Crystal Structure of Bacillus Subtilis CodW, a Noncanonical HslV-like Peptidase with an Impaired Catalytic Apparatus.

Proteins: Structure, Function, and BioinformaticsVolume 71, Issue 2 p. 1020-1026 Structure NoteFree Access Crystal structure of Bacillus subtilis CodW, a noncanonical HslV-like peptidase with an impaired catalytic apparatus Seong-Hwan Rho, Seong-Hwan Rho Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea Seong-Hwan Rho and Hyun Ho Park contributed equally to this work.Search for more papers by this authorHyun Ho Park, Hyun Ho Park Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea Seong-Hwan Rho and Hyun Ho Park contributed equally to this work.Search for more papers by this authorGil Bu Kang, Gil Bu Kang Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaSearch for more papers by this authorYoung Jun Im, Young Jun Im Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaSearch for more papers by this authorMin Suk Kang, Min Suk Kang School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorByung Kook Lim, Byung Kook Lim School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorIhn Sik Seong, Ihn Sik Seong School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorJaehong Seol, Jaehong Seol School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorChin Ha Chung, Chin Ha Chung School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorJimin Wang, Jimin Wang Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06520Search for more papers by this authorSoo Hyun Eom, Corresponding Author Soo Hyun Eom eom@gist.ac.kr Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaDepartment of Life Science, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea===Search for more papers by this author Seong-Hwan Rho, Seong-Hwan Rho Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea Seong-Hwan Rho and Hyun Ho Park contributed equally to this work.Search for more papers by this authorHyun Ho Park, Hyun Ho Park Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea Seong-Hwan Rho and Hyun Ho Park contributed equally to this work.Search for more papers by this authorGil Bu Kang, Gil Bu Kang Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaSearch for more papers by this authorYoung Jun Im, Young Jun Im Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaSearch for more papers by this authorMin Suk Kang, Min Suk Kang School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorByung Kook Lim, Byung Kook Lim School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorIhn Sik Seong, Ihn Sik Seong School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorJaehong Seol, Jaehong Seol School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorChin Ha Chung, Chin Ha Chung School of Biological Sciences, Seoul National University, Seoul 151-742, KoreaSearch for more papers by this authorJimin Wang, Jimin Wang Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06520Search for more papers by this authorSoo Hyun Eom, Corresponding Author Soo Hyun Eom eom@gist.ac.kr Department of Life Science, Cell Dynamics Research Center, Gwangju Institute of Science and Technology, Gwangju 500-712, KoreaDepartment of Life Science, Gwangju Institute of Science and Technology, Gwangju 500-712, Korea===Search for more papers by this author First published: 02 November 2007 https://doi.org/10.1002/prot.21758Citations: 5 AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat INTRODUCTION ATP-dependent proteases play vital roles in protein quality control and in regulating the levels of certain cellular proteins.1-3 Escherichia coli and other bacteria, including Bacillus subtilis, contain at least three types of multimeric ATP-dependent proteases homologous to the eukaryotic 26S proteasome: Lon, Clp and HslVU.1, 2, 4, 5 Of these, HslVU has been extensively studied as the simplest proteasome ancestor.6-10 Like the 26S proteasome, HslVU is comprised of two multimeric components: the ATPase HslU, which belongs to the AAA superfamily of ATPases,11 and the peptidase HslV, which shares common structural features with the catalytic β-type subunits of the 20S proteasome. Despite a sequence identity of only about 20%, they share the amino acids crucial for proteolysis and auto-cleavage [Fig. 1(A)]. Moreover, both use a threonine residue at the N-terminus as a catalytic nucleophile exposed by the processing of a methionine residue or a prosegment upon assembly, and are thus members of the N-terminal nucleophile (Ntn)-hydrolase family.14-16 Figure 1Open in figure viewerPowerPoint Crystal structure of CodW. A: Partial sequence alignment of the precursors of 20S proteasome core particle-related