R Eactivity of Recombinant and Mutant Vanadium Bromoperoxidase from the Red Alga Corallina Ofﬁcinalis Jayme N

Journal of Inorganic Biochemistry 91 (2002) 59–69 www.elsevier.com/locate/jinorgbio

R eactivity of recombinant and mutant vanadium bromoperoxidase from the red alga Corallina ofﬁcinalis Jayme N. Carter, Kimberly E. Beatty, Matthew T. Simpson, Alison Butler* Department of Chemistry and Biochemistry, University California Santa Barbara, Santa Barbara, CA 93106-9510, USA Received 19 November 2001; received in revised form 1 February 2002; accepted 4 February 2002

Abstract

Vanadium bromoperoxidase (VBPO) from the marine red alga Corallina officinalis has been cloned and heterologously expressed in Esherichia coli. The sequence for the full-length cDNA of VBPO from C. officinalis is reported. Steady state kinetic analyses of monochlorodimedone bromination reveals the recombinant enzyme behaves similarly to native VBPO from the alga. The kinetic Br2 HO22 parameters (Kmm51.2 mM, K 517.0 mM) at the optimal pH 6.5 for recombinant VBPO are similar to reported values for enzyme purified from the alga. The first site-directed mutagenesis experiment on VBPO is reported. Mutation of a conserved active site histidine residue to alanine (H480A) results in the loss of the ability to efficiently oxidize bromide, but retains the ability to oxidize iodide. Kinetic I2 HO22 parameters (Kmm533 mM, K 5200 mM) for iodoperoxidase activity were determined for mutant H480A. The presence of conserved consensus sequences for the active sites of VBPO from marine sources shows its usefulness in obtaining recombinant forms of VBPO. Furthermore, mutagenesis of the conserved extra-histidine residue shows the importance of this residue in the oxidation of halides by hydrogen peroxide.  2002 Elsevier Science Inc. All rights reserved.

Keywords: Vanadium; Haloperoxidase; Cloning; Expression; Mutant

1 . Introduction pong mechanism [3,4]. The oxidized halogen intermediate can act to halogenate an appropriate organic substrate or Vanadium bromoperoxidases (VBPO) are enzymes that oxidize a second equivalent of hydrogen peroxide to catalyze halide oxidation by hydrogen peroxide. These produce dioxygen in the singlet excited state (Scheme 1) enzymes, isolated from the major classes of marine algae, [5,6]. It has also been established that bromination of are thought to function in the biosynthesis of halogenated organic substrates by vanadium bromoperoxidase proceeds marine natural products [1,2], the scope of which ranges through an electrophilic (i.e. Br1 ) rather than a radical from halogenated indoles, terpenes and acetogenins to process (Br? ) process [7]. volatile halogenated hydrocarbons. In many cases the Recently the X-ray crystal structure of native VBPO halogenated compounds are of pharmacological interest from the brown alga Ascophyllum nodosum and the red due to their biological activities (e.g. antimicrobial, an- alga Corallina ofﬁcinalis have been solved to 2.0 and 2.3 tifungal, antiinﬂammatory, antiviral, etc.). Catalytically active vanadium bromoperoxidase requires one equivalent of vanadium (V) per subunit of enzyme. Using vanadium as the cofactor, these bromoperoxidases catalyze the two-electron oxidation of bromide and iodide by hydrogen peroxide in a substrate-inhibited bi-bi ping-

*Corresponding author. Tel.: 11-805-893-8178. E-mail address: [email protected] (A. Butler). Scheme 1.

