Characterization and Implications of Ca2+Binding to Pectate Lyase C*

Ca2+ is essential for in vitro activity of Erwinia chrysanthemi pectate lyase C (PelC). Crystallographic analyses of 11 PelC-Ca2+complexes, formed at pH 4.5, 9.5, and 11.2 under varying Ca2+ concentrations, have been solved and refined at a resolution of 2.2 Å. The Ca2+ site represents a new motif for Ca2+, consisting primarily of β-turns and β-strands. The principal differences between PelC and the PelC-Ca2+ structures at all pH values are the side-chain conformations of Asp-129 and Glu-166 as well as the occupancies of four water molecules. According to calculations of pKa values, the presence of Ca2+and associated structural changes lower the pKa of Arg-218, the amino acid responsible for proton abstraction during catalysis. The Ca2+ affinity for PelC is weak, as theK d was estimated to be 0.132 (±0.004) mm at pH 9.5, 1.09 (±0.29) mm at pH 11.2, and 5.84 (±0.41) mm at pH 4.5 from x-ray diffraction studies and 0.133 (±0.045) mm at pH 9.5 from intrinsic tryptophan fluorescence measurements. Given the pH dependence of Ca2+affinity, PelC activity at pH 4.5 has been reexamined. At saturating Ca2+ concentrations, PelC activity increases 10-fold at pH 4.5 but is less than 1% of maximal activity at pH 9.5. Taken together, the studies suggest that the primary Ca2+ ion in PelC has multiple functions.

the plant cell wall, causing tissue damage and cell rupture. Erwinia bacterial strains, which are deficient in one or more pectate lyase genes, have decreased virulence (3)(4)(5). Mutations affecting the extracellular secretion of the pectic enzymes also reduce virulence (6). The introduction of pectate lyase genes into Escherichia coli enables E. coli to macerate plant tissue and display soft rot-like symptoms (7). Highly purified pectate lyase preparations macerate plant tissue, thus confirming a causal relationship between the enzymes and soft rot diseases (8).
Pectate lyases constitute a family of isozymes that share 29 -91% sequence similarity (9). All isozymes cleave ␣-1,4linked galacturonic acid units of the pectate component of the plant cell wall by a ␤-elimination mechanism. In vitro catalyzed reactions have a bell-shaped pH profile, which extends from 6 to 10 with a pH optimum near 9.0 for most isozymes (10). Pectate lyases require Ca 2ϩ for in vitro activity and presumably utilize the abundant Ca 2ϩ in the plant cell wall for in vivo activity (11). Although Ca 2ϩ once was assumed to bind only the substrate (12), now it is known that Ca 2ϩ binds directly to the protein. Several three-dimensional structures of pectate lyases have been determined in the presence and absence of Ca 2ϩ (13)(14)(15)(16). All such structures reveal a novel parallel ␤-helical topology that probably is shared by all family members. The structures of Bacillus subtilis pectate lyase (15) and Bacillus sp. strain KSM-P15 pectate lyase (16) reveal details of a Ca 2ϩ binding site at the same relative location in each structure. Another structure, that of a complex between an inactive PelC 1 mutant and a pentagalacturonate substrate, has revealed three additional Ca 2ϩ sites linking the substrate to the protein (17). Although the PelC R218K-(Ca 2ϩ ) 4 -pentaGalpA structure provides many insights into the pectate lyase cleavage mechanism, the structure complicates questions regarding the function of the Ca 2ϩ ions. Are all Ca 2ϩ ions essential for in vitro catalytic activity? Are the Ca 2ϩ ions only required for substrate binding, or do any have an active catalytic role? If so, which Ca 2ϩ ion has an active catalytic role, and how might the ion participate in catalysis? Is the primary Ca 2ϩ ion involved in the polygalacturonic acid hydrolase activity that is observed in some pectate lyases (18,19)? In view of the observations that the pH of the plant cell wall increases as the pectate component is degraded (20), how does pH affect the affinity of the Ca 2ϩ ions for PelC? To address these questions, various studies with Ca 2ϩ have been initiated. The results presented herein explore the binding of the primary Ca 2ϩ ion to PelC over a broad pH range. * This work was supported by United States Department of Agriculture Grant 02-03560. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The

EXPERIMENTAL PROCEDURES
Crystal Preparation-PelC was purified and crystallized from ammonium sulfate according to published protocols (21). Unless noted, all reagents were purchased from Sigma. The PelC-Ca 2ϩ complexes were prepared by a gradual reduction in ammonium sulfate and a concurrent increase in polyethylene glycol 8000 surrounding the native crystals. Subsequently, Ca 2ϩ was added gradually. The crystals were sensitive to the order and rate of the changes made to the mother liquor, as illustrated in Fig. 1. PelC crystals were prepared at three pH values: 4.5, 9.5, and 11.2. At pH 4.5 and 11.2, the reactions were buffered by 100 mM ␥-aminobutyric acid. At pH 9.5, 100 mM amino-2-methyl-1,3-propanediol buffer was used. At each pH, crystals were prepared in the absence of Ca 2ϩ or at Ca 2ϩ concentrations of 1, 5, 10, 20, or 30 mM.