proteins accomplished using the program CLUSTALX.12 The residues considered to be crucial for proteolysis are indicated (▴): Thr-1, Asp-17 and Lys-33 in E. coli HslV; Thr-6, Asp-22 and Lys-39 in CodW.13 A downward arrow (↓) indicates the position where autolysis occurs. The N-terminal pro-segments preceding the cleavage site are highly divergent in their lengths and sequences. The aligned sequences are as follows: CodW from B. subtilis (labeled as CodW); HslVs from E.coli, H. influenzae and T. maritima (HslV_Ec, HslV_Hi and HslV_Tm, respectively); 20S poteasome β-subunit from T. acidophilum (PTB_Ta); 20S proteasome β1, β2, β5 and β7 from S. cerevisae (PTB1_Sc, PTB2_Sc, PTB5_Sc and PTB7_Sc, respectively); 20S proteasome βX, βY and βZ form H. sapiens (PTBX_Hs, PTBY_Hs and PTBZ_Hs, respectively). B: Side view of one dodecamer generated by applying a crystallographic twofold operation perpendicular to the cylindrical axis of the half-dodecamer (Mol A) in the asymmetric unit. One protomer and its symmetry mate are colored green. C: View of the half-dodecamer (W6) looking parallel to the molecular sixfold axis. An unbiased composite omit map contoured at 1σ (blue mesh) shows the ordered N-terminal pro-segment in each protomer. The N-terminus is located at the interface between two protomers (red asterisk). D: Close-up view of the pro-segment. The number in parenthesis indicates the position of the residue relative to Thr-6. CodW is the proteolytic component of CodWX, which is highly homologous to HslVU17 and has sequence identities of 52 and 55% with E. coli and H. influenzae HslV, respectively. A previous electron microscopic analysis revealed that CodW and HslV have similar molecular architectures with strict sixfold symmetry and similar dimensions (11 nm in diameter and 10 nm in height).18 In addition, both can form a hybrid protease capable of degrading SulA, suggesting CodWX and HslVU are close relatives.15 Nonetheless, recent biochemical studies suggest that CodW is distinct from HslV in several respects.15, 18, 19 First, although CodW is synthesized as a precursor with additional residues at its N-terminus, just like proteasomal β and other HslV proteases, unlike those others CodW is incapable of autolysis before the nucleophile Thr-6 upon assembly.15 It may be, therefore, that CodW utilizes its N-terminal serine as a catalytic nucleophile. Second, CodWX is an alkaline protease that is maximally active at pH 9.5, unlike HslUV which has a maximal activity at pH 8. Third, CodW by itself has no peptidase activity, even towards small peptides; it requires hydrolysis of ATP by CodX, which in turn leads to formation of a CodWX complex, for digestion of peptide and protein substrates. By contrast, uncomplexed HslV does show weak peptidase activity toward small peptides, and the binding of ATP is sufficient for the formation of HslVU, though ATP hydrolysis is required for the digestion of a protein substrate. This dissimilarity in the behaviors of CodW and HslV is extraordinary, since all of the catalytically important residues in HslV are well conserved in CodW. One possibility is that the observed disparities reflect sequence differences between HslV and CodW away from the conserved active site. If so, comparison of the structures of CodW and HslV should clarify the structural basis of the unique functional properties of CodW. Currently, a structure of CodW in complex with E. coli HslU is available.20 This, however, diffracted poorly due to lattice translocation defects in the crystal and lacks the information necessary for comparison with other HslV structures. Here, we describe the crystal structures of Bacillus subtilis CodW determined in two different space groups, P21 and C2, at pH 5.5 and 7.5, respectively. In both structures, the five extra N-terminal residues remain intact. This lack of autolysis can be explained by the impaired geometry of the canonical catalytic apparatus. MATERIALS AND METHODS Protein purification and crystallization The protein was expressed and purified as described previously.15 CodW protein concentrated to 5 mg/mL was crystallized at 293 K using the sitting drop vapor diffusion method. P21 crystals were obtained using a well solution of 55% (v/v) MPD, 0.1M sodium citrate (pH 5.5), and 8% glycerol. C2 crystals were obtained using a well solution of 30% (v/v) MPD and 0.1M HEPES-NaOH (pH 7.5) with 0.2M sodium citrate as a salt additive. X-ray diffraction data were collected on beam line 6B of the Pohang Accelerator Laboratory (PAL) in Korea. Taking advantage of the high concentration of MPD, crystals were directly frozen and cryo-cooled in a liquid nitrogen