcpE, the dissociation constant K d for the electron transfer of PetF to GcpE was estimated as ϳ10 M. Therefore, the present work reveals that PetF has the ability to transfer electrons to GcpE. To further clarify the molecular mechanism of GcpE catalysis, we must establish an assay system using the electron transfer ability of PetF to GcpE.
Recent studies have not shown that GcpE directly and/or indirectly interacts with flavodoxin (13). In this study, PetF has been identified as an interacting partner of GcpE. The results show that TeGcpE protein interacted with PetF. PetF may function as an efficient electron donor for GcpE in thermophilic cyanobacteria. Other electron carrier proteins are probably unable to function as efficient redox partners for this GcpE protein in T. elongatus BP-1, because other small electron carrier proteins, such as flavodoxin (isiB gene), that could support the reduction of GcpE are absent from the genome (16). At present, we do not know what kind of reduction system is operating for GcpE other than PetF in T. elongatus BP-1. It is possible that such flavoproteins may support the reduction of GcpE. In conclusion, we propose that the GcpE catalytic reaction of enzymatic conversion of MEcPP to HMBPP is dependent on ferredoxin as a one-electron carrier protein (Fig. 8).