Data Collection and Processing-X-ray diffraction data were collected at 20°C using a dual multiwire area detector system (San Diego Multiwire Systems, San Diego, CA). The data were processed by the San Diego Multiwire software package (22) and placed on an absolute scale by the Wilson method (23). The x-ray data statistics are summarized in Table I.
Model Refinement-Initial phases were derived from the native PelC structure determined at pH 7.5 in the absence of Ca 2ϩ (24). The reflection data were divided randomly into two sets, a working set composed of 90% of the data and a test set composed of the remaining data used for cross-validation of the refinement cycles (25). The parameter and topology files of Engh and Huber were used (26). The weight of the crystallographic term in the molecular dynamics was calculated from the empirical check procedure of X-PLOR (27) and kept constant at 50% throughout the refinement process.
The Ca 2ϩ ion for each data set was confirmed by difference Fourier methods. Each model was refined by using simulated annealing methods as implemented by X-PLOR (26). Alterations within a 10-Å region around the Ca 2ϩ were identified by simulated annealed OMIT electron density maps and refit manually using O (28) and OOPS (29). Following Powell minimization and temperature factor refinement, water molecules were assigned to peaks Ͼ 4 using MAPMAN (30) and verified visually or by using reasonable distance and geometry criteria. In the final stages, alternate cycles of occupancy and temperature factor refinement of Ca 2ϩ and selected atoms were carried out in X-PLOR until convergence was reached.
To estimate the error in the Ca 2ϩ occupancy, each of the 11 final models was refined by energy minimization using 15 different values of Ca 2ϩ occupancy ranging from 0.1 to 1.5. The total energy for each model was plotted against the occupancy of the Ca 2ϩ ion, and a nonlinear regression analysis was carried out using SigmaPlot 5.0. Each plot was fitted using a four-parameter gaussian curve, and all had R 2 values exceeding 0.99. The final Ca 2ϩ occupancy was assumed to be the value corresponding to the energy minimum of each plot, and the estimated error for Ca 2ϩ occupancy was obtained from the regression analysis. The final refinement statistics for each model are summarized in Table I.
Model Analysis-Crystallographic models were checked with PRO-CHECK (31) for geometric and crystallographic statistics. The superposition of ␣-carbons for each model was calculated and refined using least squares methods implemented by LSQMAN (32). All model figures were prepared using SETOR (33).
Estimation of pK a Values-The pK a values of individual amino acids were estimated using the computer program MEAD (34) and the atomic coordinates of the PelC structures at pH 9.5 in the absence and presence of 30 mM Ca 2ϩ . All water molecules were included in the calculations. The atoms were assigned partial charges and atomic radii using the standard CHARMM parameters (35). Interior dielectric constants of 4, 20, 30, 35, or 40 were used in the calculations.
Apparent Dissociation Constant-The Ca 2ϩ occupancy for each x-ray diffraction data set at a given pH was plotted using SigmaPlot 5.0. The data were fit to a single rectangular, two-parameter hyperbolic curve using regression analysis. The apparent dissociation constant (K d ) of Ca 2ϩ from PelC at each pH was extrapolated subsequently from the 50% occupancy value.