stream at 110 K. Data were processed using MOSFILM and SCALA software.21 Diffraction data were collected to a maximum resolution of 2.5 Å for C2 crystals and 3.0 Å for P21 crystals. Data collection, structure determination, and refinement Crystallographic computations were mostly carried out using the CCP421 and CNS22 program packages, and model building was performed using O.23 Molecular replacement searches using the P21 data were carried out to obtain the initial phases. The search model used was a dodecamer generated from the 1.9 Å structure of H. influenzae free HslV (PDB # 1G3K).24 After 12-fold noncrystallographic symmetry (NCS) averaging and prime-and-switch phasing using RESOLVE,25 it was evident that the Ser-1 to Ala-5 prosegment was intact in the structure. An improved electron density map around Ser-1 was computed from the C2 crystal data, which diffracted to a higher resolution of 2.5 Å. Combinations of protomers from the dodecameric structure determined using the P21 data set were used as search models. As a result, one half-dodecamer with a sixfold axis parallel to the c-axis was detected using a hexameric ring model (W6). The other half-dodecamer molecule (W3W3) parallel to the b-axis was then manually positioned by examining the 2Fo − Fc Fourier map computed using the calculated phases from the first W6 molecule. The two half-dodecameric CodW molecules in the asymmetric unit and the crystallographic dyad operation yielded the biological unit. Structural refinements with NCS restraints using CNS were conducted until convergence was obtained (Rcryst = 0.255, Rfree = 0.298). A large difference in isotropic B factors was observed between the two half-dodecameric molecules: 50 Å2 for W6 and 100 Å2 for W3W3. Because this implies the existence of large crystal anisotropy around a sixfold axis coincident with the crystallographic twofold-axis, 12 TLS groups assigned to each monomer were refined using REFMAC.26 As a result, the displacement parameters, which differed significantly for the two half-dodecamers, were well accounted for by the TLS parameterization, leaving residual local displacements that were very similar between protomers and to which NCS restraints could be applied. The TLS refinement gave improvements in crystallographic R and free R factors up to 3% (Rcryst = 0.226, Rfree = 0.265), and the resultant density map was also improved across the entire asymmetric unit. Solvent molecules including a Na+ ion at the conserved cation-binding site in each CodW protomer became apparent during the later stages of refinement and were added to the model. After the structural refinement was completed using C2 data at a resolution of 2.5 Å, the refined structure was transported back to the P21 data at a lower resolution of 3.0 Å and further refined until no further reduction in Rfree was observed (Rcryst = 0.245, Rfree = 0.266). The data collection and refinement statistics are summarized in Table 1. Table 1. Data Collection and Model Refinement Statistics Crystal data Native form I, pH 7.5 Native form II, pH 5.5 Data collection Space group C2 P21 Unit cell dimensions a (Å) 185.1 78.1 b (Å) 106.8 164.4 c (Å) 152.7 95.9 β (°) 112.0 111.5 Resolution range (last shell) (Å) 30.0–2.5 (2.54–2.50) 44.7–3.0 (3.16–3.00) No. of observations 280,849 169,982 No. of unique reflections 90,904 44,182 Completeness (%) 95.5 (88.5) 98.1 (97.6) Average I/σ(I) 17.2 (2.8) 7.1 (2.3) Rsymaa Rsym = Σ|Ihkl − |/Σ, where Ihkl is the scaled intensity of the hkl reflection, and is the mean value of the hkl reflection. (last shell) 0.055 (0.345) 0.089 (0.322) Refinement Rcrystbb Rcryst = Σ|Fobs − Fcalc|/ΣFobs>, where Fobs and Fcalc are the observed and calculated structure factor amplitudes of the hkl reflection. 0.226 0.245 Rfreecc Rfree was calculated in the same manner as Rcryst, but using the test set of reflections. 0.265 0.266 No. of reflections (working set) 86,324 39,712 No. of reflections (test set) 4576 4438 No. of protein atoms 16,571 16,308 No. of solvent/hetero atoms 251 106 rmsd bond lengths (Å) 0.017 0.011 rmsd bond angles (°) 1.5 1.4 a Rsym = Σ|Ihkl − |/Σ, where Ihkl is the scaled intensity of the hkl reflection, and is the mean value of the hkl reflection. b Rcryst = Σ|Fobs − Fcalc|/ΣFobs>, where Fobs and Fcalc are the observed and calculated structure factor amplitudes of the hkl reflection. c Rfree was calculated in the same manner as Rcryst, but using the test set of reflections. Validation and deposition The main chain angles, calculated using PROCHECK,27 showed that all residues fell into the most favored or additionally favored region of the Ramachandran plot. The atomic