The phylogenetically conserved histidines of Escherichia coli porphobilinogen synthase are not required for catalysis.

Porphobilinogen synthase (PBGS) is a metalloenzyme that catalyzes the first common step of tetrapyrrole biosynthesis, the asymmetric condensation of two molecules of 5-aminolevulinic acid (ALA) to form porphobilinogen. Chemical modification data implicate histidine as a catalytic residue of PBGS from both plants and mammals. Histidine may participate in the abstraction of two non-ionizable protons from each substrate molecule at the active site. Only one histidine is species-invariant among 17 known sequences of PBGS which have high overall sequence similarity. In Escherichia coli PBGS, this histidine is His128. We performed site-directed mutagenesis on His128, replacing it with alanine. The mutant protein H128A is catalytically active. His128 is part of a histidine- and cysteine-rich region of the sequence that is implicated in metal binding. The apparent Kd for Zn(II) binding to H128A is about an order of magnitude higher than for the wild type protein. E. coli PBGS also contains His126 which is conserved through the mammalian, fungal, and some bacterial PBGS. We mutated His126 to alanine, and both His126 and His128 simultaneously to alanine. All mutant proteins are catalytically competent; the Vmax values for H128A (44 units/mg), H126A (75 units/mg), and H126/128A (61 units/mg) were similar to wild type PBGS (50 units/mg) in the presence of saturating concentrations of metal ions. The apparent Kd for Zn(II) of H126A and H126/128A is not appreciably different from wild type. The activity of wild type and mutant proteins are all stimulated by an allosteric Mg(II); the mutant proteins all have a reduced affinity for Mg(II). We observe a pKa of ∼7.5 in the wild type PBGS kcat/Km pH profile as well as in those of H128A and H126/128A, suggesting that this pKa is not the result of protonation/deprotonation of one of these histidines. H128A and H126/128A have a significantly increased Km value for the substrate ALA. This is consistent with a role for one or both of these histidines as a ligand to the required Zn(II) of E. coli PBGS, which is known to participate in substrate binding. Past chemical modification may have inactivated the PBGS by blocking Zn(II) and ALA binding. In addition, the decreased Km for E. coli PBGS at basic pH allows for the quantitation of active sites at four per octamer.

Porphobilinogen synthase (PBGS) is a metalloenzyme that catalyzes the first common step of tetrapyrrole biosynthesis, the asymmetric condensation of two molecules of 5-aminolevulinic acid (ALA) to form porphobilinogen. Chemical modification data implicate histidine as a catalytic residue of PBGS from both plants and mammals. Histidine may participate in the abstraction of two non-ionizable protons from each substrate molecule at the active site. Only one histidine is speciesinvariant among 17 known sequences of PBGS which have high overall sequence similarity. In Escherichia coli PBGS, this histidine is His 128 . We performed sitedirected mutagenesis on His 128 , replacing it with alanine. The mutant protein H128A is catalytically active. His 128 is part of a histidine-and cysteine-rich region of the sequence that is implicated in metal binding. The apparent K d for Zn(II) binding to H128A is about an order of magnitude higher than for the wild type protein. E. coli PBGS also contains His 126 which is conserved through the mammalian, fungal, and some bacterial PBGS. We mutated His 126 to alanine, and both His 126 and His 128 simultaneously to alanine. All mutant proteins are catalytically competent; the V max values for H128A (44 units/mg), H126A (75 units/mg), and H126/ 128A (61 units/mg) were similar to wild type PBGS (50 units/mg) in the presence of saturating concentrations of metal ions. The apparent K d for Zn(II) of H126A and H126/128A is not appreciably different from wild type. The activity of wild type and mutant proteins are all stimulated by an allosteric Mg(II); the mutant proteins all have a reduced affinity for Mg(II). We observe a pK a of ϳ7.5 in the wild type PBGS k cat /K m pH profile as well as in those of H128A and H126/128A, suggesting that this pK a is not the result of protonation/deprotonation of one of these histidines. H128A and H126/128A have a significantly increased K m value for the substrate ALA. This is consistent with a role for one or both of these histidines as a ligand to the required Zn(II) of E. coli PBGS, which is known to participate in substrate binding. Past chemical modification may have inactivated the PBGS by blocking Zn(II) and ALA binding. In addition, the decreased K m for E. coli PBGS at basic pH allows for the quantitation of active sites at four per octamer.
