Molecular Insight into the Role of the N-terminal Extension in the Maturation, Substrate Recognition, and Catalysis of a Bacterial Alginate Lyase from Polysaccharide Lyase Family 18*

Background: The maturation and catalysis mechanisms of the PL18 alginate lyases have not yet been reported. Results: The N-terminal extension in the precursor of PL18, aly-SJ02, helped the catalytic domain fold correctly. Key residues for substrate recognition and catalysis were determined. Conclusion: The catalytic mechanism of aly-SJ02 is proposed. Significance: This study provides the foremost insight into maturation and catalysis of PL18 alginate lyases. Bacterial alginate lyases, which are members of several polysaccharide lyase (PL) families, have important biological roles and biotechnological applications. The mechanisms for maturation, substrate recognition, and catalysis of PL18 alginate lyases are still largely unknown. A PL18 alginate lyase, aly-SJ02, from Pseudoalteromonas sp. 0524 displays a β-jelly roll scaffold. Structural and biochemical analyses indicated that the N-terminal extension in the aly-SJ02 precursor may act as an intramolecular chaperone to mediate the correct folding of the catalytic domain. Molecular dynamics simulations and mutational assays suggested that the lid loops over the aly-SJ02 active center serve as a gate for substrate entry. Molecular docking and site-directed mutations revealed that certain conserved residues at the active center, especially those at subsites +1 and +2, are crucial for substrate recognition. Tyr353 may function as both a catalytic base and acid. Based on our results, a model for the catalysis of aly-SJ02 in alginate depolymerization is proposed. Moreover, although bacterial alginate lyases from families PL5, 7, 15, and 18 adopt distinct scaffolds, they share the same conformation of catalytic residues, reflecting their convergent evolution. Our results provide the foremost insight into the mechanisms of maturation, substrate recognition, and catalysis of a PL18 alginate lyase.

Alginate, which accounts for ϳ40% of the dry weight of the biomass of brown algae, is an important component of marine carbon sink (1) and plays important roles in the marine ecosystem. Alginate from seaweed is also widely utilized in the food, chemical, and pharmaceutical industries as a stabilizing, thickening, or emulsifying reagent because of its ability to form gels (2,3). Furthermore, alginate is also synthesized as an exopolysaccharide by certain bacteria (4). The alginate biofilm produced by Pseudomonas aeruginosa is an important virulence factor during lung infections in cystic fibrosis patients (5). Alginate is a natural linear polysaccharide composed of (1,4)-linked ␤-D-mannuronate and its C5 epimer, ␣-L-guluronate. These uronic acids are arranged in a homopolymeric ␤-D-mannuronate block, a homopolymeric ␣-L-guluronate block, or heteropolymeric blocks with a random arrangement of both monomers (6).
To date, the three-dimensional structures of nine alginate lyases from families PL5, 7, 14, 15, 17, and 18 have been deter-mined and deposited into Protein Data Bank (PDB). Based on these structures, the mechanisms for substrate recognition and catalysis of alginate lyases in PL5, 7, 14, and 15 have been elucidated. The catalytic triad, Tyr, His and Asn (Gln), is highly conserved in both PL5 and PL7 (12,13). However, the PL5 alginate lyase A1-III utilizes Tyr 246 as both the proton acceptor and proton donor during catalysis (14), whereas the PL7 alginate lyase A1-IIЈ uses His 191 as the proton acceptor and Tyr 284 as the proton donor (13,15). Similar to A1-IIЈ, the PL15 alginate lyase Atu3025 also uses His 311 as the catalytic base and Tyr 365 as the catalytic acid, although its overall structure is quite different from A1-IIЈ (16). The catalytic mechanism of the PL14 lyases is still largely unknown, although the structure of the PL14 lyase vAL-1 has been determined. vAL-1 exhibits endolytic activity at pH 7.0 and exolytic activity at pH 10.0, because the electric charges at the active site greatly influence the substrate binding mode (17).
For the PL18 family, only the structure of the alginate lyase from Pseudoalteromonas sp. 272 (hereafter called Aly272) has been deposited in the PDB (code 1J1T). The molecular mechanisms for substrate recognition and catalysis of the PL18 alginate lyases have not been reported. In addition, the PL18 alginate lyases have a distinct structural characteristic that differs from those in the other PL families. Most PL18 alginate lyases contain an N-terminal extension in their precursors, which is excised during enzyme maturation. The function of the N-terminal extension of the PL18 alginate lyases has never been studied, but they are predicted as a carbohydrate binding module based on sequence analysis (they showed ϳ30% identity to the putative chitin-binding domain of Streptomyces chitinases, which belong to family CBM 16) (18,19).