coordinates and structure factors of the C2 and P21 crystals have been deposited with the PDB codes 2Z3B and 2Z3A, respectively. Figures were produced using PyMOL.28 RESULTS AND DISCUSSION The structure of CodW was determined in two different space groups, P21 and C2, at pH 5.5 and 7.5, respectively. The C2 structure was refined to 2.5 Å and the P21 structure to 3.0 Å. Both structures had good geometry and nearly identical quaternary arrangements: the root-mean-square deviation for the main chain atom pairs was 0.67 Å for subunit comparisons and 0.91 Å for dodecamer comparisons, with the pore-lining residues 91–93 exhibiting the largest deviations, which were greater than 1 Å. The overall architecture of the CodW molecule is a double-doughnut-shaped dodecamer very similar to that of HslV, as was expected from the high degree of sequence conservation [Fig. 1(B,C)]. The rms deviation was 1.2 Å for protomer comparisons and 1.9 Å for dodecamer comparisons between CodW and H. influenzae HslV (PDB #1G4A). The protomers share the typical four-layered fold exhibited by HslV and all 20S proteosome subunits; two sheets of antiparallel β-strands are sandwiched between two layers of α-helices. Each protomer includes all 180 residues and a structural Na+ ion at the conserved cation-binding site formed by three main chain carbonyl oxygen atoms from residues 164, 167, and 170 [Fig. 2(A)]. The mean distance between the metal ion and oxygen atoms in our structure was 2.43 Å, which is nearly identical to the value observed in the T. maritima HslV structure.10, 30 Figure 2Open in figure viewerPowerPoint Structural basis for CodW's lack of autolysis. A: Superposition of monomeric CodW (yellow) with the NLVS-complex (grey) and the free form (cyan) of H. influenza HslV (PDB #1OFI). The ordered portion of the covalently bound inhibitor 4-iodo-3-nitrophenyl-acetyl-leucinyl-leucinyl-leucinyl-vinylsulfone (NLVS) to HslV is shown as grey lines. The red sphere indicates a sodium ion bound at the conserved cation-binding site consisting of three carbonyl groups and an adjacent water molecule. B: Superposition of the conserved active site residues in mature proteasome β/HslV-like proteases. The active site residues from the free form (PDB #1OFH), the HslU-bound form (PDB #1OFH) of H. influenzae HslV, the free T. maritime HslV (PDB #1M4Y) and the β2-subunit of the yeast 20S proteasome (PDB #1RYP) are shown. The residues are numbered as in H. influenzae. Thr-1 and Thr-2 are colored in orange for comparison. C: The active site of the T1A mutant of the β1-subunit of the yeast 20S proteasome, which lacks autolysis, is shown as balls-and-sticks (PDB #1RYP). The modeled side chain of Thr-1 is drawn in green lines overlapping the side chain of Ala-1. The Leu-(-2) to Thr-1 segment shows a typical γ-turn, which ensures the addition of Oγ to Gly-(-1)-C.29 The S1 substrate binding pocket is indicated in grey. D: The impaired geometry of the canonical catalytic apparatus in CodW is evident. The composite omit map shows a clear density consistent with an N-terminal prosegment containing Ser-1, Ser-2, Phe-3, His-4, and Ala-5 [Fig. 1(C,D)]. The backbone atoms of the Ser-2 to His-4 segment form antiparallel β-sheet hydrogen bonds with strand β3. The prosegment appears to be intact in every protomer in the asymmetric unit, unlike in the quintuple mutant of the β6-subunit of yeast 20S proteasome or T. maritima HslV, which was found to be partially processed.31 In CodW, the average main chain B-factor of the pro-segment (45Å2) was lower than that of the whole chain (51 Å2). Moreover, the CodW protein was eluted in fractions with a size of ∼240 kDa on gel filtration chromatography, which corresponds to the size of the dodecamer, and was detected as a single band on SDS-PAGE.15 The prosegment occupies the cleft corresponding to the substrate-binding site in HslV and the catalytic proteasomal β-subunits [Fig. 2(A)]. CodW has all the conserved residues known to be crucial for proteolysis, including Thr-6, Asp-22, Lys-39, Ser-130, and Gly-131 [Fig. 2(D)]. Within the 20S proteasome, every β-subunit contains these conserved residues and is processed through autolysis, which leads to activation.13, 31, 32 Indeed, the mechanism of autolysis is highly correlated with that of proteolysis. In this context, CodW's lack of autolysis is unusual, since at the sequence level CodW appears to meet all the requirements for proteolysis. On the other hand, a comparison of the geometries of the conserved active sites in CodW and the 20S proteasome/HslV-like proteases provides a clue as to why CodW does not self-cleave at the CN bond between Ala-5 and Thr-6. In