Porphobilinogen synthase (PBGS) 1 (EC 4.2.1.24) catalyzes the asymmetric condensation of two molecules of ALA to form porphobilinogen. This reaction is the first common step in the tetrapyrrole biosynthetic pathway, which is responsible for the formation of porphyrins, chlorins, corrins, and other essential cofactors (1). There is high sequence similarity among the 17 documented sequences of PBGS from a phylogenetically diverse selection of prokaryotes and eukaryotes (2-14, 16 -18, 20). 2, 3 A review of data on the PBGS mechanism has been published recently (21). All PBGS are metalloenzymes, and there is an intriguing phylogenetic variation in the use of metal ions.
Under physiologic conditions, Escherichia coli PBGS contains Zn(II) and Mg(II), each at a stoichiometry of eight per octamer (or two per active site) (22,23). The Zn(II) are believed to be analogous to the eight Zn(II) of mammalian PBGS; four Zn A participate in catalysis and four Zn B appear to bind essential cysteine residues (24 -26). Zn A and Zn B are both at the active site (27). For E. coli PBGS, it is difficult to differentiate between the two types of bound Zn(II), and Zn A Zn B is used to denote enzyme-bound Zn(II) in both sites of E. coli PBGS. The eight equivalent Mg(II) (termed Mg C ) act as allosteric activators and have no counterpart in the mammalian protein (23), but appear to be common to bacterial and plant PBGS (21,28). Although not at the active site, Mg C causes a 50-fold increase in k cat /K m , affecting both parameters favorably. Mg C increases the affinity of the protein for Zn A Zn B (23), but Mg C cannot bind to PBGS unless Zn(II) is bound (22).
It has been proposed that both Zn A and Zn B draw ligands from a putative metal binding domain in a mutually exclusive fashion and that Mg C binds elsewhere in the sequence (21,25,26,28). Fig. 1 shows the sequence of the proposed metal binding domain. The technique of extended x-ray absorption fine structure spectroscopy discriminated between Zn A and Zn B on the basis of the chemical nature of their ligands (26). The conserved tyrosine, aspartate, and histidines of the metal binding domain are candidates for some of the predominantly nonsulfur ligands to a pentacoordinate Zn A . The four cysteines of this region are proposed to tetrahedrally coordinate Zn B . In those plant and bacterial species where the cysteines are not present, there are many aspartic acid residues, and Mg B is proposed to functionally replace Zn B (17,21).
The PBGS-catalyzed reaction (shown in Fig. 2) involves the overall loss of two water molecules. Various proposed mecha-nisms all include the abstraction of one or more protons from the substrate or an intermediate by a general base (e.g. Refs. 29 -31). This general base may be histidine; published pH rate profiles show a near-neutral pK a , which is consistent with a catalytically essential histidine (28) and diethylpyrocarbonate, a group-specific reagent which targets histidine, inactivates both porcine and maize PBGS (32,33). His 128 (numbered as in the E. coli sequence and shown in Fig. 1) is the only species invariant histidine in the PBGS sequence and thus is an attractive candidate for an essential histidine.