The aly-SJ02 lyase secreted by marine Pseudoalteromonas sp. SM0524 is a bifunctional alginate lyase (20). N-terminal sequence analysis suggests that aly-SJ02 is most likely an alginate lyase belonging to the PL18 family. aly-SJ02 is a highly efficient endotype alginate lyase, releasing mainly di-and trisaccharides from alginate (20). In an engineered microbial platform for direct biofuel production from brown macroalgae, aly-SJ02 is supposed to depolymerize alginate (21). In this study, using aly-SJ02 as a model of PL18 alginate lyases, the function of the N-terminal extension in enzyme maturation was investigated by biochemical and structural analyses. The molecular mechanisms for substrate recognition and catalysis were studied by molecular dynamics simulation combined with structural and mutational analyses. Based on the results, a model for the catalysis of aly-SJ02 in alginate depolymerization is proposed.

EXPERIMENTAL PROCEDURES
Gene Cloning, Protein Expression and Purification, and Sitedirected Mutagenesis-Based on the N-terminal sequence of aly-SJ02 and the conserved sequence of the PL18 alginate lyases (22), the aly-SJ02 gene was obtained by PCR and thermal asymmetric interlaced PCR (23). After verification with high fidelity Fastpfu DNA polymerase (Transgen Biotech), the nucleotide sequence of this gene was submitted to the GenBank TM database under the accession number EU548077. The aly-SJ02 gene and its truncations, including the N-terminal extension (NTE, Ala 32 -Ser 173 ), the catalytic domain (CATD, Thr 174 -Asn 400 ), and the precursor (Ala 32 -Asn 400 ), were expressed in Escherichia coli BL21 (DE3) with pET22b, pET28a, or pGEX-4T-1 (Novagen). The recombinant proteins, which were fused with either a C-terminal poly-His tag or an N-terminal GST tag, were first purified on nickel-nitrilotriacetic acid-resin (Qiagen) or glutathione S-Sepharose 4B (GS4B; GE Healthcare) resin. The proteins were then fractionated by anion ion exchange using Source 15Q and by gel filtration with Superdex 75 resin (GE healthcare). All purified proteins were immediately stored at Ϫ80°C until use in further assays. The site-directed mutagenesis of aly-SJ02 was performed using the typical overlap extension PCR strategy and verified by sequencing. The mutants were expressed in E. coli BL21 (DE3), and the recombinant proteins were purified following the same protocol as the wild type.
The proteins separated by SDS-PAGE were blotted on a PVDF membrane (GE Healthcare). The N-terminal amino acid sequence of the proteins were analyzed by the Edman degradation method on an ABI Procise 491 protein sequencer at Peking University (Beijing, China). The method described by Ogura et al. (15) was used to determine whether the molecules of wildtype Aly-SJ02 and the N214C/T263C mutant had thiol groups.
Homology Modeling, Crystallization, Data Collection, Structure Determination, and Refinement-The homology model for mature aly-SJ02 (M-CATD, Thr 174 -Asn 400 ) was modeled using Modeler 9v7 (24) with the structure of Aly272 (PDB code 1J1T) as the template. A total of 1,000 models were created. The models with favorable objective function and DOPE scores were further validated using the SAVES server and the PSQS server. Ultimately, one model with high quality scores was selected.
The crystals of the recombinant catalytic domain (r-CATD, Thr 174 -Asn 400 ) and the recombinant precursor of aly-SJ02 were obtained using the hanging drop vapor diffusion method with 2 l of concentrated protein mixed with 2 l of well solution. The crystal of r-CATD (10 mg/ml) was grown in 100 mM phosphate-citrate (pH 4.2) and 40% (w/v) PEG-300, whereas the crystal of the precursor (6 mg/ml) was grown in 100 mM MES (pH 7.5), 25% (w/v) PEG-4000, and 150 mM ammonium sulfate. Before data collection, the crystals were equilibrated in a cryobuffer containing equal volumes of the reservoir buffer and 20% glycerol. The data were collected at the Shanghai Synchrotron Radiation Facility, Beamline BL17U, in a 100 K nitrogen stream, and they were processed with the HKL2000 program suite (25). The final model was built automatically by ARP/wARP in the CCP4i program package (26). Refinement of the structure was performed using the programs COOT (27) and PHENIX (28). The final model was evaluated using PRO-CHECK, and the structures of the r-CATD crystals and the catalytic domain in the precursor crystals (P-CATD, Thr 174 -Asn 400 ) were refined to high quality ( Table 1). All molecular graphics were created with PyMOL. To detect the metal ions in aly-SJ02, 40 M aly-SJ02 in 10 mM Tris-HCl (pH 8.0) and 100 mM NaCl was subjected to atomic absorption spectroscopy analysis using an atomic absorption spectrophotometer (TAS-990, PGENERAL). The calcium, copper, iron, manganese, and zinc in the sample were analyzed by atomic absorption spectroscopy.
