Sialic Acid Recognition by Vibrio cholerae Neuraminidase

Vibrio cholerae neuraminidase (VCNA) plays a significant role in the pathogenesis of cholera by removing sialic acid from higher order gangliosides to unmask GM1, the receptor for cholera toxin. We previously showed that the structure of VCNA is composed of a central β-propeller catalytic domain flanked by two lectin-like domains; however the nature of the carbohydrates recognized by these lectin domains has remained unknown. We present here structures of the enzyme in complex with two substrates, α-2,3-sialyllactose and α-2,6-sialyllactose. Both substrate complexes reveal the α-anomer of N-acetylneuraminic acid (Neu5Ac) bound to the N-terminal lectin domain, thereby revealing the role of this domain. The large number of interactions suggest a relatively high binding affinity for sialic acid, which was confirmed by calorimetry, which gave a Kd ∼ 30 μm. Saturation transfer difference NMR using a non-hydrolyzable substrate, Neu5,9Ac2-2-S-(α-2,6)-GlcNAcβ1Me, was also used to map the ligand interactions at the VCNA lectin binding site. It is well known that VCNA can hydrolyze both α-2,3- and α-2,6-linked sialic acid substrates. In this study using α-2,3-sialyllactose co-crystallized with VCNA it was revealed that the inhibitor 2-deoxy-2,3-didehydro-N-acetylneuraminic acid (Neu5Ac2en) was bound at the catalytic site. This observation supports the notion that VCNA can produce its own inhibitor and has been further confirmed by 1H NMR analysis. The discovery of the sialic acid binding site in the N-lectin-like domain suggests that this might help target VCNA to sialic acid-rich environments, thereby enhancing the catalytic efficiency of the enzyme. Vibrio cholerae neuraminidase (VCNA) plays a significant role in the pathogenesis of cholera by removing sialic acid from higher order gangliosides to unmask GM1, the receptor for cholera toxin. We previously showed that the structure of VCNA is composed of a central β-propeller catalytic domain flanked by two lectin-like domains; however the nature of the carbohydrates recognized by these lectin domains has remained unknown. We present here structures of the enzyme in complex with two substrates, α-2,3-sialyllactose and α-2,6-sialyllactose. Both substrate complexes reveal the α-anomer of N-acetylneuraminic acid (Neu5Ac) bound to the N-terminal lectin domain, thereby revealing the role of this domain. The large number of interactions suggest a relatively high binding affinity for sialic acid, which was confirmed by calorimetry, which gave a Kd ∼ 30 μm. Saturation transfer difference NMR using a non-hydrolyzable substrate, Neu5,9Ac2-2-S-(α-2,6)-GlcNAcβ1Me, was also used to map the ligand interactions at the VCNA lectin binding site. It is well known that VCNA can hydrolyze both α-2,3- and α-2,6-linked sialic acid substrates. In this study using α-2,3-sialyllactose co-crystallized with VCNA it was revealed that the inhibitor 2-deoxy-2,3-didehydro-N-acetylneuraminic acid (Neu5Ac2en) was bound at the catalytic site. This observation supports the notion that VCNA can produce its own inhibitor and has been further confirmed by 1H NMR analysis. The discovery of the sialic acid binding site in the N-lectin-like domain suggests that this might help target VCNA to sialic acid-rich environments, thereby enhancing the catalytic efficiency of the enzyme. Cholera is an acute bacterial infection of the intestine caused by ingestion of food or water contaminated with Vibrio cholerae serogroups O1 or O139. All toxigenic strains of V. cholerae have a neuraminidase (sialidase) encoded within a pathogenicity island in their genomes (1Jermyn W.S. Boyd E.F. Microbiology (Read.). 2002; 148: 3681-3693Crossref PubMed Scopus (121) Google Scholar). A specific role of the enzyme in pathogenesis is the removal of sialic acid from higher order gangliosides to reveal GM1, the receptor for cholera toxin (2Galen J.E. Ketley J.M. Fasano A. Richardson S.H. Wasserman S.S. Kaper J.B. Infect. Immun. 