Molecular Microbiology (2007) 65(5), 1258–1275 doi:10.1111/j.1365-2958.2007.05862.x First published online 27 July 2007 Biochemical characterization of a Neisseria meningitidis polysialyltransferase reveals novel functional motifs in bacterial sialyltransferases OnlineOpen: This article is available free online at www.blackwell-synergy.com Friedrich Freiberger,1 Heike Claus,2 Almut Günzel,1 Imke Oltmann-Norden,1 Justine Vionnet,3 Martina Mühlenhoff,1 Ulrich Vogel,2 Willie F. Vann,3 Rita Gerardy-Schahn1 and Katharina Stummeyer1* 1 Abteilung Zelluläre Chemie, Medizinische Hochschule Hannover, Carl-Neuberg-Str. 1, 30625 Hannover, Germany. 2 Institute for Hygiene and Microbiology, University of Würzburg, Josef-Schneider-Str. 2, 97080 Würzburg, Germany. 3 Laboratory of Bacterial Toxins, Center for Biologics Evaluation and Research, US FDA, 8800 Rockville Pike, Bethesda, MD 20892, USA. Summary The extracellular polysaccharide capsule is an essential virulence factor of Neisseria meningitidis, a leading cause of severe bacterial meningitis and sepsis. Serogroup B strains, the primary disease causing isolates in Europe and America, are encapsulated in a-2,8 polysialic acid (polySia). The capsular polymer is synthesized from activated sialic acid by action of a membrane-associated polysialyltransferase (NmB-polyST). Here we present a comprehensive characterization of NmB-polyST. Different from earlier studies, we show that membrane association is not essential for enzyme functionality. Recombinant NmB-polyST was expressed, purified and shown to synthesize long polySia chains in a non-processive manner in vitro. Subsequent structure–function analyses of NmB-polyST based on refined sequence alignments allowed the identification of two functional motifs in bacterial sialyltransferases. Both (D/E-D/E-G and HP motif) are highly conserved among different sialyltransferase families with otherwise little or no sequence identity. Their functional importance for enzyme catalysis and CMP-Neu5Ac binding Accepted 3 July, 2007. *For correspondence. E-mail stummeyer. katharina@mh-hannover.de; Tel. (+49) 511 532 4503; Fax (+49) 511 532 3956. Re-use of this article is permitted in accordance with the Creative Commons Deed, Attribution 2.5, which does not permit commercial exploitation. © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd was demonstrated by mutational analysis of NmBpolyST and is emphasized by structural data available for the Pasteurella multocida sialyltransferase PmST1. Together our data are the first description of conserved functional elements in the highly diverse families of bacterial (poly)sialyltransferases and thus provide an advanced basis for understanding structure–function relations and for phylogenetic sorting of these important enzymes. Introduction Neisseria meningitidis (Nm) is a leading cause of bacterial meningitis and sepsis in children and adolescents. Sporadic cases as well as outbreaks and epidemic waves are observed. Despite the availability of potent antimicrobial agents, case-fatality rates are high and survivors frequently suffer from sequelae such as limb loss and deafness (van Deuren et al., 2000). Essential virulence factors of disease causing meningococci are their extracellular polysaccharide capsules. Serogroup B strains (NmB), the primary disease-causing isolates in Europe and America, are encapsulated in a-2,8 polysialic acid (polySia). The NmB capsule was shown to mediate resistance to phagocytosis and complement-mediated bacteriolysis and it is chemically and immunologically identical to polySia found in the human host (Mackinnon et al., 1993; Hammerschmidt et al., 1994; Vogel et al., 1997; Vogel and Frosch, 1999). This mimicry also prevents the generation of effective polysaccharide-based vaccines against NmB strains. Bypass of host defence mechanisms by polySia capsules has not only been described for NmB, but is also an important virulence determinant of other disease-causing pathogens such as Escherichia coli K1 and K92, Moraxella nonliquefaciens, Pasteurella haemolytica and N. meningitidis serogroup C (Troy, 1992). The polySia capsules of NmC and E. coli K92 are connected by a-2,9 (Bhattacharjee et al., 1975) or alternating a-2,8/a-2,9 (Egan et al., 1977) glycosidically linked sialic acids, respectively. Because of their critical function in bacterial pathogenesis enzymes involved in polySia biosynthesis are interesting targets for therapeutic intervention. Biosynthesis of polySia is catalysed by membrane-associated polysialyltransferases (polySTs) at the cytoplasmic side of the inner Functional motifs in bacterial sialyltransferases 1259 membrane of Gram-negative bacteria (Masson and Holbein, 1983; Troy, 1992). The polymerization reaction proceeds by transfer of sialic acid from the donor substrate CMP-Neu5Ac to the non-reducing end of a growing polySia chain and was proposed to be processive, as no reaction intermediates could be detected (Steenbergen and Vimr, 2003). Polysialyltransferases have been cloned from E. coli K1 and K92 as well as from N. meningitidis serogroup B and C (Weisgerber et al., 1991; Vimr et al., 1992; Edwards et al., 1994; Claus