Intrinsic Tryptophan Fluorescence Assay-The apparent K d of Ca 2ϩ from PelC at pH 9.5 also was determined by changes in the intrinsic tryptophan fluorescence. The fluorescent changes were monitored on a SPEX 112 fluorometer with a 150-watt Xenon Arc as the source at an excitation wavelength of 284 nm and over an emission wavelength range of 300 -400 nm at a 2-nm interval. Water-Raman spectra were collected before each measurement for experimental consistency. The integral of each emission spectrum was used to express the fluorescence intensity of PelC at each Ca 2ϩ concentration after correcting for dilution resulting from the titrant added. For the titration of PelC, an appropriate concentration of Ca 2ϩ was added to a 2-ml fluorescence cuvette containing 0.088 mM PelC in 5 mM bis-tris propane, pH 9.5, to give a final concentration of Ca 2ϩ in the solution ranging from 10 Ϫ9 to 10 Ϫ1 M. On addition of the titrant, the sample was incubated for 2 min to ensure that equilibrium had been attained before an emission spectrum was collected. 14 fluorescence measurements were carried out at room temperature. The apparent K d value was defined as the Ca 2ϩ concentration that elicited one-half of the maximum change in fluorescence, which was determined by curve fitting with nonlinear regression to a sigmoidal dose-response curve using Prism 3.0 (GraphPad Software, San Diego, CA).
Pectolytic Activity Assay-The optimal Ca 2ϩ concentration was determined at each pH using the standard spectroscopic method for determining pectate lyase activity (8). At pH 4.5 and 11.2, the reactions were buffered with 100 mM ␥-aminobutyric acid. At pH 9.5, 100 mM amino-2-methyl-1,3-propanediol buffer was used. The selected buffer, varying concentrations of CaCl 2 , and 100 l of a 1% (w/v) sodium polypectate stock were premixed in a total volume of 990 l. The tested Ca 2ϩ concentrations ranged from 0.05 to 2 mM. To initiate the reaction, 10 l of a PelC stock (26 nM) were added and mixed. Enzymatic activity was measured by recording the absorbance increase at 232 nm every 10 s at 22°C. One unit of pectate lyase activity is defined as the production of 1 mmol of unsaturated product/min. The formation of 1 mmol of unsaturated uronide/min was taken to correspond to 1.73 units/min (36).

RESULTS
Structural Models-11 structural models of PelC have been constructed and refined using the x-ray diffraction data. Two PelC models have been refined in the absence of Ca 2ϩ at pH 4.5 and 9.5. The root mean square deviation between the ␣-carbons of the two models is 0.13 Å, indicating insignificant changes in the PelC structure over a broad pH range. Additional structures of PelC have been refined in the presence of 5, 20, and 30 mM Ca 2ϩ at pH 4.5, in the presence of 0.3, 5, and 30 mM Ca 2ϩ at pH 9.5, and in the presence of 1, 5, and 30 mM Ca 2ϩ at pH 11.2. In these models, the Ca 2ϩ occupancy ranges from 38 to 92% with the highest occupancy found at the 30 mM Ca 2ϩ concentration, pH 9.5. There are only two significant structural changes coinciding with the presence or absence of Ca 2ϩ . Major alterations are observed in the conformation of Asp-129 and in the occupancy of four water molecules near the Ca 2ϩ binding site. Consequently, in those models with a partial Ca 2ϩ occupancy, the occupancy of each Asp-129 conformation and the occupancies of the water molecules also have been refined. Essentially, each refined model is a composite of two partial models, one representing the PelC structure in the absence of Ca 2ϩ and the other representing the structure in the presence of Ca 2ϩ . The refinement statistics for all models are summarized in Table I.
Details of the Ca 2ϩ Site-The description of the Ca 2ϩ site is based on atomic details obtained from the PelC-Ca 2ϩ structure with the highest Ca 2ϩ occupancy; this is the model at pH 9.5 and 30 mM Ca 2ϩ . The details of the Ca 2ϩ site are compared with the same region in the absence of Ca 2ϩ (Fig. 2). The Ca 2ϩ site is located at the same position as the Lu 3ϩ derivative of PelC (13) and is similar to the Ca 2ϩ site observed in B. subtilis pectate lyase (15). Seven oxygen atoms form, the standard bipentagonal arrangement around the Ca 2ϩ ion with an average Ca 2ϩ -oxygen bond distance of 2.46 Å. The individual Ca 2ϩoxygen bond distances are listed in Table II. Two of the coordinated oxygen atoms derive from water molecules, Wat-1 and Wat-2. Five oxygen atoms originate from amino acid side chains. Each oxygen of the carboxylic group of Asp-131 is coordinated to the Ca 2ϩ ion, whereas only one carboxylic oxygen of Asp-129, Glu-166, and Asp-170 is coordinated. Asp-131 is invariant throughout the pectate lyase superfamily (9,37). Asp-129, Glu-166, and Asp-170 are conserved only in the pectate lyase and plant pollen subfamilies.