1992; 60: 406-415Crossref PubMed Google Scholar). The nanH gene, encoding the neuraminidase, is part of the nan-nag gene cluster that encode a series of enzymes that are potentially involved in the utilization of sialic acid released by the neuraminidase (1Jermyn W.S. Boyd E.F. Microbiology (Read.). 2002; 148: 3681-3693Crossref PubMed Scopus (121) Google Scholar). The enzyme may therefore also play a nutritional role to the bacterium by providing sialic acid as an alternative energy and carbon source (3Schneider D.R. Parker C.D. J. Infect. Dis. 1982; 145: 474-482Crossref PubMed Scopus (41) Google Scholar). Structural studies on bacterial, viral, and parasite-derived neuraminidases (4Crennell S.J. Garman E.F. Laver W.G. Vimr E.R. Taylor G.L. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 9852-9856Crossref PubMed Scopus (234) Google Scholar, 5Gaskell A. Crennell S. Taylor G. Structure (Lond.). 1995; 3: 1197-1205Abstract Full Text Full Text PDF PubMed Scopus (192) Google Scholar, 6Buschiazzo A. Tavares G.A. Campetella O. Spinelli S. Cremona M.L. Paris G. Amaya M.F. Frasch A.C. Alzari P.M. EMBO J. 2000; 19: 16-24Crossref PubMed Scopus (122) Google Scholar, 7Buschiazzo A. Amaya M.F. Cremona M.L. Frasch A.C. Alzari P.M. Mol. Cell. 2002; 10: 757-768Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar, 8Luo Y. Li S.C. Chou M.Y. Li Y.T. Luo M. Structure (Lond.). 1998; 6: 521-530Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar, 9Crennell S. Takimoto T. Portner A. Taylor G. Nat. Struct. Biol. 2000; 7: 1068-1074Crossref PubMed Scopus (344) Google Scholar, 10Varghese J.N. Laver W.G. Colman P.M. Nature. 1983; 303: 35-40Crossref PubMed Scopus (726) Google Scholar) have shown that there is a family of small enzymes (∼40 kDa) that consist of only the catalytic β-propeller domain, and a family of larger enzymes that possess, in addition, one or more carbohydrate-binding domains. A recent study on the mode of action of bacterial neuraminidases has shown that those enzymes that possess additional domains are able to hydrolyze polyvalent substrates with far greater catalytic efficiency than their monovalent counterparts (11Thobhani S. Ember B. Siriwardena A. Boons G.-J. J. Am. Chem. Soc. 2003; 125: 7154-7155Crossref PubMed Scopus (63) Google Scholar). We previously reported the crystal structure of the 83-kDa V. cholerae neuraminidase (VCNA), 1The abbreviations used are: VCNA, V. cholerae neuraminidase; ITC, isothermal titration calorimetry; STD, saturation transfer difference; Neu5Ac, N-acetylneuraminic acid; Neu5Ac2en, 2-deoxy-2,3-didehydro-N-acetylneuraminic acid. 1The abbreviations used are: VCNA, V. cholerae neuraminidase; ITC, isothermal titration calorimetry; STD, saturation transfer difference; Neu5Ac, N-acetylneuraminic acid; Neu5Ac2en, 2-deoxy-2,3-didehydro-N-acetylneuraminic acid. a three-domain protein with a canonical six-bladed β-propeller neuraminidase domain flanked by two lectin domains (12Crennell S. Garman E. Laver G. Vimr E. Taylor G. Structure (Lond.). 1994; 2: 535-544Abstract Full Text Full Text PDF PubMed Scopus (192) Google Scholar) (see Fig. 1). One lectin domain is at the N terminus, and the other is inserted between the second and third “blades” of the β-propeller. The lectin domains share the same structure despite sharing only a 23% sequence identity. A comparison with other lectin structures provides no clue as to the ligand(s) recognized by these domains, but we speculated that the lectin domains may help this secreted enzyme to bind to cells through carbohydrate interactions. This would explain the observation that VCNA remains attached to cell membranes treated with VCNA (13Sedlacek H.H. Seiler F.R. Behring Inst. Mitt. 1974; 55: 254-257Google