Aû resolution, respectively. The vanadium (V) atom in the off the coast of Santa Barbara California, and frozen at X-ray structure of A. nodosum VBPO (AnVBPO) resides in 280 8C. a trigonal bipyrimidal coordination geometry similar to the Total RNA was prepared from the alga by phenol–SDS vanadium (V) site in vanadium chloroperoxidase (VCPO) extraction methods [12] and poly(A)1 RNA was further [8]. Vanadium (V) in the form of vanadate ion, is axially isolated using the PolyATract mRNA isolation system coordinated by a histidine residue (His486 ) in AnVBPO. (Promega). A double stranded cDNA library was prepared The negatively charged cofactor is neutralized by several from poly(A)1 RNA (5 mg), using the Zap cDNA syn- hydrogen bond interactions from side chain residues at the thesis kit and the Zap cDNA cloning kit according to active site. The protein residues (Ser416 , Gly 417 , Lys 341 , manufacturer’s instructions (Stratagene). Arg349 Nh1 , Nh2 , Arg 480 Nh2 , His 418 ) act as proton donors Polymerase chain reaction (PCR) amplifications with to the oxyanion and constitute the central part of the rigid oligonucleotides designed from the known nucleotide vanadate binding site. The X-ray crystal structure of sequences of VBPO1 and VBPO2 from C. pilulifera were VBPO from C. officinalis (CVBPO) showed the vanadate performed using cDNA prepared from total mRNA iso- binding residues to be identical to residues identified in lated. The upstream primer sequence used, (59- AnVBPO, even though the structure was solved with GCCGAGGGCAGCCCATTCCATCC-39) codes for the inorganic phosphate in place of vanadate [9]. amino acid sequence [AEGSPFHP] from VBPO1. The The active sites of VBPO for both C. officinalis and A. downstream primer, (59-AAGTAGTGAACACCTG- nodosum contain an additional histidine residue not present CCATGTT-39) codes for the amino acid sequence in VCPO (i.e. His478 CVBPO, sequence reported by Isupov [NMAGVHYF] from VBPO1 [13]. The temperature et al. [9], and His411 AnVBPO). The additional histidine profile was 94 8C for 0.5 min, 51 8C for 0.5 min and 68 8C residue is not coordinated to the vanadate cofactor, but is for 2.0 min for 25 cycles in a Perkin-Elmer thermocycler. proposed to be within hydrogen bonding distance to one of The PCR reactions were performed using high fidelity pfu the vanadate oxygen atoms. It has been proposed that the Turbo DNA polymerase (Stratagene). extra histidine may influence the activated peroxovanadate The single fragment obtained by PCR amplifications intermediate by decreasing its susceptibility to nucleophilic was used to probe a cDNA library from C. officinalis. One attack by chloride during catalysis and may act as a proton million clones were screened by plaque hybridization with donor/acceptor during the enzyme reaction, thus giving 39-end labeled oligonucleotide. The oligonucleotide probe rise to the observed differences in catalytic activities of (59 - GCCGAGGGCAGCCCATTCCATCCGTCCTAC - 39) VCPO and VBPO [8,10,11]. However, the role of this was labeled with fluorescein-11-dUTP using terminal residue has not been fully determined. transferase (ECL 39-oligolabelling module, AP Biotech). In an effort to probe the role of certain active site Phage plaques were transferred to nylon membranes, and residues and to elucidate the functional significance of the the DNA fixed to the membranes by alkali fixation extra histidine residue at the active site we have cloned methods. Following prehybridization of the membranes for VBPO from the red alga Corallina officinalis. In this 2 h at 45 8C in a solution containing 53SSC (0.750 M report we describe the first sequence for the full-length NaCl, 0.075 M sodium citrate, pH 7.0), 0.02% SDS (w/v), cDNA of CVBPO, the deduced amino acid structure, and 100 mg salmon sperm DNA and 53Denhardt’s solution, the bacterial expression system used to obtain recombinant the membranes were hybridized overnight at 45 8C with enzyme. Kinetic parameters for the recombinant protein the fluorescein labeled probe. were determined and found to be similar to the native Membranes were washed several times with 13SSC, enzyme. We also report the first mutational study on the 0.1% SDS at 57 8C. Membranes were prepared for anti- active site of VBPO. As an initial approach to understand- body incubations by blocking the filters at room tempera- ing the function of the additional histidine residue and its ture in 0.1 M Tris–HCl pH 7.5, 0.15 M NaCl and 20-fold influence on halide oxidation (i.e. His480 rCVBPO; this dilution of antibody liquid block solution for 60 min (AP work, as translated from the full nucleotide sequence), the Biotech). The membranes were incubated overnight at effects of mutating the residue His480 to an alanine was room temperature with a 1:5000 dilution of an antifluores- examined. We show that His480 is important for the cein antibody conjugated to horseradish peroxidase (AP oxidation of halides by VBPO and the mutation of the Biotech) in 0.1 M Tris–HCl, pH 7.5, 0.4 M NaCl and 0.5% extra histidine residue influences enzyme–cofactor interac- (w/v) bovine serum albumin, followed by several washes tions. with the same buffer at room temperature. Oligonucleotide probe detection was performed using luminol-based substrates (AP Biotech) with the conjugated 2 . Methods and materials reporter enzyme, HRP. Chemiluminescent signals were detected using ECLீ Hyperfilmீ. Positive clones were 2 .1. cDNA clone isolation and sequencing isolated and in vivo excision of the cDNA insert-containing pBluescript phagemids were performed. The cloned Corallina officinalis alga was collected in April of 1998 inserts were sequenced using T3/T7 primers from opposite J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69 61