yeast proteasomal β-subunits, the catalytic apparatus includes Thr-1, Asp-17, and Lys-33 (Thr-1, Asp-17, and Lys-33 in H. influenzae HslV). Although this is analogous to the catalytic triad seen in serine proteases, it differs in that the N-terminal amino group and a catalytic water molecule act as a proton acceptor in proteolysis and autolysis, respectively.8, 14, 31 Superposition of the active site residues in HslV and those in the mature yeast proteasomal β2-subunit shows that the geometry of the active site is maintained precisely [Fig. 2(B)]. Even in H. influenzae HslV, whose activation processing by HslU is accompanied by a conformational change around the substrate binding cleft,33 the positions of the active site residues are almost invariant during activation. The conserved hydrogen-bonding pattern between Thr-1 and the Ser-125/Gly-126 segment appears to serve as a stabilizing factor for Thr-1 at this position, since the turn segment containing the conserved Ser and Gly residues (Ser-130 and Gly-131 in CodW) is the most rigid part of the structure with the lowest average B-factors. In that context, the crystal structures reveal that CodW has an impaired catalytic apparatus. Most importantly, Thr-6 is shifted by ∼3 Å from its expected position, which displaces the side chain of Lys-39 into an outwardly oriented position [Fig. 2(D)]. The same type of impaired geometry of the catalytic apparatus also was seen in the quintuple reactivation mutant of the yeast proteasomal β6-subunit,31, 34 where deviation of Thr-1 from its usual position in the active subunit prevents the Lys-33-Nζ and a catalytic water molecule from staying in the vicinity of Thr-1. Interestingly, this mutant was still capable of partial autolysis, perhaps because the active site residues could spontaneously recover the functional geometry. By contrast, CodW exhibits no autolytic activity, implying the impaired catalytic apparatus of CodW is not spontaneously recoverable. In their description of the structure of the T1A mutant of the yeast proteasomal β1-subunit, Ditzel et al. suggested that the formation of a γ-turn by the Leu-(-2) to Thr-1 segment is one of the major conformational constraints on self-cleavage.29 In their model, Thr-1-Oγ would be centrally positioned over the three-residue γ-turn and attack the carbonyl carbon of Gly-(-1) [Fig. 2(C)], which means that the right conformation of the prosegment with respect to Thr-1-Oγ is also essential for autolysis. In the structure of the proteasomal β1-subunit T1A mutant, the S1 specificity pocket accommodates the hydrophobic side chain of Leu-(-2) and helps to form a bulge at the Leu-(-2) to Thr-1 segment. The corresponding pocket within the structure of CodW is occupied by Lys-39, which is displaced from the active site by Thr-6 [Fig. 2(D)]. Consequently, His-4 at position -2 is displaced from the pocket and stacked against Tyr-169. Apparently, the dislocation of His-4 brings about the extension of the prosegment and prevents the formation of a γ-turn at the His-4 to Thr-6 segment. It is noteworthy that the β6-subunit reactivation mutant shows blurred electron density at residues Asn-(-2) and Gly-(-1), which implies an inherent conformational flexibility in this region. We suggest that autolysis would take place in the reactivation β6-subunit mutant whenever the conformation of the Asn-(-2) to Thr-1 segment meets the structural constraints for autolysis, that is, when it forms a γ-turn. However, the crystal structure of CodW shows neither a γ-turn nor a sign of conformational flexibility at the His-4(-2) to Thr-6 segment [Fig. 2(D)]. Therefore, the spontaneous formation of a functional canonical active site via autolysis must overcome a much higher energy barrier in CodW. Several lines of biochemical evidence imply that Ser-1 is the N-terminal catalytic nucleophile in CodW.15, 18 In the present study, however, the unambiguous density denoting Ser-1-Oγ is located at the cleft formed between two protomers and is close to the entrance of the proteolytic chamber, but it shows no relevant interaction with neighboring residues. This may be attributable to the low pH of the crystallization buffer. In fact, CodWX becomes inactive at pH 5.5. Otherwise, a significant conformational rearrangement around the active site would be inevitable upon activation by CodX. The interiors of the catalytic chambers of the proteolytic components of HslV, ClpP, and the 20S proteasome are hydrophobic and favor denatured polypeptide chains (data not shown).35 By contrast, the catalytic chamber of CodW is strongly basic, but it is apparent that the chamber will become less polar