Scholar) and a report that VCNA had two binding sites with different affinities (14Luben G. Sedlacek H.H. Seiler F.R. Behring Inst. Mitt. 1976; 59: 30-37Google Scholar). Here we report the x-ray structures of two complexes of VCNA with the substrates α-2,3-sialyllactose to 1.9 Å and α-2,6-sialyllactose to 1.6 Å. These studies reveal that the N-terminal lectin domain binds the α-anomer of Neu5Ac (α-Neu5Ac). The number of direct and water-mediated interactions between sialic acid and the lectin domain, when compared with other sialic acid recognizing proteins, suggest a high binding affinity for VCNA. Using a non-hydrolyzable thiosialoside, Neu5,9Ac2-2-S-(α-2,6)-GlcNAcβ1Me (see Fig. 2), the affinity was measured as a Kd ∼ 30 μm by isothermal titration calorimetry (ITC). To explore the details of ligand recognition at the N-terminal lectin domain, we used saturation transfer difference (STD) NMR spectroscopy to map the ligand epitopes with the same thiosialoside used in the ITC experiment. In the α-2,3-sialyllactose complex, we found the neuraminidase inhibitor Neu5Ac2en is bound at the catalytic site suggesting that VCNA can synthesize its own inhibitor, as reported for the influenza virus neuraminidase (15Burmeister W.P. Henrissat B. Bosso C. Cusack S. Ruigrok R.W. Structure (Lond.). 1993; 1: 19-26Abstract Full Text PDF PubMed Scopus (151) Google Scholar), the Trypanosoma rangeli (16Amaya M.F. Buschiazzo A. Nguyen T. Alzari P.M. J. Mol. Biol. 2003; 325: 773-784Crossref PubMed Scopus (61) Google Scholar), and the Newcastle disease virus hemagglutinin-neuraminidase (9Crennell S. Takimoto T. Portner A. Taylor G. Nat. Struct. Biol. 2000; 7: 1068-1074Crossref PubMed Scopus (344) Google Scholar). We also confirmed this for VCNA by 1H NMR. Protein Expression—Escherichia coli HB101 transformed with the expression vector pCVD364 that produces VCNA was a gift from Prof. E. Vimr, University of Illinois, Urbana. The enzyme, retained in the periplasm, was purified as described previously using the osmotic shock fluid method (17Vimr E.R. Lawrisuk L. Galen J. Kaper J.B. J. Bacteriol. 1988; 170: 1495-1504Crossref PubMed Google Scholar). Because of the limited yield from the osmotic shock method, another vector, pET30 b(+), which expressed the enzyme in the cytoplasm and gave a higher yield, was developed and used to transform E. coli BL21(DE3) cells. The cDNA of the entire nanH gene lacking its signal sequence was amplified by the PCR using pCVD364 as template. The forward primer was CGTCCATATGGCACTTTTTGACTATAACGCAACGGG, incorporating an NdeI site (underlined), and the reverse primer was CCGGCTCGAGTTAGTTTTGCGATAAGGTCAGCTTCTG, incorporating an XhoI site (underlined) after the stop codon of nanH gene. These primers generated a 2274-bp fragment encoding the protein without signal sequence. Amplification of the gene was carried out in a 50-μl volume containing 0.1 μm each primer, a 50-ng template, 100 μm dNTPs, and 3 units of Pyrococcus furiosus DNA polymerase (Promega) in 1× Pfu DNA polymerase buffer, which was supplied with the enzyme. PCR conditions were 95 °C for 30s for 1 cycle followed by 30 cycles at 95 °C for 30 s, 55 °C for 60 s, and 72 °C for 3 min. A final extension time of 72 °C for 7 min was then applied. The resulting gene fragment was cleaned and double-digested with NdeI and XhoI before ligation to the expression vector pET30b(+), which was predigested with the same enzymes. The ligation mix was then used to transform E. coli DH5α cells. Positive clones were identified from LB agar plates containing 30 μg/ml kanamycin by colony PCR using the original PCR primers, and the sequence was confirmed by DNA sequencing prior to transforming the expression host E. coli BL21(DE3). Expression was performed by inoculating single colonies into LB medium containing 30 μg/ml kanamycin. Cultures were grown first at 37 °C to an optical