directions by an automated DNA sequencer (Iowa State NH43 VO . Cell lysis was performed using lysozyme treat- University, DNA sequencing facility). ment (100 mg/ml) followed by sonication. Identification of the 59-coding region for the VBPO Soluble protein was purified by modification of the cDNA was carried out by PCR amplification with cDNA protocol described for recombinant VBPO from C. piluli- template using primers designed from the 59-nucleotide fera [13]. In all anion-exchange purification steps KBr salt sequences of VBPO1 and VBPO2, and downstream se- was used instead of NaCl. Active fractions eluting from the quence information obtained by cDNA library screening. DEAE FF anion-exchange column were heat shock treated The forward primer in the reaction was 59-at658C for 20 min followed by immediate centrifugation ATGGGTATTCCAGCTGACAAC-39 and the reverse at 16 000 g. The supernatant was used in all subsequent primer used was 59-CGTAGCCACCTCAGTCACCAG-39. purification steps. In addition, the gel-filtration resin The temperature profile of the PCR reaction was 94 8C for Superose 6 (Pharmacia) was used in FPLC as a last 0.5 min, 52 8C for 1.0 min and 68 8C for 2.5 min for 25 purification step. cycles. Ampilified PCR products were excised from a 0.7% agarose gel and cloned into the pGEM-T easy vector 2 .4. Production of mutant H480A (Promega) for further sequence analysis. Mutant H480A was produced using the Quickchange 2 .2. Nucleotide and protein sequence alignments site-directed mutagenesis kit (Stratagene). The mutation analysis was made directly in the pTrc2.CVBPO expression vector according to the Quickchange protocol using the primers Nucleotide sequences for VBPO from C. officinalis and 59 - TCGCCGAGGGCAGCCCATTCGCCCCGTCCTAC- C. pilulifera were aligned with CLUSTAL W without manual GGAAGCGGCCA-39 and the complementary primer 59- adjustment using BIOEDIT sequence alignment editor [14]. TGGCCGCTTC CGTAGGACGGGGCGAATGGGCTGC- The amino acid sequence for VBPO from C. officinalis was CCTCGGCGA-39. After confirming the presence of the also aligned with other VHPO using CLUSTAL W. Sequence desired mutation by sequencing, the vector was trans- identity analysis was performed using the GENEDOC pro- formed into the E. coli strain BL21 for expression experi- gram, v. 2.5 [15]. ments. Purification of mutant H480A was as described for rCVBPO. 2 .3. Expression and purification of recombinant vanadium bromoperoxidase 2 .5. Conditions for steady state kinetic analysis of native, recombinant and mutant CVBPO The expression vector pTrcHis2B (Invitrogen) was doubly digested with NcoI and Pst1 followed by de- The reaction buffer for the steady state kinetic analyses phosphorylation of the vector ends. Using PCR, Nco1 and of monochlorodimedone (MCD) bromination consisted of Pst1 restriction sites were generated at the 59 and 39 ends, 80 mM sodium phosphate at pH 6.50 with varied con- respectively of the VBPO gene. The forward primer used centrations of potassium bromide (0.5–50 mM) and suffi- was 59-CTAGCCATGGGTATTCCAGCTGAC-3]]]] 9 and the cient sodium sulfate to give a final ionic strength of 0.8 M. reverse primer was 59-TTCCTTCTGCAGCAATAAACG-]]] The hydrogen peroxide concentrations were varied from 5 TTGGCCTAA-39. The restriction sites are underlined in mM to 5 mM. All reactions were performed at 25 8C. The each primer. The PCR product was doubly digested with concentration of hydrogen peroxide was determined spec- 2 Nco1 and Pst1 restriction enzymes to generate the restric- trophotometrically by the formation of triiodide (I3 ) as tion ends on either side of the VBPO gene. Following described [16]. purification of the gene by a 1% agarose gel, the gene was Iodoperoxidase activity for rCVBPO and mutant H480A 22 ligated to the Nco1 and Pst1 sites of the pTrcHis2B were measured by following the conversion of I to I3 by 21 21 expression vector. A transformation reaction of the ligation hydrogen peroxide at 350 nm (eM 526 400 cm M ). product was performed into the E. coli strain BL21. Enzyme free reactions were used as background controls. A small starter culture of the transformed bacteria with The enzyme activity was determined indirectly by measur- the resultant plasmid was used to inoculate a 1-l culture of ing the consumption of H22. O The reaction buffer con- LB-medium containing 100 mg/ml of ampicillin. Recom- sisted of 80 mM sodium phosphate, pH 6.5, with varied binant expression was induced with 1 mM IPTG after 6 h concentrations of potassium iodide (0.25–100 mM). The of growth at 37 8C. Protein expression experiments were hydrogen peroxide concentrations were varied from 5 mM performed at 20, 30 and 37 8C for various lengths of time to 800 mM. All reactions were performed at 25 8C. In to optimize protein expression levels. addition, steady state kinetic reactions for mutant H480A The induced cells were harvested by centrifugation at were performed in the presence of excess vanadate (100 5000 g for 20 min at 4 8C followed by resuspension of the mM) to account for possible affects on protein–cofactor cells in 50 mM Tris–SO4 buffer, pH 8.0, with 1 mM interaction introduced by the mutation. 62 J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69

