|Home | About | Journals | Submit | Contact Us | Français|
The X-ray crystal structure of the 6-pyruvoyltetrahydropterin synthase (PTPS) homolog from Streptomyces coelicolor, SCO 6650, was solved at 1.5 Å resolution. SCO 6650 forms a hexameric T-fold that closely resembles other PTPS proteins. The biological activity of SCO 6650 is unknown, but it lacks both a required active-site zinc metal ion and the essential catalytic triad and does not catalyze the PTPS reaction. However, SCO 6650 maintains active-site residues consistent with binding a pterin-like substrate.
Biopterin is an essential cofactor in mammalian nitric oxide synthase and aromatic hydroxylation reactions, but is not commonly found in bacteria. Glycosylated biopterin analogs are synthesized by a number of photosynthetic bacteria and may be involved in photoreception or UV protection (Wachi et al., 1995 ; Yamazawa et al., 1999 ). Tetrahydrobiopterin (BH4) is biosynthesized in three steps from GTP by the successive actions of GTP cyclohydrolase I (GCH I), 6-pyruvoyltetrahydrobiopterin synthase (PTPS) and sepiapterin reductase (SR). PTPS is a zinc-containing protein that catalyzes two successive Amadori rearrangements to convert 7,8-dihydroneopterin triphosphate to 6-pyruvoyltetrahydropterin.
The hexameric rat PTPS assumes a ‘tunnel’ fold (T-fold; Nar et al., 1994 ; PDB code 1b66), which has also been reported for GTP cyclohydrolase I (GCH I; Nar et al., 1995 ; PDB code 1gtp), dihydroneopterin aldolase/epimerase (DHN aldolase; Sanders et al., 2004 ; PDB code 1rrw) and urate oxidase (Retailleau et al., 2004 ; PDB code 1r51). The T-fold of PTPS contains a dimer of trimers that form a stacked toroid. The active site of PTPS lies at the interface of three subunits (A, A′ and B; Ploom et al., 1999 ; see gray residues in Fig. 1 ). Three histidine imidazole side chains from subunit A (His23, His48 and His50) and the hydroxyl group at C2′ of the substrate coordinate an essential divalent zinc metal ion in the active site. An Asp88-His89 duo from subunit B activates Cys42 of subunit A to initiate the catalytic cycle (see Fig. 1 ). The endocyclic nitrogen at N3 and the exocyclic amino group at C2 hydrogen bond to the side-chain carboxylate of Glu107, while the backbone amide of the preceding residue (Thr106) hydrogen bonds to the pyrimidine carbonyl O atom (at C3). Additionally, a second glutamic acid side chain from subunit A (Glu133) forms a hydrogen bond to the C1′ hydroxyl of the substrate (Fig. 1 ). While subunit A′ is not directly involved in catalysis, it lines one side of the active site and the backbone carbonyl of Met70 forms a hydrogen bond to the exocyclic amino group, presumably stabilizing the bound substrate.
The genome sequence of Streptomyces coelicolor (Bentley et al., 2002 ) revealed the presence of a PTPS homolog (SCO 6650), which is nestled in a cluster of genes that appears to be conserved in all strains of actinomyces sequenced to date. In addition to PTPS, this cluster also contains a GTP cyclohydrolase II homolog, SCO 6655, which we have shown to catalyze the conversion of GTP to 2-amino-5-formylamino-6-ribosylamino-4(3H)-pyrimidinone 5′-phosphate (FAPy; Spoonamore et al., 2006 ). A biological imperative for the co-localization of these genes is not clear.
As part of a program to assign the biological function of this cluster, SCO 6650 was cloned, overexpressed in Escherichia coli and purified as either His6-tagged or native protein. Both the native and His6-tagged proteins crystallized readily and the structure of the selenomethionine-substituted His6-tagged protein was solved to 1.5 Å resolution. The model reveals that SCO 6650 lacks nearly all the catalytically essential residues but retains the residues required for binding a pterin-like substrate.
Restriction endonucleases were from New England Biolabs. Expression vectors pET28a and pET29a and expression strain E. coli BL21 (DE3) were from Novagen. The pGEM-T Easy kit was from Promega. LB refers to Lennox Broth. Pfu Turbo and DH5α were from Invitrogen. Purification resins were obtained from GE Healthcare. All other chemicals were obtained from VWR Scientific and Sigma–Aldrich.