Ca 2ϩ -induced Structural Changes-The presence of a highly occupied Ca 2ϩ ion does not cause any major structural change in the native PelC structure. The root mean square deviation of the ␣-carbons in the PelC structures at pH 9.5, in both the presence and the absence of Ca 2ϩ , is 0.17 Å. The observed differences include two side-chain movements in the area of the Ca 2ϩ binding site (Fig. 2). Upon Ca 2ϩ binding, the side chain of Asp-129 makes an 89°torsional rotation around the -1 bond. This rotation moves the carboxylic oxygen of Asp-129 closer to the Ca 2ϩ ion, resulting in a positional displacement of over 2 Å. The refined occupancies of Ca 2ϩ and the two conformations of Asp-129 are summarized in Table III. Glu-166 makes a rotational change of 58°along the -3 torsional bond, but the movement is sufficiently small as to preclude the refinement of an alternate conformation. Nevertheless, the movement of Glu-166 is critical for the establishment of an extensive network of hydrogen bonds and electrostatic interactions in the active site region. In the presence of Ca 2ϩ , Glu-166 is locked into a singular position by the coordination of one carboxylic oxygen to Ca 2ϩ and the other oxygen, with a strong electrostatic interaction with Lys-190 at a distance of 2.7 Å. In the absence of Ca 2ϩ , Glu-166 does not form strong interactions with neighboring atoms.
Changes in the water structure also are observed. In the absence of Ca 2ϩ , there are four water molecules in the region, Wat-1, Wat-2, Wat-3, and Wat-4. The position and occupancy of Wat-1 and Wat-4 are independent of the presence or absence of Ca 2ϩ . Wat-1 coordinates to Ca 2ϩ when the ion is present but does not alter its position when the cation is absent. Similarly, Wat-4 forms a hydrogen bond with the guanidinium nitrogen of Arg-218 irrespective of the Ca 2ϩ occupancy. Wat-2 forms hy-  Arg-218 to any other atom, does not exist in the absence of Ca 2ϩ . Wat-3 also shifts position depending upon the Ca 2ϩ occupancy. Wat-3 is displaced by the carboxylic oxygen of Asp-129 when the residue rotates into the coordination sphere of the Ca 2ϩ ion. Finally, two new water molecules, Wat-6 and Wat-7, appear, forming hydrogen bonds with water molecules in the region. Overall, there is a net gain of two water molecules within an 8-Å sphere around the Ca 2ϩ . Effect of Ca 2ϩ on the pK a of Arg-218 -In the pectate lyase superfamily, an invariant arginine analogous to Arg-218 in PelC is believed to initiate the abstraction of the proton in the initial stages of the ␤-elimination reaction. The pK a of Arg-218 in the PelC structures at pH 9.5, in the presence (30 mM) and absence of Ca 2ϩ , has been estimated by macroscopic electrostatic calculations using different values of the dielectric constant. Using a dielectric constant of 20, the pK a of Arg-218 is estimated to be 13.0 in the absence of Ca 2ϩ and 11.60 in the presence of Ca 2ϩ . Although the value of the dielectric constant influences the absolute pK a value, the relative shifts in the pK a are qualitatively similar. At all dielectric constant values, the presence of Ca 2ϩ lowers the pK a of Arg-218, suggesting that the primary Ca 2ϩ ion plays a role in priming the catalytic arginine for the proton abstraction step.