2 .6. Steady state kinetic analysis sequences obtained from cDNA screening and PCR experiments. Sequencing of at least fifteen isolated clones from The initial rates, v, as a function of hydrogen peroxide both directions was performed to check for ambiguities in or halide concentrations were fit to the Michaelis–Menten the gene sequences and the possible presence of isoforms. expression of the form v5V[HO]/22K1[H 22 O ] by an A single gene sequence was obtained. The resultant clone halide iterative process, where V and K are functions of Km of CVBPO is 1794 nucleotides in length, where the open H22 O halide and Kmm[17]. The kinetic parameters, K and reading frame of the gene encodes a protein of 598 amino HO22 Km were obtained from appropriate fits of initial rate acids (Fig. 1). The calculated molecular mass of the data as a function of hydrogen peroxide and bromide/ protein is 65 458 Da, similar to the molecular mass of the iodide concentrations using Cleland’s programs PINGPONG monomer unit (64 kDa) estimated by SDS–PAGE electro- [17]. All measurements were performed in duplicate or phoresis. triplicate. Results presented are the least square fits to the The cDNA of CVBPO when aligned to the gene data and not the best fit by the PINGPONG program. sequences of bpo1 and bpo2 from C. pilulifera was found to have 96 and 93% sequence identity, respectively. Examination of the ORF for the gene bpo1, bpo2 and 2 .7. Enzyme activity assays cvbpo showed that the codon usage between the algae is similar and does not appear to contribute to the observed In-gel activity staining was performed for native, recom- amino acid differences. binant and mutant forms of CVBPO. SDS–polyacrylamide The translated amino acid sequence for CVBPO was gel electrophoresis was carried out using 10% Ready-Gels examined with respect to other known amino acid se- (Bio-Rad) under native conditions (in the presence of quences of vanadium-dependent bromoperoxidases. Align- SDS, but without reducing agent or boiling). The presence ments to the bromoperoxidases isolated from C. pilulifera of bromoperoxidase activity in the gel can be assessed by showed greater than 90% sequence identity, with 100% incubating the gel in 0.1 M sodium phosphate buffer, pH conservation of all active site residues (Fig. 2). Recently, 6.5, 0.1 M KBr, 100 mM vanadate, 1 mM o-dianisidine the amino acid sequence of VBPO from Corallina of- and 2 mM H O . 22 ficinalis (CVBPO protein) was reported with the X-ray Specific activity measurements for the bromination of crystal structure of the enzyme [9]. The reported amino monochlorodimedone were performed for rCVBPO and acid sequence was determined using protein sequencing mutant H480A at varied pH (4.5–7.0) with 0.1 M citrate– methods as well as direct comparisons of the electron phosphate buffer. Reactions were performed under con- density map to the known sequences of VBPO from C. ditions of low substrate concentrations (0.5 mM H O , 5 22 pilulifera [13]. The alignment of the two amino acid mM Br2 ) and high substrate concentrations (5 mM H O , 22 sequences of CVBPO (translated and protein) revealed a 100 mM Br2 ) in the presence of 100 mM vanadate. Protein 91% sequence identity and showed 52 total sequence concentrations were determined according to bicinchoninic differences. In addition, all amino acids associated with the acid assay purchased from Pierce Chemical. active sites of these enzymes are completely conserved. Alignments of amino acid sequences from the marine brown algae Ascophyllum nodosum and Fucus distichus to 3 . Results recombinant CVBPO (rCVBPO) are shown (Fig. 2). Both VBPOs from the brown algae show 31% sequence identity 3 .1. DNA and amino acid sequence analysis of to recombinant rCVBPO. In addition to what was shown in recombinant CVBPO the alignment between the coralline red algae, all of the residues associated with the vanadate cofactor are 100% The cloning strategy used to obtain the gene for CVBPO conserved. Alignments of all known amino acid sequences was prompted by earlier studies which showed that the of VBPOs show the presence of highly conserved consen- vanadium-dependent bromoperoxidases from C. pilulifera sus sequences, GSPxHPSYxSGHA and LTxxGEI/LNK- and C. officinalis exhibited similar reactivity towards LAxNxAxGRxMxGxHYxxDxxxxLLLGE (Fig. 2). The antiserum generated against purified VBPO from C. piluli- conserved consensus sequence contains amino acid res- fera [18]. In addition, protein amino acid analyses of other idues reported to be associated with the anion-binding site vanadium-dependent haloperoxidases from marine sources of the vanadate cofactor as shown by X-ray crystal- have shown a high degree of conservation of active site lography [8,9,19]. residues associated with the binding site of the vanadate cofactor [8,9,13]. The use of primers and DNA probes designed to target 3 .2. Expression of C. officinalis VBPO and H480AinE. the active site sequences of vanadium-dependent coli haloperoxidases afforded a single clone for VBPO from the marine red alga Corallina officinalis. The complete gene For initial expression experiments of CVBPO in E. coli, sequence for CVBPO was constructed using nucleotide the gene was inserted into the expression vector J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69 63

Fig. 1. Nucleotide sequence and deduced amino acid sequence of Corallina officinalis vanadium bromoperoxidase (accession no. AF218810). The nucleotide sequences used in primer design for PCR amplification reactions are shown underlined. The nucleotide sequence used as a probe to screen the cDNA library is shown underlined with an arrow. pTrcHis2B (Invitrogen) under the control of a trc promot- Purification of the enzyme from bacterial cell lysates was er. very similar to the procedure described for recombinant Expression experiments were carried out at various VBPO from C. pilulifera, and yielded approximately 0.1– temperatures to optimize the amount of soluble protein 0.3 mg of protein per liter of culture with specific activities expressed. Experiments performed at 37 8C produced around 150–200 mmol MCD brominated/min/mg. These minimal amounts of soluble protein with the majority of yields are similar to what was reported for recombinant expressed protein forming inclusion bodies. Alteration of VBPO from C. pilulifera using a bacterial expression the growth temperature after induction of expression system [13]. proved to be important in order to increase the ratio of soluble to insoluble protein. Optimal soluble expression 3 .3. Kinetic analyses for recombinant CVBPO and the conditions for the enzyme were found to be 16 h at mutant H480A 20–25 8C, following induction with 1 mM IPTG. Recombinant CVBPO is produced as the apo-protein 3 .3.1. Recombinant CVBPO form in E. coli, and requires the addition of the vanadate The steady state rates of MCD bromination by rCVBPO cofactor for activity during the purification procedures. were investigated as a function of hydrogen peroxide 64 J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69

Fig. 2. Alignment of the amino acid sequence of vanadium bromoperoxidase (CVBPO) for Corallina officinalis with other known sequences of VBPOs. The sequences obtained from GenBank were aligned using CLUSTALW multiple alignment program and displayed using GENEDOC. Sequence identity analysis was performed using GENEDOC. Amino acid residues associated with the active site vanadium cofactor are marked with asterisks (*) above the sequence, and the highly conserved consensus sequences are underlined. Abbreviations: CVBPO translated, C. officinalis translated from nucleotide sequence, accession number AF218810; CVBPOprotein, C. officinalis obtained by protein sequencing, reported by Isupov et al. [9], accession no. 1QHB; VBPO1, C. pilulifera VBPO1, accession no. BAA31261; VBPO2, C. pilulifera VBPO2, accession no. BAA31262; ANVBPO, A. nodosum VBPO, accession no. 1QI9A-B; FDVBPO, Fucus distichus VBPO, accession no. AAC35279. concentration (5–200 mM) and bromide concentration produce a set of parallel lines (Fig. 3). Parallel lines were (0.5–5 mM) at pH 6.5. Plots of the initial rates of MCD also observed for initial steady state rates versus bromide bromination versus hydrogen peroxide and bromide con- concentrations at fixed hydrogen peroxide concentrations centrations showed typical Michaelis–Menten type satura- (5–50 mM). The parallel plots are consistent with the bi-bi tion kinetic behavior (data not shown). ping-pong mechanism reported previously for the class of Double reciprocal plots of the initial steady state rates of vanadium-dependent haloperoxidases [3,4,20]. Br H22 O MCD-bromination versus hydrogen peroxide concentra- The values of Kmmand K for rCVBPO at pH 6.5 tions at low fixed bromide concentrations (0.5–2.0 mM) were determined from the best fit using Cleland’s PINGPONG J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69 65

ﬁxed hydrogen peroxide concentrations. The calculated I2 value for the kinetic constant for iodide Km at pH 6.5 was 1.8 mM.