Histidine at the active site of PBGS might also serve as a Zn(II) ligand. Extended x-ray absorption fine structure spectroscopy studies of bovine PBGS indicate that Zn A (which binds directly to the substrate (25)) contains at least one imidazole ligand (26). Diethylpyrocarbonate inactivation of porcine PBGS can be protected against by the addition of Zn(II), suggesting that the modified histidine is a Zn(II) ligand (32). His 126 in the E. coli sequence has previously been proposed to be a ligand to Zn A (25). His 126 and His 128 may form the nucleation site for binding Zn A . Recently His 126 was proposed to be the key "switch" in determining a preference for binding Zn(II) (mammals and some bacteria) or Mg(II) (plants and some bacteria) to the region illustrated in Fig. 1. 4 To investigate whether His 126 or His 128 are involved in catalysis and/or Zn A ligation, we performed site-directed mutagenesis on E. coli PBGS to replace His 126 and His 128 with alanine, both in turn and simultaneously. Alanine was selected because the side chain methyl group of alanine is unable to serve either as a general base or as a Zn(II) ligand. A small side chain was chosen to have a minimal effect on protein folding. We report here the kinetic and other physical properties of the mutant proteins relative to wild type E. coli PBGS. These results also provide significant new information about wild type E. coli PBGS.

MATERIALS AND METHODS
Materials-ALA-HCl, potassium P i , bis-tris propane, p-(dimethylamino)benzaldehyde, and equine hemin (type III) were purchased from Sigma. ␤ME was purchased from Fluka Chemical Corp. and distilled under vacuum prior to use. HgCl 2 , ZnCl 2 , MgCl 2 (ultrapure), and high purity KOH were purchased from Aldrich. [4-13 C]ALA was customsynthesized by C/D/N Isotopes. Centrifree and Centriprep ultrafiltration devices were purchased from Amicon Corp. House distilled water was further purified by passage through a Milli-Q water purification system (Millipore). DNA plasmid purification kits were purchased from Qiagen. Oligonucleotides were synthesized in house by the Fox Chase Cancer Center Oligonucleotide Synthesis Facility. All other chemicals were reagent grade.
Assay, Kinetic Characterization, and pH Rate Profile Determination-PBGS activity assays as well as K m and V max determinations were as described previously (28). Most assay buffers contained 0.1 M potassium P i , pH 7.0, and 10 mM ␤ME. Except where noted, the assay also contained 10 M Zn(II) and 1 mM Mg(II). In most cases, the assay was started by the addition of ALA-HCl to a final concentration of 10 mM, which lowered the pH to 6.8. For determination of K m and V max , ALA was varied from 0.1 to 10 mM, and the final pH of the assay was controlled by adding HCl such that the total ALA-HCl plus HCl was equivalent to 10 mM. The buffer bis-tris propane was used for the pH profiles. Protein to be assayed in the absence of Mg(II) was prepared by passing twice through a Bio-Rad P6 spin column previously equilibrated with 0.1 M potassium P i , pH 7.0, 10 mM ␤ME, 10 M Zn(II) as per the manufacturer's directions. This procedure removes free Mg(II), but less than 1 equivalent of bound Mg(II) is retained. A 100 -500-fold dilution into an assay mixture, followed by a 10-min preincubation, releases this Mg(II) and results in free Mg(II) of 0.5-1 M. A unit of activity is defined as the production of 1 mol of porphobilinogen/h.

Construction of Plasmid pCR261f1 and Site-directed Mutagenesis-
The plasmid pCR261 was a kind gift of Dr. C. Roessner of Texas A & M, College Station, TX. This plasmid is pUC19 with a 1.6-kilobase NruI fragment containing the E. coli PBGS coding sequence inserted into the SmaI site. The Muta-Gene kit from Bio-Rad was used for site-directed mutagenesis and required insertion of the f1 origin into the plasmid pCR261 (34). We obtained a fragment containing the f1 origin flanked by EcoRI recognition sequences as a kind gift of John Taylor in Dr. G. D. Markham's laboratory in this institution. This fragment was excised with EcoRI and ligated into pCR261. The resulting plasmid (pCR261f1) was transformed into the E. coli strain CJ236. The uracil-containing single strand plasmid was isolated according to the protocol of the Muta-Gene kit. Second strand synthesis was performed using an oligonucleotide with the intended amino acid substitution as well as a silent, novel, and unambiguous restriction site (see Table I). After transformation of the double-stranded mutated plasmid into E. coli strains HB101 or XL1Blue, colonies were screened by performing restriction digests on plasmids purified from selected colonies. Mutant plasmids were sequenced by the Fox Chase Cancer Center DNA Sequencing Facility using an 18-base sequencing primer annealed to a site ϳ100 bases upstream of the mutation. The sequence confirmed the mutation and ensured that there were no unintentional sequence alterations near the mutation (ϳ400 bases).