Protein Concentration Determination and Enzyme Assays-The protein concentration was determined by a BCA protein assay kit (Thermo). The alginate lyase activity was determined by monitoring the increase in absorbance at 235 nm (A 235 ) that is caused by the production of unsaturated uronic acids as the lyase cleaves glycosidic bonds in the polymer chain (7). A mixture of 80 l of buffer (50 mM Tris-HCl, pH 8.0), 100 l of alginate substrate (10 mg/ml), and 20 l of enzyme was incubated at 40°C for 10 min. After incubation, the mixture was boiled for 5 min to terminate the reaction. One unit of enzyme activity was defined as the amount of enzyme that increased the A 235 by 0.1 per min. To determine the K m value of aly-SJ02 and its mutants, the substrate concentration was varied from 0.06 to 6 mg/ml. The kinetic parameters of aly-SJ02 and its mutants were derived by a nonlinear regression fit directly to the Michaelis-Menten equation using the Origin8 Pro SR4 software.
Model Docking and Molecular Dynamics Simulation-AutoDock 4.2 (29) was used to conduct aly-SJ02 and tetrasaccharide rigid body docking. The structure of the tetrasaccharide, a mannuronate tetramer, was downloaded from the PDB (code 4F13) (30). The P-CATD molecule was restrained within a grid box (70 ϫ 34 ϫ 38 points in each dimension). The docking searches were executed using the Lamarckian genetic algorithm with a maximum number of 25,000,000 energy evaluations and the default settings for all other options. A total of 20 candidate solutions were returned and ranked by binding energy. The selection of the final results was detailed under "Results and Discussion." To study the conformational flexibility of lid loop 1 and loop 2, energy minimization, and molecular dynamics simulation were performed using the NAMD program suite (31). The model was placed in a 55 ϫ 55 ϫ 46 Å rectangular box under periodic boundary conditions. Sodium and chlorine ions were randomly dispersed, and they were located at least 5 Å from the model and at least 5 Å from one another. The entire system was minimized over the following 1 ns. Then the equilibrated sys-tem was subjected to a 50-ns MD simulation at 300 K. The MD simulation ran on 20 cores of the Inspur Tiansuo at the Supercomputing Center of Shandong University. The average wall clock time was ϳ2 h for a 1-ns simulation.
Circular Dichroism Spectra-Wild-type aly-SJ02 and its mutants were subjected to circular dichroism spectroscopy assays at 25°C on a J-810 spectropolarimeter (Jasco). The CD spectra of the aly-SJ02 proteins at a final concentration of 25 M were collected from 250 to 195 nm at a scan speed of 200 nm/min with a path length of 0.1 cm. All aly-SJ02 proteins were in buffer containing 10 mM Tris-HCl (pH 8.0) and 100 mM NaCl.

RESULTS AND DISCUSSION
Sequence Analysis of aly-SJ02-The gene encoding aly-SJ02 was amplified from Pseudoalteromonas sp. SM0524. The open reading frame of aly-SJ02 is 1203 bp in length and is predicted to encode a PL18 precursor protein with 400 amino acids. According to the sequence alignment and a search in the NCBI Conserved Domain Database (32), the precursor of aly-SJ02 contains a signal peptide (Met 1 -Ala 31 ), an NTE (Ala 32 -Gly 155 ), a linker (Ser 156 -Ser 173 ), and a PL18 CATD (Thr 174 -Tyr 395 ) (Fig. 1A). The aly-SJ02 lysate secreted by Pseudoalteromonas sp. SM0524 only contains the CATD; the NTE is cleaved off at the bond between Gly 157 and Ser 158 in the linker (20). When the aly-SJ02 gene was expressed in E. coli, N-terminal amino acid sequencing of the recombinant aly-SJ02 showed that the NTE can be cleaved off at two sites in the linker: between Gly 157 and Ser 158 or between Asp 160 and Gly 161 . It has been reported that the NTEs of other PL18 alginate lyases are also cleaved off in the linker region during enzyme maturation (19). The exact cleavage sites are different among these PL18 alginate lyases, although they share high amino acid sequence identities (more than 70% between one another) (Fig. 1B). The cleavage site in AlyPEEC is between Gly 163 and Gly 164 (18). However, AlyA has two cleavage sites, one between Ser 147 and Gly 148 and the other between Asp 161 and Gly 162 , but the resulting 33-and 30.5-kDa forms showed little difference in enzymatic activity (19). The PL7 alginate lyases contain three conserved motifs, SA2 (RXEXR), SA3 (YXKAGXYXQ), and SA4 (QXH), which compose the active center and are crucial for substrate recognition and catalysis (33,34). According to sequence alignment and secondary structure prediction, the PL18 alginate lyases also have three similar conserved motifs. The conserved amino acid sequences in these motifs are RHEYK (SA2), YFKFGNYLQ (SA3), and QHH (SA4) (Fig. 1B).