at pH 9.5, where CodW has its maximal activity. The entrance to the proteolytic chamber of CodW is formed by a hydrogen-bonded turn and a short 310-helix and contains a cluster of conserved positively charged residues, Lys-92, Arg-95, and Lys-96 (Arg-86, Arg-89, and Arg-90 in E. coli HslV), whose functions are not yet clear. The average diameter of the pore, which measures 20.5 Å between Cα atoms, is narrow enough to prevent folded proteins from reaching the proteolytic core. The average diameter of the pore is nearly same in the E. coli, H. influenzae and T. maritime enzymes, but it changes slightly upon HslU binding.24, 36 It has been suggested that this "gating" motion is an allosteric activation/inactivation mechanism of HslV by HslU and is analogous the activation mechanism in yeast proteasomes. However, the conformational change at the pore entrance reportedly has no relation to the activation of HslV, and thus its function remains to be discovered.37 Although the pore leading the proteolytic core of CodW remains open in the P21 structure, in the C2 structure it is blocked by the presence of a poly-anion interacting with Lys-92 in each protomer [Fig. 3(A,B)]. Though presumed to be a citrate ion, the identity of the poly-anion remains unidentifiable because of the rotational averaging of sixfold symmetry. It is also unclear how many of the Lys-92 residues in the six protomers have an extended conformation to simultaneously coordinate the anion. It is noteworthy that this type of interaction has not been seen in any other HslV structure. Figure 3Open in figure viewerPowerPoint Obstruction of the pore leading to the proteolytic core by Lys-92. A: A 2Fo − Fc map contoured at 0.8σ within 2 Å from the pore-lining residues is overlaid on the structure (grey mesh). A Fo − Fc map contoured at 3σ shows an unidentified poly-anion from either the co-purification or crystallization buffer at the sixfold symmetry axial pore (blue mesh). B: Disordered side chains of the pore-lining residues in the P21 crystal. In HslUV/proteasome-like proteases, unfolding and translocation of the polypeptide are mediated by the ATPase subunit, and the unfolded chain is transported to the protease subunit in a vectorial fashion.38 A denatured polypeptide chain traveling from the outlet of the ATPase subunit to the active site in the proteolytic chamber can be regarded as a random flight chain with a free end. If some structural elements were to bias the random walk of the polypeptide chain, the efficiency of the translocation of the substrate to the active site cleft would be greatly improved. In fact, the possible existence of a ratcheting mechanism directing substrates to the proteolytic chamber has been anticipated.39 The antechambers of the 20S proteasome, which each have a volume of ∼59 nm3, may provide an example, as they must be able to maintain the polypeptide in an unfolded form as it passes through them. Prokaryotic HslV proteins do not have these antechambers. Instead, they have a flexible insertion containing a cluster of conserved positively charged residues. Moreover, the removal of a positive charge or the introduction of a negative charge at Arg-86 of E. coli HslV inactivates the enzyme,40 which implies that this residue interacts with the substrate polypeptide and that its electrostatic nature is important in the translocation of the substrate. Interestingly, the Δ86–91 mutant was found to be functionally indistinguishable from wild-type enzyme,40 which means the pore-lining segment likely facilitates substrate translocation in a passive manner. In CodW, the Lys-92 residues in some of the six protomers likely constitute a binding site for the C-terminal carboxylate of the substrate. The study presented here revealed the intact structure of the N-terminal prosegment of CodW and showed that its lack of autolysis can be attributable to the impaired geometry of its catalytic apparatus. However, the structures failed to clarify the catalytic mechanism of CodW. Although Ser-1 was previously proposed to be an N-terminalnucleophile, the local structure around it does not provide all the required features of the active site of a protease. Furthermore, restoration of the canonical active site and subsequent autolysis will require input of additional energy to overcome the conformational constraints observed in the current structures. The activation of CodW thus appears to be accompanied by a conformational change either in the canonical catalytic apparatus or Ser-1. It therefore seems plausible that the binding of CodX is critical to the activation of CodW. In summary, the formation of a functionally