density of 0.5–0.6 at 600 nm before subjecting cells to a heat shock of 42 °C for 20 min. The temperature of the cultures was then lowered to 25 °C for 10 min before the addition of 1 mm isopropyl-1-thio-β-d-galactopyranoside to initiate protein expression. Cultures were then left shaking at 25 °C overnight. Cells were collected by centrifugation (8000 × g, 4 °C), and pellets were frozen at -80 °C until required. Protein Purification—The enzyme used in the NMR experiment and in the co-crystals of α-2,6-sialyllactose was obtained using the second vector. The purification protocol of the enzyme was the same for the two expressions. The enzyme was recovered from the shock fluid (or cell lysis) by precipitation with 50% (w/v) ammonium sulfate. The pellet was dissolved in the minimum volume of 10 mm Tris-HCl, pH 7.6, the undissolved material was removed by centrifugation, and the clear sample was dialyzed exhaustively against the same buffer containing 100 mm NaCl. The sample was then loaded onto a POROS HQ20 anion exchange column. The protein was eluted from the column using a buffer system containing 20 mm Tris-HCl, pH 7.6, and 250 mm NaCl. Fractions were checked for neuraminidase using as substrate and were a column in 20 mm Tris-HCl, pH 7.6, 100 mm after against the same with the were grown using in a at °C and obtained by of VCNA in 20 mm Tris-HCl, pH 7.6, 10 mm mm and to 1 of a mm of α-2,3-sialyllactose or α-2,6-sialyllactose in 20 mm Tris-HCl, pH 7.6, and 1 of the containing mm and a 10 mm of and We that the of the the of the to the that might bind to the enzyme. The and mm in 20 mm Tris-HCl, pH and were collected on at the All were collected conditions using in the buffer as a were and using the Scopus Google Scholar) and into with 1994; PubMed Scopus Google Scholar). The to the are in and of of the is of the as but a of of protein of ligand of of water protein protein for the and of the ligand of the crystal structure at a the is of the as but a of protein for the and of the protein in a and using J. A. 1994; Scopus Google Scholar) was used to the structures of the complexes using the structure of the temperature VCNA as the of domains was carried out in the R.W. J. T. G.L. 1998; PubMed Scopus Google Scholar) against the measured in the Å. of and using the were carried out with with the M. A. PubMed Scopus Google Scholar). are in The were in difference were in both structures 1). were using in and water were to the α-2,3-sialyllactose and α-2,6-sialyllactose complex for the α-2,3-sialyllactose and α-2,6-sialyllactose complexes have been with the and 1H NMR to the of of 10 mm α-2,3-sialyllactose in the of VCNA in of mm buffer containing mm and 10 mm was by 1H NMR spectroscopy time on a at were with and of a of with a of The pH for of VCNA was reported to between and G.L. J. PubMed Scopus Google Scholar). A series of were in after the addition of the VCNA to the α-2,3-sialyllactose were at and were each of 10 mm α-2,3-sialyllactose and 10 mm Neu5Ac2en in the same buffer mm buffer, pH containing mm and 10 mm were as NMR sample containing μm VCNA and μm of the thiosialoside in of mm buffer containing mm and 10 mm and a sample containing μm of thiosialoside in only of the buffer were of of the thiosialoside was using the NMR experiment of the thiosialoside in complex with VCNA, the protein was at a of in the of the and at with a of of with a between each resulting in a saturation time of with saturation of and were also All NMR were at with and a of 1 was using the sequence Taylor M. J. 1992; PubMed Scopus Google Scholar), and a was used to the protein The time of the of thiosialoside in the of VCNA were measured using the these which were performed in a of of with a were were to the by the of the in the NMR with signal of a the and were a that was for 10 was using the expression system as described of VCNA was in the