The gene encoding the S. coelicolor PTPS homolog (sco6650) was amplified with Primer 1 (5′-GGAATTCCCATGGCATATGTTCAGCATCACCGTCCGCGATCAC-3′) and Primer 2 (5′-TTTTAAGCTTATTACAGCGCACGCTCGTAACTCGC-3′) from S. coelicolor genomic DNA and subcloned into the pGEM T-Easy vector prior to cloning into the NdeI/HindIII site of pET28a and pET29a overexpression plasmids for the expression of His6 or native protein, respectively. The sequence of the final construct revealed that a nonconservative mutation had inadvertently been introduced (via Primer 2) during initial cloning. To correct this error, the gene was amplified from the cloning vector using Primer 1 and Primer 3 (5′-GCAAGCTTATTACAGCGCACGCTCGTAACTCGCCCAGGC-3′) and recloned into the expression vectors. DNA sequencing confirmed that the error had been corrected.
Native protein was prepared by transformation of E. coli BL21 (DE3) with the pET29 expression vector bearing SCO 6650. Transformants were grown overnight in LB containing 34 µg ml−1 kanamycin prior to the inoculation of large-scale (1 l) growth containing 34 µg ml−1 kanamycin. Expression was induced by the addition of 0.1 mM isopropyl β-d-1-thiogalactopyranoside at an OD600nm of ~1; cells were harvested after 4 h. For purification, cell pastes were resuspended into 25 ml 0.02 M Tris–HCl pH 8 containing 0.5 mM DTT and supplemented with one tablet of Complete EDTA-free protease inhibitor (Roche). Cells were lysed by sonication and then centrifuged to remove cellular debris. The cleared lysate was applied onto a Q-Sepharose Fast Flow column (2.5 × 14.5 cm) that had been equilibrated with 0.02 M Tris–HCl pH 8 containing 0.5 mM DTT; the protein was eluted with a linear gradient of 0–0.5 M KCl in the loading buffer; SCO 6650 eluted at ~0.25 M KCl. Fractions that contained SCO 6650, as determined by SDS–PAGE, were pooled. Pooled fractions were dialyzed against 4 l 0.02 M Tris–HCl pH 8.0 containing 0.5 mM DTT with one buffer change. SCO 6650 was concentrated with a 10 000 Da molecular-weight cutoff centrifugal concentrator (YM-10 membrane, Amicon) and flash-frozen in liquid nitrogen until needed. Protein stock was quantified using the BCA protein-assay kit (Pierce) with BSA as a standard.
Selenomethionine-labeled N-terminally His6-tagged protein was expressed in M9 minimal media supplemented with selenomethionine as described previously (Doublié, 1997 ). The protein was purified essentially as described for the native protein above, except that an additional gel-filtration step was included as follows. Dialyzed and concentrated Q-Sepharose fractions were applied onto a Sephacryl S-300 column (2.6 × 60 cm) equilibrated with 0.02 M Tris–HCl pH 8 containing 0.5 mM DTT. The protein was eluted with the same buffer.
Metal analysis of the selenomethionine-labeled protein by ICP-OES (Garratt–Callahan) revealed 2.1 equivalents of selenium in each monomer, which is consistent with the post-translational removal of the N-terminal methionine, resulting in the presence of two Se atoms in each monomer.
Conditions yielding crystals of SeMet-containing His6-SCO 6650 were found using the sparse-matrix screening kits Crystal Screen I and II from Hampton Research. The hanging-drop method was utilized in these experiments. The SeMet histidine-tagged SCO 6650 was crystallized as follows: a 3 µl aliquot of the protein solution, which typically contained 40 mg ml−1 protein in 0.02 M Tris–HCl pH 8.0 and 0.5 mM DTT, was combined with 3 µl of a solution containing 2 M NaCl and 10% PEG 6000. The well volume was 0.5 ml. Crystals grew at room temperature in 7 d and were flash-frozen using 3.2 M NaCl, 10% PEG 6000 for cryoprotection.