TABLE 1 Diffraction data and refinement statistics of P-CATD and r-CATD
Structural Analysis of aly-SJ02-In addition to the intact aly-SJ02 gene, we also expressed the precursor (Ala 32 -Asn 400 ) and the CATD (Thr 174 -Asn 400 ) for crystallization. The crystal structure of r-CATD (the recombinant CATD) was determined to 2.1 Å, and the crystal structure of P-CATD (the CATD in the recombinant precursor) was determined to 1.7 Å. Unfortunately, the structure of the NTE (Ala 32 -Gly 155 ) in the precursor could not be solved. We also could not obtain crystals of mature aly-SJ02 (the product of the intact gene expressed in E. coli), most likely because it was not suitably stable for crystallization. With the PL18 alginate lyase Aly272 (PDB code 1J1T) as the initial search model, the structures of r-CATD and P-CATD were solved. Two molecules were found in the asymmetric unit of r-CATD, with one molecule rotated 180°to the other ( Fig.  2A). Our gel filtration analysis indicated that r-CATD was present as a monomer in solution, suggesting that the r-CATD dimer observed in the structure resulted from crystal packing. The overall structures of r-CATD and P-CATD have an root mean square deviation of only 0.202 Å. Moreover, they are nearly identical to that of Aly272 (Fig. 2, B and D), because there is only one amino acid residue difference between aly-SJ02 (S173-N400) and Aly272. Because the crystal structure of mature aly-SJ02 could not be solved, the structure of mature aly-SJ02 (hereafter called M-CATD, Thr 174 -Asn 400 ) was obtained by homology modeling based on the structure of Aly272.
The CATD of aly-SJ02 shows a ␤-jelly roll structure composed mainly of two anti-parallel ␤-sheets (sheets A and B) (Fig.  2C), which is a structural motif commonly shared by PL7 and PL18 alginate lyases. The SA2, SA3, and SA4 motifs form a catalytic groove. Similar to other endo-type alginate lyases (14,34,35), both ends of the catalytic groove of aly-SJ02 are not blocked (Fig. 2B), indicating that aly-SJ02 is an endo-type alginate lyase. This corresponds to the previous result that aly-SJ02 released di-, tri-, and tetra-saccharides from alginate (20). Two loops cover over the catalytic groove. Atomic absorption spectroscopy analysis confirmed that a Ca 2ϩ ion is present in the aly-SJ02 molecule. The structural analysis showed that this metal ion is chelated by residues Asp 273 , Asp 281 , Val 283 , Asn 286 , and Ile 288 , as well as a water molecule in the structure (Fig. 2E). The chelation site of Ca 2ϩ is far away from the active center, suggesting that this metal ion is not directly involved in the polysaccharide degrading reaction. To determine whether the Ca 2ϩ has an effect on enzyme activity, we added EDTA to the reaction mixture to chelate the Ca 2ϩ from the enzyme. The deprivation of Ca 2ϩ from aly-SJ02 did not cause any detectable structure changes, but it resulted in a 50% reduction in enzymatic activity (Fig. 2, F and G). This outcome indicates that the Ca 2ϩ in aly-SJ02 is important for maintaining enzymatic activity, even though it does not directly participate in the catalytic reaction.