Expression and Purification of Wild Type and Mutant PBGS-The E. coli strain CR261 was used to generate wild type E. coli PBGS as reported previously (28). The mutant plasmids were transformed into HU1000 (hemB, thr1, leuB6, thi1, lacY1, tonA21, supE44, Ϫ and F Ϫ ) (a 4 J. S. Ventura and M. P. Timko, personal communication. The four cysteines are proposed to be ligands to Zn B , whereas the tyrosine, aspartate, and/or histidine(s) of this region are potential ligands to Zn A (see text). The residues mutated in this study are underlined in the E. coli sequence, and the substitutions are illustrated directly above these residues in parentheses. The numerous aspartates in PBGS of eukaryotic photosynthetic organisms are proposed to bind Mg B (11).
FIG. 2. The PBGS-catalyzed reaction. The asymmetric condensation of two molecules of ALA involves removal of four non-ionizable hydrogen atoms and the two C-4 carbonyl oxygen atoms. A-side ALA loses both C-3 methylene protons, and P-side ALA loses one proton from the C-5 methylene group and one from the amine group. The detailed chemical mechanism remains a matter for speculation, but it is likely to involve acid-base chemistry at the active site (21).
kind gift of Dr. Charlotte Russell of City College New York), an E. coli strain that does not have a functioning PBGS, requires hemin for normal growth and is heme-permeable. After initial growth on LB ϩ hemin, complementation was determined by streaking individual colonies in parallel on LB plates with and without hemin (4 g/ml). In the absence of hemin, HU1000 cells grow as microcolonies at 37°C which appear ϳ36 h after plating. Under the same conditions, HU1000 cells transformed with a plasmid containing a functional copy of hemB grow as normal sized colonies at 37°C which appear ϳ18 h after plating. Purification of the mutant proteins was as described previously for wild type E. coli PBGS (28)  Wild type PBGS was eluted with 200 ml of 2 mM potassium P i , pH 6.8, 30 M Zn(II), 1 mM Mg(II), 10 mM ␤ME. Elution of the mutant proteins H126A and H126/128A required a wash of 200 ml of wild type elution buffer containing 5% ethylene glycol. The pooled protein was concentrated and applied to a Sephacryl S-300 column as described previously (28). Amino acid analysis to confirm the conversion of histidine to alanine was carried out by Dr. William Abrams of the University of Pennsylvania School of Dental Medicine. 13 C NMR Spectroscopy to Quantify the Active Site Stoichiometry-[4-13 C]ALA was added incrementally to a 1.6-ml solution containing 4.8 mol of E. coli PBGS subunits in 0.1 M bis-tris propane, pH 8, 10 mM ␤ME, 10 M ZnCl 2 , and 1 mM MgCl 2 . The stock [4-13 C]ALA was standardized by rapid quantitative PBGS-catalyzed conversion to [3,[5][6][7][8][9][10][11][12][13] C]porphobilinogen which was quantified using Ehrlich's reagent. 13 C NMR spectra were obtained on a Bruker AM300 spectrometer as described previously (28) and processed with a 20-Hz Lorenzian line broadening function. F is defined as the integral of a well resolved signal at 123 ppm arising from C-3 of free [3,5-13 C]porphobilinogen, indicative of one carbon of free porphobilinogen. The C-5 signal of free [3,5-13 C]porphobilinogen (120.5 ppm) and the C-3 signal of bound [3,[5][6][7][8][9][10][11][12][13] C]porphobilinogen (121 ppm) are not fully resolved, but together indicate one carbon of free plus one carbon of bound. The integral of this signal is denoted as T; T Ϫ F ϭ bound. Integrals were quantified by the cut and weigh method which is valid because C-3 and C-5 are both quaternary carbons with equivalent relaxation times and because free and bound porphobilinogen are in fast exchange (35). Because there is no measure of F when F Յ T/10, the data could not be fitted reliably to a binding curve.