3 .3.2. Bromoperoxidase activity of mutant H480A Kinetic analysis of the haloperoxidase activity for mutant H480A was performed to evaluate the affect the mutation had on overall catalysis. Recombinant CVBPO was used as a comparative wild type control in the measurements. When the haloperoxidase activity of mutant H480A was measured at pH 6.50 using chloride, bromide and iodide as substrates it was found that the mutant had no chloroperoxidase activity, but retained the ability to oxidize bromide and iodide. The specific activity of H480A for bromide oxidation was 4% (i.e. 6.3 mmol MCD Fig. 3. Primary double reciprocal plot of the rate of MCD bromination as brominated/min/mg) of the rCVBPO activity at the same a function of hydrogen peroxide concentration at fixed, low bromide pH (i.e. 6.5) in the presence of excess vanadium cofactor concentrations at pH 6.50 for rCVBPO; d, 0.500 mM Br2 ; s, 0.750 mM (100 mM). Excess vanadium was included in the 222 Br ; ., 1.00 mM Br ; ,, 2.00 mM Br . Each line is the linear least haloperoxidase reactions to compensate for possible altera- squares fit to data. tions in enzyme–cofactor interactions introduced by the mutation. When purified native CVBPO, rCVBPO, and Br H22 O program [17]. The calculated values for Kmmand K mutant H480A were analyzed by native PAGE and in-gel was 1.20 mM and 17.0 mM, respectively. The kinetic activity assays performed with substrates bromide, hydro- constants for native CVBPO from C. officinalis, isolated gen peroxide and o-dianisidine (Fig. 4), it was shown that and purified from the same batch of alga as the mRNA for the rCVBPO (lane 2) and mutant H480A (lane 3) ran Br H22 O VBPO was isolated, i.e. Kmmof 1.29 mM and K of similarly in the gel to native CVBPO (lane 1). These 27 mM, were similar to those of the recombinant enzyme, results suggest the recombinant and mutant form of showing explicitly that the native and recombinant protein CVBPO retain a similar structure to native CVBPO. In behave similarly. The pH optimum and kinetic constants addition, lane 3 of the native gel analysis reveals that the reported previously for native VBPO isolated from C. mutant H480A retains substantially less bromoperoxidase officinalis are also in agreement with the values we report activity under the conditions of incubation than native here (Table 1) [21]. CVBPO and rCVBPO. The steady state rates of triiodide formation by rCVBPO Specific activity measurements for H480A and were also examined as a function of hydrogen peroxide rCVBPO, as a function of pH, were performed to de- (5–200 mM) and iodide (0.4–15 mM) concentration. termine the pH optimum of haloperoxidase activity (Fig. Double reciprocal plots of the initial steady state rates of triiodide formation versus hydrogen peroxide concentration at low fixed iodide concentration produces a set of parallel lines (data not shown). Parallel lines were also observed for initial steady state rates versus iodide concentrations at

Table 1 Kinetic parameters obtained from the rates of MCD bromination and triiodode formation at pH 6.50; kinetic parameters were obtained using the PINGPONG program

Br2 HO22 I2 VBPO sample KKmm K m (mM) (mM) (mM) rCVBPO 1.2 17.0c 1.8 20.0d Native CVBPOa 1.3 27 – Native CVBPOb 1.0 60.0 – Mutant H480A ND 200 33.0 Fig. 4. In-gel activity analysis of CVBPO isolated from Corallina ND, not determined. officinalis or E. coli (recombinant protein). Purified protein was separated a Enzyme isolated in our laboratory. on 10% SDS–polyacrylamide gels under native conditions. The proteins b From Sheffield et al. [21]. are visualized in the gel by activity staining with o-dianisidine, bromide, c Determined by bromide kinetics. and hydrogen peroxide at pH 6.5. Lanes: 15native VBPO from Corallina d Determined from iodide kinetics. officinalis (1 mg); 25rCVBPO (1 mg); 35mutant H480A (36 mg). 66 J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69

Fig. 5. Specific activity measurements of (•) rCVBPO and (s) mutant Fig. 6. H480A primary double-reciprocal plot of the rate of triiodide H480A as a function of pH. Reaction conditions were 3 nM rCVBPO or formation as a function of hydrogen peroxide concentration at fixed, low d 22s , 225 nM H480A, 100 mM bromide, 5 mM H22 O , and 100 mM vanadate. iodide concentrations at pH 6.50; ,5mMI ; ,6mMI ; ,10mM The reaction rates were determined in duplicate. The buffers used are I22 ; j, 12 mM I . Each line is the linear least squares fit to data. listed in Materials and methods.