ITC buffer composed of mm Tris-HCl, pH 7.6, mm 10 mm and A sample of the thiosialoside was in the ITC were performed using the The isothermal binding was obtained by a mm ligand into a μm VCNA The heat released or was obtained by the heat of the of were by mm ligand into the ITC buffer, and the of binding were for of prior to analysis. were and with the The from the at the and which the the structures of the and enzymes are the and the α-2,3-sialyllactose complex structures a of Å for with the domains of and Å for the N-terminal lectin domain, the catalytic domain, and the lectin domain, The α-2,3-sialyllactose and α-2,6-sialyllactose complexes are and with a of Å for density of both complexes the bound in the of the N-terminal lectin domain The binding is on the same of the enzyme with to the catalytic site Å 1). The of in to the binding and the out to All of the the and of the to the with the lectin domain Fig. was no in the density for part of the suggesting that it is or more that the had been the in comparison to the enzyme. The of that its and against the of the of to its to with the of and by to with the of the and of the of VCNA with Neu5Ac and Neu5Ac2en an a water and in a NMR to the of Neu5,9Ac2-2-S-(α-2,6)-GlcNAcβ1Me to VCNA 4 the between and of the 1H NMR and NMR of the thiosialoside in complex with VCNA. a of saturation between and were only the saturation time for the experiment is between and to the Neu5Ac carbohydrate of the thiosialoside are in the however no significant signal of the is to the attached to and the and of of the ligand are in this these suggest that the of the thiosialoside binds to the VCNA domain but the is in with the enzyme in this signal is for the attached at of Neu5Ac suggesting that this has with the domain. In the attached to the of Neu5Ac have the with the VCNA, as by the of this signal in the A comparison of the for the of the thiosialoside, that have comparison of their binding J. E.R. J. 2003; PubMed Scopus Google Scholar), that when the (Neu5Ac) is to then the for the the of the and the were of the order of of the by using the thiosialoside gave the from two of and × a Kd 37 μm. of the Neu5Ac2en in the difference density map for the α-2,3-sialyllactose complex showed the inhibitor Neu5Ac2en bound in the site The structure reported here details of the interactions at the site and the of water in the Fig. and the interactions. in other an with the of Neu5Ac2en, and the of is to of a with a acid in has been in the catalytic of neuraminidases (15Burmeister W.P. Henrissat B. Bosso C. Cusack S. Ruigrok R.W. Structure (Lond.). 1993; 1: 19-26Abstract Full Text PDF PubMed Scopus (151) Google Scholar, Taylor M. J. 1992; PubMed Scopus Google Scholar, M. J. Chem. 1994; PubMed Scopus Google Scholar, Amaya M.L. Buschiazzo A. Alzari Frasch A.C. J. Am. Chem. Soc. 2003; 125: PubMed Scopus Google Scholar). part of the a with of the well as direct between the ligand and the enzyme, there are water-mediated interactions in the site of the α-2,3-sialyllactose complex and of The in a on one by the of and and on the other by the of and The α-2,6-sialyllactose complex Neu5Ac2en in the catalytic site but a the in a to the of Neu5Ac2en in the α-2,3-sialyllactose 1H NMR Neu5Ac2en the the of the of α-2,3-sialyllactose by VCNA as by 1H NMR in the of the 1H NMR from to it is to the of the substrate α-2,3-sialyllactose and the of the of the and the and of each of the Neu5Ac are in this of the The first shown in Fig. at for the of at and for at the the of the for the in until at however there is a in to the of This is in with previously with the α-anomer of Neu5Ac released as the first of the which then to to a final of Neu5Ac Fig. also between and the a signal at in this of the at between and A of Neu5Ac2en in the same buffer without VCNA identified this signal to to the of The of the Neu5Ac2en as