RESULTS
General Characterization of the Mutants H126A, H128A, and H126/128A-Site-directed mutagenesis was used to generate the E. coli PBGS mutant proteins H126A, H128A, and H126/128A (see Fig. 1). All proteins (mutant and wild type) were able to complement the hemB Ϫ strain HU1000, as evidenced by normal growth of the transformed strains in media without supplementation by hemin. The mutant proteins are soluble and have the same quaternary structure as wild type PBGS, as evidenced by their behavior on a Sephacryl column. A summary of the protein purification of H126A is presented in    Table II. The purification characteristics of all mutants were similar. Only the low overall yield of activity (ϳ15%) was different from purification of wild type where our yield is typically 40 -50% (28). SDS-polyacrylamide gel electrophoresis of all proteins showed the purified mutant proteins to be homogenous and to migrate identically to wild type E. coli PBGS (data not shown). Total amino acid analysis was used to confirm the histidine content; the results are 6.15 Ϯ 0.15 for wild type, 5.0 Ϯ 0 for H128A, and 3.9 Ϯ 0.2 for H126/128A, each an average value of two independent preparations and very close to the expected values of 6, 5, and 4. Therefore, the observed catalytic activity is unlikely to be due to wild type protein present as the result of a recombination event during expression. H126A was not subjected to total amino acid analysis.

The Response of Wild Type and Mutant PBGS to Zn(II) and Mg(II)-
The specific activities of H128A, H126A, and H126/128A measured at 10 M Zn(II) and 1 mM Mg(II) were 56.4, 59.4, and 38.9, respectively. To test whether the conserved histidines are involved in Zn(II) binding, the response of enzyme activity to Zn(II) and Mg(II) were measured. Fig. 3A shows the effect of added Zn(II) on the activity of wild type and mutant proteins in the absence and presence of Mg C . Only the mutant protein H128A shows a significantly increased K d for Zn A Zn B relative to wild type PBGS. Under the conditions of Fig. 3A, in the absence of Mg C , all the mutant proteins show reduced activity relative to wild type PBGS. However, when Zn(II) is added in the presence of Mg C , the activities of the mutant proteins are close to or exceed (H126A and H128A) that of wild type.
We previously established that Mg C is an allosteric activator of wild type PBGS which increases the V max , reduces the K m for ALA, stabilizes the quaternary structure, and reduces the K d for Zn A Zn B (23). The multiple functions of Mg C are retained in the mutant proteins but the magnitudes of the effects vary. The kinetic parameters for mutant and wild type PBGS (determined at 10 M Zn(II), Ϯ 1 mM Mg(II) 5 ) are shown in Table III. Mg C causes V max to increase by 3.4 -13.2-fold and K m to decrease by 5-40-fold for the mutant proteins. Native polyacrylamide gel electrophoresis reveals similar stabilization of the quaternary structure of the mutant proteins by Mg C . Fig. 3B shows the effect of added Mg(II) on the activity of wild type and mutant PBGS and can be used to estimate the K d for Mg C . H126A, H128A, and H126/128A all have a higher K d for Mg C (0.4, 0.3, and 0.5 mM, respectively) than wild type PBGS (Ͻ0.1 mM).