et al., 1997). Identity is highest between the pair of E. coli (82% identity) and meningococcal enzymes (65% identity) and is lower between the two genera (33% identity). Interestingly, the neisserial enzymes are elongated by a C-terminal domain not present in the E. coli polySTs. While only few studies included the neisserial polySTs (Masson and Holbein, 1983; Swartley et al., 1997; Steenbergen and Vimr, 2003), the E. coli enzymes have been studied more extensively (Whitfield, 2006; Troy, 1992). They are unable to start de novo biosynthesis of polySia, but elongate exogenously added acceptors including oligo- and polysialic acids, sialylgangliosides and synthetic acceptors (Steenbergen and Vimr, 1990; Cho and Troy, 1994; McGowen et al., 2001). The nature of the priming endogenous acceptor is still unknown, but recently gene products NeuE and KpsC were shown to be required for de novo synthesis of polySia in E. coli (Andreishcheva and Vann, 2006) and the functional complex of the E. coli K92 polysialyltransferase was found to be larger than a monomer (Vionnet et al., 2006). Construction of chimeric proteins between the closely related E. coli K1 and K92 enzymes, synthesizing a-2,8and alternating a-2,8/a-2,9-linked polySia, respectively, mapped the region responsible for linkage specificity to primary sequence elements located between amino acids 53 and 85 in both enzymes (Steenbergen and Vimr, 2003). However, so far no data on isolated proteins are available. Attempts to purify or solubilize polySTs failed and resulted in inactivation of the enzymes and it was therefore proposed that membrane association is required for polyST activity (Steenbergen and Vimr, 2003; Vionnet et al., 2006). The lack of purified native or recombinant protein furthermore prevented detailed structure function analyses of these important enzymes. No conserved motifs have been described for bacterial sialyltransferases. This is different from the eukaryotic sialyltransferases, where four conserved motifs have been shown to be involved in binding of donor and acceptor substrates and in enzyme catalysis (Drickamer, 1993; Livingston and Paulson, 1993; Geremia et al., 1997; Jeanneau et al., 2004). Moreover, based on primary sequence similarities all eukaryotic sialyl- and polysialyltransferases can be grouped into a single family GT-29 of the CAZy database (http://www.cazy.org) while the bacterial enzymes are distributed into four families (Coutinho et al., 2003). Bacterial polysialyltransferases are found in CAZy family GT-38, while families GT-42, GT-52 and GT-80 contain sialyltransferases that sialylate bacterial lipooligosaccharide (LOS). Structural information is available for the Campylobacter jejuni sialyltransferase cst-II (member of CAZy family GT-42) (Chiu et al., 2004) and the sialyltransferase PmST1 of Pasteurella multocida (member of CAZy family GT-80) (Ni et al., 2006). While cst-II belongs to the glycosyltransferase-A-like (GT-A) structural group, PmST1 has a GT-B-like fold. In the current study we focused on the characterization of the polyST from N. meningitidis serogroup B and throughout succeeded with soluble expression and purification of recombinant NmB-polyST. We thereby demonstrate that membrane association is not a prerequisite for the formation of functional enzyme. We also show that removal of the C-terminal extension present in NmB but not in the homologous E. coli enzymes, completely abolished enzymatic activity, proving it as an essential functional domain. Using site-directed mutagenesis and refined protein alignment strategies, we identified two functionally important motifs, which are highly conserved in a number of bacterial (poly)sialyltransferases of otherwise unrelated sequences. With these data we provide the first evidence for the conservation of catalytic features among bacterial sialyl- and polysialyltransferases and thereby improve the basis for design of sialyltransferasespecific drugs. Results Recombinant expression of NmB-polyST Detailed structure–function analyses of bacterial polysialyltransferases were hitherto prevented by insufficient supply of the enzymes. Expression levels were described to be low and polyST activity was found associated with bacterial membranes (Shen et al., 1999; Steenbergen and Vimr, 2003; Vionnet et al., 2006). As attempts to solubilize polyST were either unsuccessful or accompanied by inactivation of the enzyme, all analyses have been performed with crude membrane fractions as enzyme source. With the aim of obtaining purified enzyme in yields sufficient to perform structure–function studies, we began the current work with a systematic search for conditions that would allow production of recombinant NmB-polyST. To test the influence of N-terminal fusion partners on NmB-polyST expression and activity, constructs were generated either with short N-terminal epitope tags (T7, Strep II) or with additional large fusion parts like NusA and maltose-binding protein (MBP) (Fig. 1A). The constructs were expressed in E. coli BL21 (DE3) and soluble and insoluble fractions of the bacterial lysates were analysed © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1260 