PelC Affinity for Ca 2ϩ -The Ca 2ϩ occupancy in each of the PelC-Ca 2ϩ complexes shown in Table III has been plotted as a function of Ca 2ϩ concentration at each pH value. The occupancy data have been fit to a single rectangular, two-parameter hyperbolic curve using regression analysis. The Ca 2ϩ concentration, which corresponds to the 50% occupancy level of the fitted curve, is an estimate of the Ca 2ϩ affinity for the protein.
The apparent K d of the PelC-Ca 2ϩ complex is estimated to be 0.132 (Ϯ0.004) mM at pH 9.5, 1.09 (Ϯ0.29) mM at pH 11.2, and 5.84 (Ϯ0.41) mM at pH 4.5 (Fig. 3). The apparent K d at pH 9.5 has been verified by intrinsic tryptophan fluorescence techniques. The addition of Ca 2ϩ to native PelC induces a small but measurable change in the intensity of the emitted fluorescence but does not shift the wavelength of the emitted light. The apparent K d from the fluorescence data is estimated from the titration curve (Fig. 4). At pH 9.5, the K d of 0.133 (Ϯ0.045) mM estimated from fluorescence methods is in close agreement with a K d of 0.132 (Ϯ0.004) mM measured from the x-ray diffraction analyses of the Ca 2ϩ occupancy. At all pH values, the PelC affinity for Ca 2ϩ is relatively weak.
Optimal Ca 2ϩ Concentration for Enzymatic Activity-The pH profiles of all pectate lyases are determined routinely at Ca 2ϩ concentrations between 0.1 and 0.4 mM. Because the PelC affinity for Ca 2ϩ at pH 4.5 is weak, a higher concentration of Ca 2ϩ may be needed to activate PelC at the lower pH range. The enzymatic activity of PelC at pH 4.5 has been determined in the presence of various Ca 2ϩ concentrations ranging from 2.5 to 40 mM. The data demonstrate that the enzymatic activity at pH 4.5 is greatest at 30 mM Ca 2ϩ (Fig. 5). Saturation of the Ca 2ϩ binding site increases the enzymatic activity by 10-fold. Nevertheless, at pH 4.5, PelC still has very low activity, less than 1% of the maximal activity measured at pH 9.5.

DISCUSSION
The primary Ca 2ϩ site of PelC shares the same location and coordination as the primary Ca 2ϩ site for the heavy atom derivative Lu 3ϩ , which is used to phase the initial structure. The Ca 2ϩ coordination involves three amino acids and two water molecules arranged in the typical bipentagonal coordination. The primary Ca 2ϩ ion has been observed in two other pectate lyase structures, that of B. subtilis pectate lyase (15) and Bacillus sp. strain KSM-P15 pectate lyase (16). Although the location and coordination numbers are the same, there are small differences between the Ca 2ϩ binding sites. Asp-129 in PelC corresponds to Gln-182 in B. subtilis pectate lyase. As an uncharged residue, Gln-182 is unable to coordinate to the Ca 2ϩ   FIG. 3. Titration of the occupancy of the primary Ca 2؉ site in PelC. The refined Ca 2ϩ occupancies in the PelC-Ca 2ϩ complexes are plotted as a function of the Ca 2ϩ concentration diffused into crystals at pH 9.5 (f), 11.2 (q), and 4.5 (OE). The occupancy data were fit to a single hyperbolic curve using regression analysis. The apparent K d is estimated to be 0.132 mM at pH 9.5, 1.09 mM at pH 11.2, and 5.84 mM at pH 4.5. For many years, Ca 2ϩ was believed to interact only with the substrate in order to induce a substrate conformation that could be recognized by the pectate lyases (12). Therefore, it is not surprising to find that PelC-Ca 2ϩ has an apparent K d in the tenth millimolar range even at the optimal pH for enzymatic activity. Nevertheless, the weak association between PelC and Ca 2ϩ is not so weak as to be insignificant, particularly in the context of the plant cell wall, where Ca 2ϩ is believed to be abundant (11). The actual concentration of Ca 2ϩ in the plant cell is difficult to determine because Ca 2ϩ is distributed differentially in various areas of the plant tissue and is dependent upon plant type and stage of growth (38). One estimate for soybean hypocotyl cell walls places the Ca 2ϩ concentration between 10 and 50 mM (39). The x-ray diffraction method is somewhat unconventional for measuring Ca 2ϩ affinity, but it has been used previously in the quantitation of Ca 2ϩ binding to internalin B (40). Moreover, the K d at pH 9.5 determined by the x-ray diffraction method has been corroborated with intrinsic tryptophan fluorescence measurements. The present research does not address whether the affinity of Ca 2ϩ for PelC increases in the presence of the substrate. Such a question is difficult to answer because additional Ca 2ϩ ions are bound at the PelC-oligosaccharide interface (17).