N-terminal Extension-mediated Maturation of aly-SJ02-Unlike the alginate lyases from other families, most PL18 alginate lyases have an NTE in their precursors, which is predicted as a carbohydrate binding module in the NCBI Conserved Domain Database. However, the mature aly-SJ02 enzyme and the other characterized PL18 alginate lyases all contain only a catalytic domain. Therefore, it is impossible that the N-terminal extensions function as a carbohydrate-binding module in these enzymes during catalysis. Enzymatic activity analysis showed that the activity of r-CATD toward alginate was ϳ20% lower than that of M-CATD, which suggests that there may be some delicate differences in the structures of the active centers of r-CATD and M-CATD. Residues Arg 219 , Gln 257 , His 259 , Tyr 347 , Tyr 353 , and Lys 349 , located in the active center, are conserved in the alginate lyases from both PL7 and PL18 families, indicating that they may be crucial for substrate recognition or catalysis. We performed a detailed conformational comparison of these conserved residues in the active centers of r-CATD, P-CATD, and M-CATD. In P-CATD and M-CATD, these res-idues have the same conformation, suggesting that P-CATD folds correctly (Fig. 3A). For r-CATD, although these residues in molecule B of the asymmetric unit have nearly the same conformation as those in P-CATD (except a tiny swing between Arg 219 and Tyr 353 in the two molecules) (Fig. 3B), the conformation of molecule A differs greatly from P-CATD. The guanidino group in the side chain of Arg 219 in molecule A deviates ϳ3.5 Å from that in P-CATD. Similarly, the imidazolyl group of His 259 in molecule A has a 4.0 Å swing (Fig. 3C). These conformational differences indicate that without the NTE, the CATD of aly-SJ02 may fold incorrectly, thereby affecting its activity toward alginate. Thus, the N-terminal extension in the aly-SJ02 precursor is likely to function as an intramolecular chaperone to help the orderly folding of the CATD. Our biochemical assays also support this hypothesis. When the NTE and the CATD were co-expressed with two different vectors in an E. coli cell, they formed a protein complex by interaction, which eluted as a single peak that appeared earlier than both the NTE and the CATD peaks in gel filtration (Fig. 3, D and E). In contrast, the NTE and the CATD, when expressed separately in E. coli cells, could not form a protein complex when they were mixed in vitro, which resulted in two peaks during gel filtration (Fig. 3E). Altogether, these results suggest that the NTE in the aly-SJ02 precursor may facilitate the folding of the CATD. After the folding of the precursor, the NTE is removed by a protease that cuts in the linker region, and the CATD becomes a mature enzyme. Overall structure of the aly-SJ02 catalytic domain. A, overall structure of r-CATD. Two r-CATD monomers are located in the asymmetric unit, forming a face to face pseudo-complex. Molecule A is in blue, and molecule B is in yellow. B, superposition of the structures of Aly272 and r-CATD. Aly272 is in cyan, and the two molecules of r-CATD are in blue and yellow, as before. C, overall structure of P-CATD. The final model of the aly-SJ02 catalytic domain shows a glove-like ␤-jelly roll structure composed mainly of two anti-parallel ␤-sheets (SA and SB). D, superposition of the structures of Aly272 and P-CATD. Aly272 is in cyan, and P-CATD is in green. E, detailed view of the Ca 2ϩ chelated by aly-SJ02. The Ca 2ϩ ion is located far away from the active center. F, effect of EDTA on the activity of aly-SJ02. The lyase activity decreased with increasing EDTA. After the Ca 2ϩ was completely chelated out from the enzyme by EDTA, the enzyme activity did not further diminish with an increase of EDTA from 2 to 5 mM. G, circular dichroism spectra of aly-SJ02 with different concentrations of EDTA.

Gating Function of the Lid Loops in aly-SJ02 for Substrate
Entry-Alginate lyases typically have loops covering the active center, which are called the lid loops (17,34,36). A study on the lid loops of PL7 alginate lyase A1-IIЈ suggests that the flexibility of the lid loops is important for substrate accommodation in the active center (15). The lid loop of PL5 alginate lyase A1-III can move with a maximum distance of 13.4 Å between the open form (apoenzyme) and the closed form (holoenzyme) in the orthorhombic crystal system (30). aly-SJ02 also has two lid loops (loop 1, Ala 208 -Gly 217 ; loop 2, Ala 260 -Thr 265 ) over the substrate-binding pocket (Fig. 4A). The B factor for the loop 2 region is quite high in the structure (Fig. 4A), suggesting that  1J1T). The conserved residues in the two molecules share the same spatial conformation. B, conformational comparison of the conserved amino acid residues in the active centers of P-CATD (blue) and molecule B from the r-CATD asymmetric unit (yellow). The conserved residues in molecule B share almost the same conformation as those in P-CATD, except a tiny swing between Arg 219 and Tyr 353 in the two smolecules. C, conformational comparison of the conserved amino acids residues in the active centers of P-CATD (blue) and molecule A from the r-CATD asymmetric unit (red). Molecule A differs greatly from P-CATD in the conserved amino acid residue conformations. D, SDS-PAGE analysis of the co-expressed (Co-exp) NTE and CATD. In co-expression, the NTE (without His tag) and CATD (with His tag) formed a complex, which was purified by His tag affinity chromatography. The expressed NTE without the His tag could not bind nickel-nitrilotriacetic acid. E, gel filtration analysis of the interaction between the NTE and the CATD of aly-SJ02. The co-expressed NTE and CATD formed a complex that eluted in a single peak, and the separately expressed NTE and CATD did not when they were mixed in vitro. this loop may oscillate flexibly. To investigate whether the lid loops of aly-SJ02 have a gating function in