pH Activity Relationships of H128A, H126/128A, and Wild Type PBGS-Histidine residues are frequently implicated to be responsible for near-neutral pK a values of enzymes. We determined the kinetic parameters K m and V max at various pH values for H128A, H126/128A, and wild type PBGS in the presence of 10 M Zn(II) and 1 mM Mg(II). 5 The results are illustrated in Fig. 4. The catalytic efficiency (k cat /K m ; Fig. 4A) FIG. 4. The effect of pH on the kinetic parameters of wild type PBGS, H128A, and H126/128A. q, wild type; f, H128A; å, H126/128A. A, k cat /K m (M Ϫ1 s Ϫ1 ) versus pH; B, K m (mM) versus pH (note y axis is log scale); C, V max (units/mg) versus pH. These experiments were performed as described under "Materials and Methods" for the determination of kinetic constants. The pH shown is measured after the addition of 10 mM ALA⅐HCl to the assay mix. Assays also contained 10 mM ␤ME, 10 M Zn(II), and 1 mM Mg(II). shows a pK a of ϳ7.5 which is maintained in H128A and H126/128A. This pK a arises from the pH dependence of the K m (Fig. 4B) and is the same in all proteins studied. In contrast to the pK a effect, the K m values for the mutant proteins differ by more than an order of magnitude from wild type PBGS; V max for H128A is reduced about 30% from wild type PBGS at neutral pH, and the V max for H126/128A is reduced about 50% from wild type at alkaline pH (Fig. 4C). These results indicate that neither His 126 nor His 128 are involved in the ratedetermining step of the enzyme-catalyzed reaction, but implicates them in substrate binding. The Stoichiometry of E. coli PBGS Active Sites-The high K m of ALA for E. coli PBGS at neutral pH in the absence of Mg C (ϳ3 mM) previously prevented our quantifying the stoichiometry of enzyme active sites (28). Although it is well established that mammalian PBGS has four active sites per octamer, the stoichiometry of active sites on E. coli PBGS is a matter of some controversy (36). Herein we document that the K m for ALA at pH 8 in the presence of saturating Zn(II) and Mg(II) is tight enough (ϳ50 M) that 13 C NMR spectroscopy can be used to quantify the active site stoichiometry. Fig. 5 illustrates the integrated signal area of free and bound [3,[5][6][7][8][9][10][11][12][13]  Temperature Dependence of H126/128A and Wild Type PBGS-We previously reported an E a (energy of activation) of 64.2 KJ mol Ϫ1 for E. coli PBGS in the absence of Mg C (28). This is similar to the bovine PBGS E a of 77.1 KJ mol Ϫ1 (37); bovine PBGS is unresponsive to Mg(II). Fig. 6 shows Arrhenius plots of the temperature dependence for H126/128A and wild type PBGS, obtained in the presence of 1 mM Mg(II). The Arrhenius plot of H126/128A in the presence of Mg(II) is clearly biphasic. Below 35°C E a is 26.5 KJ mol Ϫ1 and above 35°C E a is 2.1 KJ mol Ϫ1 , 13-fold less than E a at lower temperatures. This result implies a change in the rate-determining step at 35°C or a change in the K d for Zn(II) or Mg(II). The biphasic character of the data for wild type E. coli PBGS is less pronounced. Below 35°C E a is 19.7 KJ mol Ϫ1 , above 35°C E a is 2.1 KJ mol Ϫ1 . The 50°C data point for both plots is not included in the calculation of E a and may reflect a second transition. One surprising observation is that at 37°C Mg C causes an 11-fold decrease in E a for wild type PBGS; the magnitude of this effect is in sharp contrast to the Mg C induced 2-fold increase in V max . This suggests that Mg C causes a change in the rate-determining step of the reaction. DISCUSSION Site-directed mutagenesis has been used to probe the roles of one species-invariant histidine His 128 of E. coli PBGS and its partially conserved neighbor His 126 . Previous chemical modification data had suggested an essential histidine (32,33). The most striking result is the retained ability of the mutant proteins H126A, H128A, and H126/128A to catalyze the formation of porphobilinogen, indicating that these histidines are not essential for catalysis. Under approximately physiologic conditions the V max and K m values for the mutant proteins vary little from wild type E. coli PBGS, suggesting that the rate-determining step(s) of the reaction are unaltered by the mutation. Thus, one may ask why these histidines are so highly conserved. Perhaps these residues are involved in a step in catalysis whose rate is several orders of magnitude faster than the rate-determining step; the overall rate of PBGS is slow (k cat is ϳ1 s Ϫ1 ). The PBGS-catalyzed reaction includes several essential steps. P-side ALA (see Fig. 2) must bind and form a Schiff base to Lys 246 . Zn A must bind to facilitate binding of A-side ALA. Beyond catalyzing formation of these early complexes, the protein may play a passive role, simply providing the scaffolding for binding substrates and metal ions. The conserved residues may contribute only secondary interactions to catalysis.