Selenium MAD data were collected at 100 K using an ADSC Quantum-4 CCD detector on beamline X26C at Brookhaven National Laboratory. Data were collected at three wavelengths (see Table 1 for statistics) and reduced using HKL-2000 (Otwinowski & Minor, 1997 ). Data to 2.6 Å resolution were used for phasing. Nine selenium positions were located and initial phases were calculated using SOLVE (Terwilliger & Berendzen, 1999 ). Density modification, phase extension to 1.5 Å resolution and initial structure building were performed using RESOLVE (Terwilliger, 2003 ). The structure was refined using REFMAC5 (Murshudov et al., 1997 ) and rebuilt with Coot (Emsley & Cowtan, 2004 ). There is residual difference electron density on the faces and the empty center of the hexamer along the pseudo-threefold axis. Disc-shaped pieces of electron density, presumably from partially occupied histidine residues of the His6 tag (sequence MGSSHHHHHHSSGLVPRGSH), can be seen stacked on Trp25 residues from each chain located at the top and bottom of the central cavity. Except for these partially occupied stacked histidine residues, the residual electron density could not be interpreted and since neither the residue number nor chain identification of these histidine residues could be determined, they were not included in the model.
Structurally, SCO 6650 is a member of the T-fold protein family, which also includes GTP cyclohydrolase I, dihydroneopterin aldolase and urate oxidase (Murzin et al., 1995 ). The asymmetric unit of the His6-tagged protein crystals contains a hexamer, which is likely to be the biologically relevant oligomerization state of the protein. Gel-filtration studies with the native SCO 6650 are consistent with the presence of a hexamer (data not shown). The protein is composed of two stacked trimers. A 12-stranded antiparallel β-barrel forms the central tunnel in the protein; the overall diameter of the trimer is 55 Å (Fig. 2 ) and the hexamer is 53 Å in length.
The X-ray crystal structures of several PTPS homologs have been reported (Nar et al., 1994 ; Ploom et al., 1999 ) or deposited in the Protein Data Bank (PDB codes 1y13, 2g64, 2a0s, 2dtt, 2dj6 and 2oba). The crystallization of the E. coli homolog (YgcM) was reported recently (Seo et al., 2008 ). With the exceptions of the Pyrococcus horikoshii and E. coli PTPS homologs, which are trimers, PTPS proteins are hexamers. Site-directed mutagenesis studies have confirmed that the active site of PTPS has contributions from both toroids (A, A′ and B in Fig. 2 ), suggesting a hexameric biological unit (Bürgisser et al., 1994 , 1995 ; Ploom et al., 1999 ). The trimeric PTPS homologs lack several key catalytic residues (Cys42, Asp88 and His89; see Fig. 2 ) which would be contributed by a subunit from an adjacent toroid (subunit B in Fig. 2 ). However, we note that Asp88 and His89 are conserved in both the Py. horikoshii and E. coli sequences.
Structural similarities between SCO 6650 and Py. horikoshii (PDB code 2dtt), Pseudomonas aeruginosa (2oba), Rattus rattus (1b66), Plasmodium vivax (2a0s) or Pl. falciparum (1y13) PTPS homologs were confirmed by structural overlays. Interestingly, despite low sequence identities (15–21%) and similarities (29–40%), the root-mean-squared deviation of Cα atoms varies between 1.6 and 2.1 Å for the oligomers or 1.3 and 1.6 Å for the monomers (Krissinel et al., 2004 ).
A multiple sequence alignment comparing the PTPS sequences from R. rattus, Ps. aeruginosa and Py. horikoshii with that of SCO 6650 is shown in Fig. 3 . In each case, an X-ray crystal structure of the protein is available, allowing the locations and roles of conserved residues to be evaluated. While the alignment reveals substantial sequence similarities, several residues that are known to be required for the action of PTPS are absent from SCO 6650 (see Fig. 1 ).