Although the x-ray diffraction method has many drawbacks over conventional methods for determining Ca 2ϩ affinity for a protein, the approach offers one significant advantage in that the actual Ca 2ϩ ion for which K d is being measured can be visualized. The diffraction studies of the PelC R218-(Ca 2ϩ ) 4 -pentaGalpA complex as well as preliminary fluorescence data indicating a negative Hill coefficient raised the possibility that native PelC might have more than one Ca 2ϩ site, each with differing affinities, at exogenous Ca 2ϩ concentrations of 10 mM or less. No positive electron density, consistent with partially occupied Ca 2ϩ ions, is present at any other site at which the exogenous Ca 2ϩ concentration is 30 mM or less. This observation is consistent with the R218-(Ca 2ϩ ) 4 -pentaGalpA complex structure in which the other Ca 2ϩ ions coordinate to only one or two amino acids, suggesting that the affinity of these Ca 2ϩ ions for the protein may be very weak (17).
The Ca 2ϩ affinity for PelC is highest at pH 9.5, the optimal pH for in vitro enzymatic activity. The Ca 2ϩ affinity decreases 10-fold at pH 11.2, possibly as a consequence of the affinity of Ca 2ϩ for water. At pH 11.2, there is a large population of hydroxyl ions that will compete effectively with the amino acids for direct coordination to Ca 2ϩ . The nearly 100-fold reduction in Ca 2ϩ affinity for PelC at pH 4.5 can be explained by the increased protonation of the aspartic and glutamic acid groups that otherwise would coordinate with the Ca 2ϩ ion. The detection of any Ca 2ϩ affinity at pH 4.5 is somewhat surprising, given that in vitro pectate lyase activity had not been detected previously under standard assay conditions. However, the standard assays routinely employ Ca 2ϩ concentrations that are too low at pH 4.5 to saturate the Ca 2ϩ site as deduced by the x-ray diffraction analyses. This observation raises the question as to what extent pectate lyase activity at pH 4.5 is dependent upon the Ca 2ϩ concentration. To answer this question, we have determined pectate lyase activity in the presence of higher Ca 2ϩ concentrations and found that saturation of the Ca 2ϩ binding site increases the enzymatic activity by 10-fold. Nevertheless, the activity of PelC at pH 4.5 still is very low, as it is less than 1% of the maximal activity measured at pH 9.5. Therefore, the low pectate lyase activity at pH 4.5 must be dependent upon additional factors.
The Ca 2ϩ binding motif found in PelC differs from the high affinity Ca 2ϩ sites found in EF-hand Ca 2ϩ -binding proteins (41)(42)(43)(44)(45). The typical EF-hand Ca 2ϩ binding site consists of a continuous or discontinuous helix-loop-helix motif on which the coordinating amino acids are located. In the continuous type (Fig. 6A), the helix-loop-helix motif is composed of a contiguous sequence of 12 amino acids. In the discontinuous type (Fig. 6B), the helix-loop-helix is composed of two noncontiguous sequences of amino acids. In contrast to the typical EF-hand Ca 2ϩ binding motif, the Ca 2ϩ binding pocket on PelC is composed entirely of ␤-strands and ␤-turns (Fig. 6C). The coordinating amino acids are located on two adjacent rungs of the parallel ␤-helix. Three of the four amino acids that coordinate the Ca 2ϩ ion are located on ␤-turns that connect two ␤-strands. Only one amino acid, Asp-170, is located within a ␤-strand. Despite the differences in the surrounding topology of the Ca 2ϩ sites, the coordination around the Ca 2ϩ ion in PelC appears to be identical in number and average bond distance to typical helix-loop-helix proteins that have much higher Ca 2ϩ affinities.