with respect to protein–metal interactions. Without the 5). Recombinant CVBPO showed optimal bromoperoxid- availability of a crystal structure of the mutated active site ase activity around pH 6.0 using low substrate concen- it seemed conceivable that the mutation of His480 could 2 trations (0.5 mM H22 O , 5 mM Br ) and high substrate affect enzyme–cofactor interactions by altering hydrogen 2 concentrations (5 mM H22 O , 100 mM Br ). Mutant bond interactions to the cofactor, or by inducing changes in H480A was found to have maximum bromoperoxidase the electrostatic distribution potential at the active site activity between pH 6.0 and 6.5 under conditions of high center. During the purification procedure for rCVBPO and 2 substrate concentrations (5 mM H22 O and 100 mM Br ), mutant H480A from E. coli, the apo-enzymes are activated however with reduced specific activity as compared to with vanadate at pH 8.3. Formation of the rCVBPO rCVBPO (Fig. 5). At all pH values tested, mutant H480A holoenzyme is indicated by no further increases in bromoperoxidase activity could not be detected under haloperoxidase activity upon incubation in the presence of conditions of low substrate concentrations (0.5 mM H22 O , excess cofactor. Mutant H480A, following activation with 5 mM Br2 ). vanadate and purification, was found to have variable haloperoxidase activity depending on the presence or 3 .3.3. Iodoperoxidase activity of mutant H480A absence of excess cofactor in reaction solutions (Fig. 7). The severely reduced bromoperoxidase activity of The rate of MCD bromination by mutant H480A was H480A prompted the kinetic analysis of iodide oxidation. shown to increase when added excess cofactor was present The steady state kinetics of iodoperoxidase activity cata- in the reaction versus reactions in the absence of excess lyzed by mutant H480A was studied by the rate of cofactor (Fig. 7a and b). Reactions where mutant H480A triiodide formation, as described above for rCVBPO. and excess vanadate were preincubated prior to the addi- Triiodide formation was examined as a function of hydro- tion of bromide and hydrogen peroxide did not further gen peroxide concentrations (10–500 mM) and iodide increase the reaction rate (Fig. 7c). This result suggests the concentrations (1–50 mM) at pH 6.5. Plots of initial rates increased activity in the presence of excess cofactor is not of triiodide formation versus hydrogen peroxide concen- dependent on equilibrium being reached between mutant trations or iodide concentrations showed saturation-type H480A and excess vanadate in solution at this pH. Michealis–Menten kinetics (not shown). Double reciprocal Although, when mutant H480A was preincubated with 2 plots of initial steady state rates of I3 formation versus excess vanadate and hydrogen peroxide prior to the hydrogen peroxide or iodide concentrations produced sets addition of bromide a further increase in the reaction rate of parallel lines (Fig. 6). Using Cleland’s PING PONG was observed (Fig. 7d). These results suggest that the kinetic program the calculated kinetic constants for hydro- increase in activity for H480A is a combined affect of HO22 I2 gen peroxide Kmmand iodide K were 200 mM and vanadate and the first substrate in the catalytic cycle, 33 mM, respectively (Table 1). hydrogen peroxide. Although the specific activity of bromide and iodide oxidation is marginally improved in 3 .3.4. Effects of the vanadium (V) cofactor on the the presence of excess cofactor, the haloperoxidase activity haloperoxidase activity of mutant H480A of H480A is still only a small percentage of recombinant The reduced specific activity of H480A was considered CVBPO activity. J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69 67

4 . Discussion

4 .1. Kinetic analyses of recombinant CVBPO

In this study CVBPO from the red alga C. ofﬁcinalis was cloned and kinetically characterized. The cloned gene for CVBPO was used to study the active site mutant H480A in an effort to understand the role of this residue in

the activation of H22 O and the oxidation of halides (Fig. 8). Steady state kinetic analysis of rCVBPO for bromide and iodide oxidation indicates that the enzyme performs according to a substrate-inhibited bi-bi ping-pong mechanism, as was previously reported for this class of enzymes. The substrate kinetic parameters for the iodoperoxidase and bromoperoxidase-catalyzed reactions by rCVBPO are Fig. 7. Improved speciﬁc activity of mutant H480A in the presence of similar to reported values of other vanadium-dependent HO22 excess vanadate and hydrogen peroxide. (a) MCD activity assay con- haloperoxidases from marine algae [22–24]. The Km taining H480A (550 nM), 50 mM phosphate buffer, pH 6.50,100 mM for rCVBPO at pH 6.5 does not depend on the nature of

KBr, and 5 mM H22 O . (b) MCD activity assay containing H480A (550 the halide present, i.e. either bromide or iodide. These nM), 50 mM phosphate buffer, pH 6.50, 100 mM KBr, 5 mM H O and 22 results would support the proposed mechanism where 100 mM vanadate. (c) Same conditions as in (b), reaction was initiated by the addition of hydrogen peroxide following a 10-min preincubation of hydrogen peroxide ﬁrst coordinates the vanadate cofactor the enzyme and 100 mM vanadate. (d) Same conditions as in (b), reaction to form the peroxovanadate intermediate prior to oxidation I2 was initiated by the addition of halide following a ten min preincubation of the halide. The value of Km for classical vanadium of the enzyme, vanadate, and hydrogen peroxide. bromoperoxidases is limited to VBPO isolated from brown

Fig. 8. Active site superposition of vanadium bromoperoxidase (C. ofﬁcinalis) and vanadium chloroperoxidase (C. inaequalis) near the VCPO vanadate cofactor (shown in red, upper middle of ﬁgure) based on secondary structural alignments. Top left arrow points towards the overlay of the extra-histidine in VBPO and the phenylalanine side chain VCPO. Figure was created with Swiss-PDB-viewer [29]. 68 J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69