substrate entry and product release, we performed a molecular dynamics simulation for the aly-SJ02 P-CATD structure with a particular focus on these two loops. The space between loops 1 and 2 tended to oscillate in size during the 50 ns of dynamic motion (supplemental Movie S1), resulting in "open" and "closed" states. Distance measurements of the Asn 214 and Thr 263 side chains showed that the space between the loops can increase to 11.5 Å in the open state from 3.2 Å in the closed state, which makes it possible for an alginate molecule to enter the substrate-binding pocket of aly-SJ02. Unlike the nearly rigid body motion of the lid loop of A1-III (30), the enlargement of space between the lid loops of aly-SJ02 is mainly caused by conformational changes in the side chains of amino acids on the loops (supplemental Movie S1). As shown in Fig. 4B, the distance between Asn 214 on loop 1 and Thr 263 on loop 2 is the minimum between the lid loops. To further confirm the significance of the two loops for substrate entry, an N214C/T263C mutant with Asn 214 and Thr 263 replaced by cysteine residues was constructed to introduce a rigid interaction between loop 1 and loop 2 by the formation of a disulfide bond. No thiol groups were detected in the N214C/T263C mutant, indicating that a disulfide bond formed between the lid loops. The mutation almost completely abolished the activity of aly-SJ02, and the reduction of disulfide bond formation by the addition of DTT significantly increased the mutant activity (Fig. 4C). This result indicates that maintaining lid loop flexibility is essential for substrate entry into the substrate-binding pocket of aly-SJ02. Altogether, our results indicate that the lid loops of aly-SJ02 can alternate between open and closed states to serve as a gate for substrate entry during catalysis. In the open state, the space between the loops is large enough for alginate entry. After the substrate enters the pocket, loops 1 and 2 can move closer to one another to form the closed state, which may promote subsequent alginate lysis. As soon as the catalytic reaction is completed, the lid loops return to the open state to release the products, preparing for the next catalytic reaction.
Analysis of the Important Residues Involved in Substrate Recognition-To investigate how a substrate binds the active site of aly-SJ02, we used the program AutoDock to simulate a model of the aly-SJ02-tetrasaccharide complex without any artificial residue restrictions, except that a tetrasaccharide (a mannuronate tetramer) downloaded from PDB (PDB code 4F13) was forced into the catalytic cavity of aly-SJ02. A total of 20 solutions were recorded, and only the first ranked solution, which carries the lowest binding energy reported by AutoDock, was collected for further analysis.
The subsites in aly-SJ02 are labeled with "Ϫn" to represent those that bind the nonreducing terminus of the oligosaccharide and "ϩn" to represent those that bind the reducing terminus (Fig. 5A). Based on the location of the oligosaccharide in the deepest cleft in the binding pocket, the tetrasaccharide is positioned at subsites Ϫ1, ϩ1, ϩ2, and ϩ3. Accordingly, the constituent alginate residues are designated as A Ϫ 1, A ϩ 1, A ϩ 2, and A ϩ 3 from the nonreducing end. The cleavage would occur between A Ϫ 1 and A ϩ 1 (37). Generally, the positively charged residues located in the active center of aly-SJ02 form a highly positive potential active center (Fig. 5B), which is quite adaptable for binding the negatively charged alginate and/or for neutralizing the negative charge of the carboxyl groups on alginate. Uronic acid residue A Ϫ 1 is accommodated at subsite Ϫ1 by residues Gln 355 and Lys 364 . Amino acid substitution of these two residues showed that the Q355A mutant was inactive and the K364A mutant retained 85% activity, suggesting that A Ϫ 1 is recognized mainly by Gln 355 through a hydrogen bond with the carboxyl group and that Lys 364 assisted in binding with the carboxyl group. At subsite ϩ1, the A ϩ 1 saccharide residue is recognized by residues on SA3, SA4, L1, and L2. The carboxyl group is recognized by Gln 257 and His 259 , O2 by Thr 263 , O3 by Asn 216 , and O4 by Thr 353 . Amino acid substitutions for Gln 257 and His 259 resulted in inactivation of the enzyme, whereas mutants of Thr 263 and Asn 216 still retained ϳ50% activity (Fig.  5C), which suggested that the conserved residues Gln 257 and His 259 are more important in A ϩ 1 recognition. At subsite ϩ2, the uronic acid residue is recognized by residues on SA2, SA3, and SA4. The carboxyl group is recognized by Arg 219 and Lys 349 , and the O3 and O4 are recognized by His 259 . Substitutions of Arg 219 and Lys 349 generated inactive mutants, showing the importance of these residues in A ϩ 2 recognition. At subsite ϩ3, the carboxyl group of A ϩ 3 is recognized by Lys 223 and Tyr 347 . Ala substitutions of Lys 223 and Tyr 347 led to inactivation of the enzyme, whereas K223R and Y347F retained 13 and 76% activity, respectively. The K m values of the active mutants above were all significantly increased (Fig. 5E), indicating that these mutations decreased the affinity of aly-SJ02 to the substrate and therefore led to a reduction in activity. Therefore, the strictly conserved residues, Arg 219 , Lys 211 (in SA2), Gln 257 , His 259 (in SA3), Tyr 347 , Lys 349 , and Gln 355 (in SA3), mainly interact with the uronic acid carboxyl groups of the substrate. Amino acid substitutions of the conserved residues at subsites ϩ1 and ϩ2 resulted in inactivation of the enzyme, whereas mutations at subsites Ϫ1 and ϩ3 still retained partial enzyme activity. These results indicate that substrate recognition in aly-SJ02 mainly happens at subsites ϩ1 and ϩ2, but certain secondary interactions occur at subsites Ϫ1 and ϩ3, which is consistent with the previous result that aly-SJ02 mainly released di-and trisaccharides from alginate (20).