Why then does PelC bind Ca 2ϩ weakly, whereas proteins with a helix-loop-helix motif bind Ca 2ϩ with high affinity? In reviews of Ca 2ϩ -binding proteins (44 -46), no one has been able to find a direct correlation between the metal ion binding site and its affinity for the protein. Pidcock and Moore (46) observed other features that may affect the strength of binding. Among these features are the nature of the electrostatic interactions, including extensive hydrogen bond networks, and large metal ion-induced protein conformational changes. As discussed previously, PelC has an extensive hydrogen bond network that is altered in the presence of Ca 2ϩ . The Ca 2ϩ -induced conformational changes are relatively minor in PelC and less likely to influence the affinity of Ca 2ϩ . We propose that the answer for PelC may lie in the orientation of the Ca 2ϩ toward the lone pairs of electrons on the carboxylate oxygen. High affinity for a Ca 2ϩ ion usually is found when Ca 2ϩ is coordinated directly to both oxygen atoms or to the syn conformation of the lone electron pair on one of the carboxylate oxygens (47). In PelC, the carboxylate oxygen atoms of two amino acids, Glu-166 and Asp-170, coordinate to Ca 2ϩ in the anti conformation of the lone electron pair. Asp-131 forms a direct interaction with the Ca 2ϩ , and only Asp-129 binds in the syn conformation. Thus one possible explanation for the weaker Ca 2ϩ affinity in PelC may be the weak coordination to the anti conformation in two of the four coordinating amino acids.
In the presence of a substrate, PelC is known to bind to three additional Ca 2ϩ ions, each of which has fewer coordinating ligands to the protein. None of these Ca 2ϩ ions are observed in the PelC structure in the absence of a substrate. The presence of only one Ca 2ϩ , which coordinates more tightly than the other sites through an invariant and several highly conserved amino acids, suggests that this Ca 2ϩ has a special role. One role, suggested by the present study, is to reduce the pK a of Arg-218 so that the amino acid is better suited chemically to act as a proton acceptor in the ␤-elimination reaction mechanism. Another possibility, suggested by the PelC R218K-(Ca 2ϩ ) 4 -penta-GalpA complex, is that this Ca 2ϩ is critical for aligning the substrate properly in the active site. Other studies suggest that the occupancy of the primary Ca 2ϩ site determines whether PelC functions as a hydrolase or a lyase (48). When the structure of the PelC-Ca 2ϩ complex is superimposed upon the polygalacturonase structures (48), the PelC amino acids coordinated to the primary Ca 2ϩ ion are analogous to the catalytic residues in polygalacturonases. The PelC amino acids are in the proper orientation to carry out hydrolase activity but only in the absence of the Ca 2ϩ ion. Thus the weak affinity of PelC for the primary Ca 2ϩ ion, as determined herein, is critical for understanding how Ca 2ϩ can regulate the function of PelC.
Although structure-function studies have focused primarily on microbial pectate lyases, pel genes have been identified in a wide range of plants with at least 25 isozymes found by gene annotation in Arabidopsis thaliana. 2 Because the plant cell wall pH is ϳ5.75, it has been difficult to understand how any pectate lyase functions in such an environment. The present study indicates that pectate lyases are active even at a low pH if the exogenous Ca 2ϩ concentration is sufficiently high. If the Ca 2ϩ concentration is low, other studies suggest that the enzymes function as hydrolases. The occupancy of the primary Ca 2ϩ site appears to determine whether the pectate lyase functions as a hydrolase or a lyase. Although additional studies are needed to confirm this hypothesis, the weak affinity of the primary Ca 2ϩ ion for PelC observed herein is a necessary feature for the Ca 2ϩ ion to serve as a regulatory switch.
FIG. 6. Stereoviews of three types of Ca 2؉ binding motifs. A, the Ca 2ϩ region of ␣-lactalbumin exemplifies the continuous helix-loophelix Ca 2ϩ binding motif, which is associated with tight binding. B, the Ca 2ϩ region of phospholipase A2 typifies the discontinuous helix-loophelix Ca 2ϩ binding motif. C, the ␤-turn/␤-strand Ca 2ϩ binding motif of PelC exhibits weak binding affinity. In all figures, the polypeptide backbone of the protein is represented as a green ribbon, the Ca 2ϩ ions as yellow spheres, the oxygen atoms as red rods, and a disulfide bridge as a yellow bar.