I2 algal sources (A. nodosum, Km 0.82 mM) [25]. The kinetics of iodide oxidation by H480A suggest the pres- kinetic constant for iodide for rCVBPO was found to be ence of the extra histidine at the active sites of VBPOs is I2 similar to the reported Km for the vanadium-dependent necessary for the formation of the peroxovanadate inter- iodoperoxidase from the marine brown alga Pelvetia mediate as well as the effective activation of the bound I2 canaliculata, PcI and PcII, Km 2.1 and 2.4 mM, respec- peroxide towards halide oxidation. tively [23]. Mutant H480A possessed similar specific activity values for iodoperoxidase activity when compared to vanadium 4 .2. Bromoperoxidase activity of mutant H480A iodoperoxidase from Laminaria ochroleuca and S. polyschides at the same pH, where specific activities ranged Mutation of the extra-histidine residue in rCVBPO from 0.5 to 5 U/mg [24]. These results raise questions significantly decreased activity towards bromide oxidation, regarding the active site coordination of the vanadate and essentially transformed H480A into a vanadium-de- cofactor for classical vanadium iodoperoxidases. Currently, pendent iodoperoxidase. The observed low residual bromo- primary amino acid sequence data are not available for peroxidase activity for mutant H480A could only be vanadium iodoperoxidases, thus it is unclear at this time observed under conditions of increased enzyme and sub- what the structural and mechanistic differences are be- strate concentrations. Similar results were observed for the tween vanadium bromoperoxidases and vanadium active site mutant R360A of VCPO involved in charge iodoperoxidases. The iodoperoxidase activity observed for neutralization of the vanadate cofactor [11]. Mutant mutant H480A suggests that there are primary residues at R360A of VCPO possessed residual chloroperoxidase the active sites of vanadium iodoperoxidases that may activity but retained the ability to oxidize the less elec- differ from the normally conserved consensus sequence for tronegative bromide, and demonstrated high substrate vanadium haloperoxidases. kinetic constants. The decreased chloroperoxidase activity for the R360A mutant was attributed to a weakened 4 .4. Effects of the vanadium (V) cofactor on the activation of the bound peroxide species [11]. The pH haloperoxidase activity of mutant H480A dependence on bromoperoxidase activity for H480A as compared to rCVBPO suggests that the extra histidine The improved specific activity of mutant H480A in the residue does not govern the pH at which overall catalysis presence of added vanadate and peroxovanadate cofactor occurs (Fig. 5). suggests mutation of the histidine alters the interaction of vanadate at the active site of the enzyme. A likely result of 4 .3. Mutant H480A iodoperoxidase activity the histidine mutation is a change in the electrostatic potential distribution around the active site, thereby reduc- HO22 The value of Km in the iodoperoxidase reaction for ing the charge neutralization of the negatively charged HO22 HO 22 H480A, Kmm5200 mM, was similar to K for vanadate cofactor. The improved haloperoxidase activity vanadium-iodoperoxidases from the brown algae P. observed following preincubation of the enzyme with HO22 canaliculata and Saccorhiza polyschides (PcI Km 100 vanadate and hydrogen peroxide implies that only a HO22 HO 22 mM, PcII Km1m200 mM and SpV K 478 mM). fraction of H480A is in the holoform of the enzyme, and HO22 Although Km for H480A is an order of magnitude possibly a more complete formation of the activated HO22 higher than Km for rCVBPO (Table 1). These results enzyme may occur in the presence of a preformed perox- suggest mutation of the extra-histidine residue in proximity ovanadate complex. These results are similar to what was to the active-site center affects hydrogen peroxide binding observed for the H404A mutant of VCPO, where this I2 to the enzyme. In contrast, Km for H480A was greater mutant was also shown to have increased haloperoxidase I2 than Km for both native vanadium iodoperoxidases and activity upon preincubation with vanadate and hydrogen I2 vanadium bromoperoxidases. The increased value of Km peroxide [10]. It is also conceivable that the improved for H480A may suggest that mutation of the extra-histidine halogenating activity observed following preincubation residue weakens activation of the bound peroxide species with vanadate and hydrogen peroxide reflects a greater for oxidation by halides. Previous speculations are that the stability of a peroxovanadate species over vanadate at the extra histidine present in VBPO contributes to the observed mutated active site. X-ray crystal structure analysis of differences in oxidative power between VCPO and VBPO. VCPO peroxovanadate structure showed hydrogen perox- The extra histidine is proposed to alter the protonation ide bound side-on to the vanadate in a distorted tetragonal state of the bound peroxide that would be attacked by the conformation, which is different than the trigonal bipyrimi- incoming halide ion [8,10,11]. The peroxovanadate moiety dal structure of the native holoenzyme [19,28]. in