Residue Glu 221 at SA2, which is quite conserved in both PL7 and PL18 alginate lyases, most likely has no interaction with the substrate, because the distance between them is Ͼ5 Å. However, the mutation of Glu 221 to Ala or similar residues (Asp or Gln) resulted in severe activity loss (Fig. 5D). Structural analysis showed that the distance from Glu 221 to Lys 349 is 2.9 Å, and that from Glu 221 to Arg 219 is 3.7 Å, indicating that there may be interactions between them. Therefore, Glu 221 is most likely essential to maintain the conformation of Lys 349 and Arg 219 , which are both responsible for the recognition of the carboxyl group of Aϩ2 (Fig. 5F).
In addition, analysis of the circular dichroism spectra of aly-SJ02 and its mutants indicated that there are no detectable structural changes in the mutants (Fig. 6). Therefore, the changes in the activity and the K m of the mutants resulted from residue substitutions rather than from overall structural changes.
Catalytic Mechanism of PL18 Alginate Lyase aly-SJ02-Alginate lyases catalyze the degradation of alginate by a ␤-elimination mechanism where the glycosidic 1-4 O-linkage is cleaved between subsites ϩ1 and Ϫ1, and a double bond forms between C4 and C5 on uronic acid A ϩ 1 (38). In the model of the aly-SJ02-tetrasaccharide complex, the distance between C5 of A ϩ 1 and Tyr 353 is 3.1 Å, and the distance between O4 of A ϩ 1 and Tyr 353 is 3.3 Å (Fig. 5G). Mutations of Tyr 353 resulted in inactivation of the enzyme (Fig. 5D). Therefore, residue Tyr 353 most likely functions as both a catalytic base and acid during aly-SJ02 catalysis.
In summary, our results for aly-SJ02 indicated that the lid loops are critical for substrate entry; the conserved residues at the active center, Arg 219 , Lys 223 , Gln 257 , His 259 , Tyr 347 , and Lys 349 , recognize and stabilize the carboxyl group of the substrate; and Tyr 353 acts as both a catalytic base and acid in catalysis. Based on our results, a model for the catalysis of aly-SJ02 in alginate degradation is proposed as follows (Fig. 7): (i) When A, schematic drawing of the sugar-binding subsites in alginate lyase. The nonreducing end of the substrate is drawn on the left, and the reducing end is on the right. The cleavage site is indicated by an arrow. B, the surface model of the aly-SJ02-tetrasaccharide complex. aly-SJ02 is shown as an electrostatic surface, and the tetrasaccharide is displayed as cyan sticks. C, the amino acid residues interacting with the tetrasaccharide in the aly-SJ02-tetrasaccharide model. The residues are presented as sticks in blue, and the tetrasaccharide is shown as lines in green. D, enzymatic activities of the aly-SJ02 mutants. The activity of wild-type aly-SJ02 is considered to be 100%. E, K m values of the mutants of aly-SJ02. The K m value of wild-type aly-SJ02 is considered to be 100%. F, interactions of Glu 221 with Arg 219 and Lys 349 . The distance between Glu 221 and the substrate is Ͼ5 Å. The distance from Glu 221 to Lys 349 is 2.9 Å, and that from Glu 221 to Arg 219 is 3.7 Å, suggesting interactions between them. G, the detailed amino acid environment around the cleavage site in the aly-SJ02-tetrasaccharide model. Tyr 353 is located at an appropriate site to act as both a catalytic base and acid.