Mutation of His 126 and/or His 128 to alanine perturbs the K m for ALA; the K m is determined by the more loosely bound of the two identical substrates, the ALA known to require Zn A in order to bind at the active site (25,38). This increased K m is consistent with our active site model because the increase may be due to altered affinity for Zn A or to alteration of the A-side ALA binding pocket. Despite the value of K m for the mutant proteins, they maintain the pK a of 7.5 in the k cat /K m profile. This pK a cannot be due to His 126 nor His 128 ; the possibility that this pK a arises from another histidine is diminished by the fact that some PBGS (e.g. Bradyrhizobium japonicum PBGS) contain only these two histidines. Since this pK a must be related to substrate binding, it may help verify one of two published models for A-side ALA binding to Zn A , shown in Fig. 7. One model shows Zn A bound through the carbonyl oxygen of ALA (Fig. 7A) (39), whereas an earlier model shows bidentate ligation through both the amino group and the carbonyl oxygen ( Fig. 7B) (25,40). The bidentate model arises in part from 13 C and 15 N NMR spectra of the protein-product complex which showed the amino group of bound porphobilinogen to be deprotonated (40). The pK a of 7.5 may reflect deprotonation of the amino group of enzyme-bound ALA which would be necessary for bidentate chelation. However the pK a for deprotonation of the amino group of free ALA is 8.4.
His 126 and/or His 128 have been introduced as potential Zn(II) ligands. A review of the literature indicates that the mutation of Zn(II) ligands can have different effects on different proteins. For bacteriophage T4 gene 32 protein, mutation of the Zn(II) binding residues Cys 87 or His 64 result in a mutant protein which does not bind Zn(II) or single-stranded DNA and shows some structural instability in the Zn(II) binding region (41). However, mutation to a Zn(II) ligand of mammalian sorbitol dehydrogenase (42) or E. coli methionyl-tRNA synthetase (43) results in a reduced affinity for Zn(II) which can be overcome by adding excess Zn(II) to the assay. This is like the H128A mutant of PBGS. It appears that the removal of one ligand of four or five may be insufficient to eliminate metal binding because Zn(II) can accommodate a variable number of ligands equally well. Recent ab initio calculations of Bock et al. (19) show little energetic differences among complexes with Zn(II) bound to four, five, or six first coordination sphere water molecules. Although removal of one or two ligands may not disrupt Zn(II) binding, simultaneous mutation of two neighboring ligands may have a dramatic effect. Metalloproteins often contain a nucleation site for metal ion binding composed of two amino acids separated by three or fewer intervening amino acids in the primary structure (15). In E. coli PBGS, His 126 and His 128 comprise one possible Zn(II) nucleation site. The affinity of H126/128A for Zn(II) indicates that these two histidines do not form the nucleation site for Zn(II) binding.
PBGS from some species have been reported to require Mg(II) but not Zn(II) (e.g. Ref. 17). Our current results raise the question whether the Mg C effect can be so large as to mask a requirement for Zn(II). For instance, Mg C has a 21-fold effect on V/K of wild type PBGS and a 134-fold effect for H126A. Thus, a relatively minor change in the PBGS sequence can dramatically affect the magnitude of the allosteric effect of Mg C . In addition, Mg C enhances the binding affinity for Zn A Zn B , amplifying the apparent allosteric effect at low Zn(II).
Thus, care must be taken in evaluating the metal ion requirements for PBGS from all species.