VCPO is predicted to be protonated, where as the extra histidine in VBPO is predicted to prevent protonation of the bound peroxide [10,11]. Model studies of V complexes 5 . Conclusion with hydrogen peroxide also imply that the protonation state of the bound peroxide species may be important in The conserved consensus sequences found for VBPO the mechanism of halide oxidation by VHPOs [26,27]. The from marine sources and their similar architecture to the J.N. Carter et al. / Journal of Inorganic Biochemistry 91 (2002) 59 –69 69 active sites of VCPO has provided a framework to obtain R eferences and characterize the recombinant forms of VBPO. This study is the first mutational analysis reported for VBPOs [1] D.J. Faulkner, Natural Prod. Report 15 (1998) 113–158. and the first full-length sequence of Corallina officinalis. [2] G.W. Gribble, Acc. Chem. Res. 31 (1998) 141–152. Mutation of the extra histidine residue present in rCVBPO [3] E. de Boer, R. Wever, J. Biol. Chem. 236 (1988) 12326–12332. [4] R.R. Everett, H.S. Soedjak, A. Butler, J. Biol. Chem. 265 (1990) showed the importance of this residue in the oxidation of 15671–15679. halides by hydrogen peroxide. The mutated enzyme was [5] R.R. Everett, A. Butler, Inorg. Chem. 28 (1989) 393–395. essentially transformed into a vanadium-dependent [6] R.R. Everett, J.R. Kanofsky, A. Butler, J. Biol. Chem. 265 (1990) iodoperoxidase. Iodoperoxidase activity of mutant H480A 4908–4914. showed the mutant bound H O ten times less tightly than [7] H.S. Soedjak, J.V. Walker, A. Butler, Biochemistry 34 (1995) 22 12689–12696. native and recombinant CVBPO. In addition, mutant I- [8] M. Weyand, H. Hecht, M. Kiesz, M.F. Liaud, H. Vitler, D. H480A showed a greater Km value than what is observed Schomburg, J. Mol. Biol. 293 (1999) 595–611. for VBPO. The extra-histidine residue present in VBPOs [9] M.N. Isupov, A.R. Dalby, A.A. Brindley, Y. Izumi, T. Tanabe, G.N. may contribute to the intricate balance of charges and Murshudov, J.A. Littlechild, J. Mol. Biol. 299 (2000) 1035–1049. hydrogen bond networks present in vanadium haloperoxid- [10] R. Renirie, W. Hemrika, R. Wever, J. Biol. Chem. 275 (2000) 11650–11657. ases that enable the activation of hydrogen peroxide for the [11] W. Hemrika, R. Renirie, S. Macedo-Ribeiro, A. Messerschmidt, R. oxidation of halides yet, a picture is emerging where by the Wever, J. Biol. Chem. 274 (1999) 23820–23827. differences in VBPO and VCPO activity is not as simple as [12] R.D. Palmiter, Biochemistry 13 (1974) 3606. replacement of one amino acid for another. The subtle [13] M. Shimonishi, S. Kuwamoto, H. Inoue, R. Wever, T. Ohshiro, Y. differences between these two enzymes in architecture Izumi, T. Tanabe, FEBS Lett. 428 (1998) 105–110. [14] T.A. Hall, Nucl. Acids. Symp. Ser. 41 (1999) 95–98. surrounding the vanadate sites provide a finely tuned [15] K.B. Nicholas, H.B. Nicholas, D.W. Deerfield, EMBnet News 4 distribution of charges, which reflects the ability to oxidize (1997) 14. halides. It is expected that the continued investigations of [16] F. Bjorksten, Eur. J. Biochem. 5 (1968) 133–142. site-directed mutants of VBPO (i.e. His480Phe is in [17] W. Cleland, Methods Enzymol. 63 (1979) 103–138. progress) and the generation of laboratory evolved mutants [18] N. Itoh, A.K.M. Hasan, Y. Izumi, H. Yamada, Biochem. Int. 15 (1987) 27–33. of VBPO will contribute to the understanding of the [19] A. Messerschmidt, R. Wever, Proc. Natl. Acad. Sci. USA 93 (1996) mechanistic differences between vanadium-chloro, bromo, 392–396. and iodoperoxidases. [20] H.S. Soedjak, A. Butler, Biochim. Biophys. Acta 1079 (1991) 1–7. [21] D.J. Sheffield, T. Harry, A.J. Smith, L.J. Rogers, Phytochemistry 32 (1993) 21–26. [22] M. Almeida, M. Humanes, R. Melo, J.A. Silva, J.J.R. Frausto da A cknowledgements Silva, H. Vilter, R. Wever, Phytochemistry 48 (1998) 229–239. [23] M. Almeida, M. Humanes, R. Melo, J.A. Silva, J.J.R. Frausto da We are grateful for the support from NSF Grant CHE- Silva, R. Wever, Phytochemistry 54 (2000) 5–11. 9529374 (A.B.), Sea Grant NA66RG0447, project R/MP- [24] M. Almeida, S. Filipe, M. Humanes, M.F. Maia, R. Melo, N. 76 and R/MP-69 (A.B.), and Sea Grant Traineeship to Severino, J.A.L. da Silva, J.J.R. Frausto da Silva, R. Wever, Phytochemistry 57 (2001) 633–642. J.N.C. funded by a grant from the National Sea Grant [25] H. Vilter, Bot. Mar. 26 (1983) 451–455. College Program, National Oceanic and Atmospheric [26] B.J. Hamstra, G.J. Colpas, V.L. Pecoraro, Inorg. Chem. 37 (1998) Administration, and the US Department of Commerce. We 949–955. thank Shane Anderson of the Marine Science Institute at [27] G.J. Colpas, B.J. Hamstra, J.W. Kampf, V.L. Pecoraro, J. Am. Chem. the University of California for the collection of algae. The Soc. 118 (1996) 3469–3478. [28] A. Messerschmidt, L. Prade, R. Wever, Biol. Chem. 378 (1997) views expressed herein are those of the authors and do not 309–315. necessarily reflect the views of NOAA or any of its [29] N. Guex, M.C. Peitsch, Electrophoresis 18 (1997) 2714–2723. subagencies. The US Government is authorized to re- produce and distribute for governmental purposes.