the a after however its until it after We have shown that the N-terminal lectin domain of VCNA binds through direct and water-mediated suggesting a high binding The of is from the lectin binding suggesting that this lectin is to sialic with different The is also from the sialic with at to as we found with the thiosialoside in the NMR and ITC The of the lectin domain of VCNA is to that of the N-terminal lectin domain, share only 23% sequence identity. these studies have revealed the nature of ligand(s) recognized by the domain. of the two lectin domains that the with the ligand in the domain in the domain, by in from the for ligand It is therefore that the domain binds sialic acid. sialic from have been and structures of in complex with Neu5Ac are from J. Mol. Biol. PubMed Scopus Google Scholar, A. C. J. S. L. J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar), M. Mol. Cell. 1998; 1: Full Text Full Text PDF PubMed Scopus Google Scholar, W.S. J. Cell. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar), Cusack S. Nature. 1988; PubMed Scopus Google Scholar, T. S.C. Structure (Lond.). Full Text Full Text PDF PubMed Scopus Google Scholar, O. G. S.C. J. 2002; PubMed Scopus Google Scholar, V. M. D.R. Takimoto T. Portner A. Taylor G. J. PubMed Scopus Google Scholar, G. S.C. EMBO J. 2002; PubMed Scopus Google Scholar), and S. T. W.G. Protein Sci. 6: PubMed Scopus Google Scholar, A. Nat. Struct. Biol. 1994; 1: PubMed Scopus Google Scholar, C. C. E. J. Biol. Chem. 2000; Full Text Full Text PDF PubMed Scopus Google Scholar). All of these sialic acid recognizing domains have different and however bind to the same ligand by recognizing different of the VCNA another to the In the complexes the at to sialic in or with The VCNA lectin domain has interactions with sialic acid compared with the other sialic with the involved in We the affinity of the VCNA lectin domain by ITC using a non-hydrolyzable thiosialoside, to of substrates by VCNA, which gave a of μm. The calorimetry that this is by a that an The NMR confirmed the interactions for the sialic acid and showed that there is between the lectin domain and the We can therefore that a Kd ∼ 30 μm the binding affinity of the lectin domain for which is a relatively high affinity for carbohydrate has a Kd ∼ 100 μm with 1982; PubMed Scopus Google Scholar), but this also from interactions of the virus has a Kd ∼ 1 mm with J.E. PubMed Scopus Google Scholar), as the sialic acid binding domain of the O. G. S.C. J. 2002; PubMed Scopus Google Scholar). affinities in the with sialic acid containing have been with cholera toxin A. E. PubMed Scopus Google Scholar), T. J. G. 2003; PubMed Scopus Google Scholar), toxin T. J. G. 2003; PubMed Scopus Google Scholar), and a lectin PubMed Scopus Google Scholar), but in sialic acid is one with other carbohydrate to the The of the inhibitor Neu5Ac2en in the site of VCNA co-crystallized with α-2,3-sialyllactose was on the substrate used in the 1H NMR spectroscopy was used to the of the of α-2,3-sialyllactose by VCNA and these confirmed that Neu5Ac2en was as a of the the it was that the of Neu5Ac2en a and then in of Neu5Ac2en provides a for the in the signal at The of VCNA to its own inhibitor the as has been by these provides a for the Scopus Google Scholar) of sialic acid at high In Scopus Google Scholar) this was to the binding of that has a affinity for the The of Neu5Ac2en from the complex of ligand and the that the enzyme the α-2,6-linked substrate at the of the substrates 1983; PubMed Scopus Google Scholar). the of a sialic acid binding domain on VCNA suggests a for its role in the pathogenesis of The sialic acid binding site the enzyme to the catalytic domain in to its receptor thereby enhancing its catalytic efficiency and revealing GM1, the receptor for cholera toxin. It may also that the enzyme to the of the fluid and action in the small