While the putative active-site cleft in SCO 6650 is similar to that described for PTPS (Ploom et al., 1999 ), the residues that line the site are different. In the following discussion, we utilize the R. rattus X-ray crystal structure (PDB code 1b66) for comparison to SCO 6650, since significant biochemical data highlight the importance of various residues in the active site of the protein (Bürgisser et al., 1995 ; see Fig. 1 for schematic comparison). The R. rattus PTPS structure shows that a zinc divalent metal ion is bound to the imidazole side chains of three His residues (23, 48 and 50) as well as the 2′-OH of the substrate. The zinc-coordinating residues are required for activity (Bürgisser et al., 1995 ). In SCO 6650, an Ala residue is found at position 30. A His→Leu site-directed variant at the corresponding position in rat PTPS (His50) loses the ability to bind the zinc ion and is catalytically inactive. Zinc is not observed in the crystal structure of SCO 6650, nor could it be detected by ICP-OES analysis of protein preparations, even when the growth medium was supplemented with 0.1 mM ZnSO4 (data not shown). A trio of residues in the active site of PTPS, Cys42, His89 and Asp88, are thought to be involved in initiating catalysis. In SCO 6650, the Cys42–His89–Asp88 triad is replaced by Ala24–Tyr68–Asn67. Collectively, the substitutions that are apparent in SCO 6650 relative to rat PTPS (summarized in Fig. 1 ), where biochemical studies have confirmed the importance of these residues, suggest that SCO 6650 cannot catalyze the conversion of 7,8-dihydroneopterin triphosphate to 6-pyruvoyltetrahydropterin (Bürgisser et al., 1995 ). In fact, we have been unable to detect this activity with SCO 6650 under conditions where the mouse PTPS is active (data not shown).
While the catalytic residues required for PTPS function are not conserved in SCO 6650, the residues involved in binding of a pyrimidine-containing substrate are retained. In this analysis, ligand-bound models of PTPS from Py. horikoshii (Bagautdinov et al., 2007 ; PDB code 2dtt), R. rattus (Ploom et al., 1999 ; PDB code 1b66), Pl. vivax (PDB 2a0s) and Pl. faciparum (PDB code 1y13) were utilized. Each of these proteins has a Ser/Thr-Glu motif (Thr106-Glu107 in rat PTPS) that hydrogen bonds to the pyrimidine ring. In the available substrate-bound PTPS structures, the carboxylate side chain of the glutamate residue (Glu107 of rat PTPS) hydrogen bonds the exocyclic amino group at C2 and the endocyclic N3 nitrogen. The backbone amide of the Ser/Thr residue (Thr106 of rat PTPS) hydrogen bonds to the carbonyl O atom at C4. A similar duo of residues (Thr84-Glu85, see Fig. 1 ) are observed in an analogous position in the SCO 6650 structure. A backbone carbonyl (Met70 in rat PTPS) also hydrogen bonds to the exocyclic amino group. The backbone carbonyl of Val50 in SCO 6650 appears to be in the same position as the backbone carbonyl of Met70. Furthermore, the carboxylate side chain of Glu133, which has a catalytic role in rat PTPS (Bürgisser et al., 1995 ), interacts with the 1′-OH group of the substrate; Glu120 of SCO 6650 is in the same position. Therefore, in contrast to the catalytic triad and zinc ion-binding site, there is striking conservation in SCO 6650 of the residues that interact with the substrate, suggesting that it utilizes a substrate that is structurally similar to 7,8-dihydroneopterin triphosphate (see Fig. 1 for summary of interactions). We should note that a similar constellation of residues (corresponding to Thr84-Glu85) has been noted in the T-fold proteins GCH I and DHN aldolase (Hennig et al., 1998 ) for binding of the pyrimidine ring of the substrate.
The authors wish to thank Annie Dahlgran for cloning sco6650 and purifying some of the preparations used in this work. Reid McCarty and Alberto Rascon are acknowledged for gifts of mouse PTPS and E. coli GCH I, respectively. Data were measured on beamline X26C of the National Synchrotron Light Source (NSLS). Financial support for NSLS comes principally from the Offices of Biological and Environmental Research and of Basic Energy Sciences of the US Department of Energy and from the National Center for Research Resources of the National Institutes of Health. VB acknowledges support from the National Institutes of Health (GM072623). JES was partially supported by an NIH predoctoral training grant (GM08804). In addition, the research of VB was supported (in part) by a Career Award in Biomedical Sciences from the Burroughs Wellcome Fund.