aly-SJ02 approaches an alginate chain, the lid loops are in the open state, and the negatively charged alginate chain enters the positively charged active site. The conserved residues in the active site recognize the carboxyl groups of the substrate. After the substrate is recognized in the active site, the lid loops convert to the closed state for the subsequent catalytic reaction. (ii) FIGURE 6. Circular dichroism spectra of aly-SJ02 and its mutants. There were no detectable structural changes among aly-SJ02 and its mutants. The strong electronic attraction of the carboxylate anion makes the proton on C5 easily attacked by a base catalyst. The general base catalyst Tyr 353 abstracts the proton from C5 of the uronic acid, and a transition state forms with C5 becoming a carbanion. The positively charged residues at subsite ϩ1 form a positive environment to stabilize and neutralize the carbanion. The stability of the carbanion is a key factor for this reaction. (iii) Simultaneously, protonated Tyr 353 , as a general acid catalyst, donates the proton to the glycoside bond, whereas the C5 carbanion begins to form a double bond with C4. (iv) The glycosidic 1-4 O-linkage between A ϩ 1 and A Ϫ 1 is cleaved, and a double bond forms between C4 and C5 at A ϩ 1. The lid loops return to the open state, and the reaction products are released.
Structural Comparison of Alginate Lyases from Different Families-The carbohydrate-active enzymes have huge substrate diversity in a highly selective manner using only a limited number of available folds, indicating that they were therefore subjected to multiple divergent and convergent evolutionary events (39). A phylogenetic tree was constructed for the characterized alginate lyases in families PL5, 7, 14, 15, and 18 where the sequences fell into five branches (Fig. 8A). Although the characterized PL14 alginate lyases are from eukaryota or viruses, the other characterized alginate lyases are all from bacteria. The crystal structures of these alginate lyases adopt three different scaffolds. PL5 alginate lyase A1-III has an ␣/␣-barrel scaffold, the alginate lyases from families PL7, 14, and 18 adopt a ␤-jelly roll as their basic scaffold, and PL15 lyase Atu3025 has an ␣/␣-barrel ϩ anti-parallel-␤-sheet-fold (Fig. 8B). Despite the distinct structural scaffolds of PL5 A1-III and PL7 ALY-1, the catalytic residues of A1-III, Tyr 246 , His 192 , Asn 191 , and Arg 239 , share a similar spatial arrangement with the ALY-1 catalytic residues, Tyr 195 , His 119 , Gln 117 , and Arg 72 (12). We compared the active centers of the alginate lyases from families PL5, 7 (A1-IIЈ as the representative), 14, 15, and 18 (Fig. 8C). The spatial arrangements of the catalytic residues in the active centers of alginate lyases from families PL 5, 7, 15, and 18 are nearly identical, and the swing distances of the corresponding residues are less than 1 Å. Because their alginate substrate is a constant linear polysaccharide composed of mannuronic and its isomer guluronic acid, these bacterial alginate lyases may have evolved to degrade alginate in a similar manner, which reflects the convergent evolution of bacterial alginate lyases. In contrast,  (3A0N) in yellow, PL15 alginate lyase Atu3025 (3A0O) in blue, and PL18 alginate lyase aly-SJ02 (4Q8K) in pink. The putative catalytic residues are shown as red sticks. C, conformation comparison of the conserved residues in the alginate lyase active centers. The color of the conserved amino acid residues in the active center of each lyase corresponds to that of the overall structure in B. The distances between the residues are labeled. The spatial arrangements of the catalytic residues in the active centers of alginate lyases from families PL 5, 7, 15, and 18 are approximately the same, whereas those in the PL14 alginate lyase vAL-1 are quite different.
although PL14 alginate lyase vAL-1 has a similar ␤-jelly roll scaffold as the alginate lyases from families PL7 and 18, the arrangement of the putative catalytic residues of vAL-1 is quite different from PL7 and 18, implying that PL14 alginate lyases may utilize a different catalytic mechanism from the other alginate lyases. Because PL14 alginate lyases are from viruses or eukaryota, they may depolymerize alginate in a different manner from bacteria.
Conclusions-The alginate lyase aly-SJ02 from the marine bacterium Pseudoalteromonas sp. SM0524 is a member of the PL18 family. Mature aly-SJ02 containing only a catalytic domain displays a ␤-sandwich fold. The NTE in the aly-SJ02 precursor may help mediate protein folding during maturation. Structural and mutational analyses revealed that certain conserved residues at subsites ϩ1 and ϩ2 that stabilize or neutralize the negative charge of uronic acid are essential for substrate recognition. Moreover, these analyses indicated that Tyr 353 may function as both a catalytic base and acid. A model involving a four-step elimination reaction for alginate degradation catalyzed by aly-SJ02 is proposed based on our results. Our study provides the foremost insight into the maturation, substrate recognition, and catalysis mechanisms of a PL18 alginate lyase, which are helpful for further study of the biological roles and biotechnological applications of the PL18 alginate lyases.