F. Freiberger et al. Fig. 1. Influence of N-terminal fusion tags on NmB-polyST expression and activity. A. Schematic representation of NmB-polyST fusion proteins. NmB-polyST is shown as white box while short epitope tags (His, T7, Strep II) and large fusion partners (MBP, NusA) are given as black and grey boxes respectively. The length of the black ruler represents 200 amino acids. B. Western blot analysis of NmB-polyST fusion proteins. Proteins were expressed in E. coli BL21(DE3) and equal amounts of soluble (s) and insoluble (i) fractions were separated by SDS-PAGE. C-terminally epitope-tagged fusion proteins were detected by Western blot analysis with anti-His-tag antibody. C. Enzymatic activity of NmB-polyST fusion proteins in the soluble fractions. PolyST activity was analysed using the radiochemical activity assay. Reactions were incubated at room temperature and aliquots were assayed for radiolabelled polySia at the indicated time points. Each value represents the average of three independent determinations with the standard deviation indicated. D. Specific activities of NmB-polyST fusion proteins were standardized by NmB-polyST expression levels, which were determined by immunoblotting and infrared fluorescence detection. for expression and activity of NmB-polyST. Although soluble protein could be detected for all constructs (Fig. 1B), the addition of large N-terminal fusions considerably increased the amount of active protein in the soluble fractions (Fig. 1C). Compared with polySTs carrying only short N-terminal epitope tags, additional fusion of NusA or MBP increased the soluble activity of NusA– and MBP–polyST two- and threefold, respectively. Also, the specific activity of both fusion proteins was increased twofold compared with enzymes with short tags (Fig. 1D). Subsequent experiments were therefore carried out with either NusA or MBP fusion proteins. Definition of the minimal active domain of NmB-polyST It has been reported for other bacterial sialyltransferases that N- or C-terminal truncations significantly increase protein solubility by eliminating membrane interaction domains (Chiu et al., 2004; Ni et al., 2006). Consequently, the second series of experiments was designed to define the minimal catalytic domain of NmB-polyST. N- and C-terminal truncations of the enzyme were generated as NusA fusion proteins carrying a C-terminal His-tag for detection. Full-length and truncated proteins were expressed in E. coli and soluble and insoluble cell fractions were tested for expression and enzymatic activity of NmB-polyST. As depicted in Fig. 2, approximately 50% of the wild-type polyST was soluble (Fig. 2B) and enzymatically active (Fig. 2D), whereby the detected activity was threefold higher in the soluble than in the insoluble fraction. Removal of 23 (D23NmBpolyST) and 33 (D33NmB-polyST) amino acids from the N-terminus had only slight effects on solubility and activity of NmB-polyST. However, deletion of the first 64 amino acids (D64NmB-polyST) shifted the majority of the expressed protein to the insoluble fraction and no enzymatic activity was detected in soluble or insoluble fractions. Primary sequence analysis revealed that NmB-polyST carries a C-terminal extension of 95 amino acids that is not present in the polySTs of E. coli K1 and E. coli K92 (Fig. 2A). To investigate the role of this additional © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1261 Fig. 2. Analysis of expression and activity of wild-type and truncated NmB-polyST. A. Schematic representation of bacterial polySTs. The N-terminal domain homologous in polysialyltransferases of E. coli K1, K92 and N. meningitidis serogroup B is shown as white box, with conserved amino acid stretches indicated in grey. The 95 amino acid comprising C-terminal domain of NmB-polyST, which is not conserved in the E. coli enzymes, is depicted in black. B and C. Western blot analysis of wild-type and truncated NmB-polySTs. NusA fusion proteins of full-length and N-terminally (B) and C-terminally (C) truncated NmB-polyST were expressed in E. coli BL21(DE3). Soluble (s) and insoluble (i) fractions (re-suspended in the same volume and buffer as the soluble fractions) were prepared and equal amounts were analysed by SDS-PAGE and Western blotting using anti-His-tag antibody. D. Enzymatic activity of soluble and insoluble fractions was analysed using the radiochemical polyST assay. The data were standardized by NmB-polyST expression levels that were determined by immunoblotting and infrared fluorescence detection, and are given relative to the enzymatic activity of the soluble wild-type fraction. Each value represents the average of three independent determinations with the standard deviation indicated. protein part, we generated a set of truncated NmBpolySTs lacking the C-terminal domain either partially (NmB-polySTD22, NmB-polySTD60) or completely (NmB-polySTD94, NmB-polySTD95, NmB-polySTD97). As displayed in Fig. 2C, all variants could be expressed as soluble proteins at similar or, in the case of constructs with entirely deleted C-terminal domain, even higher levels than the full-length enzyme. However, each C-terminal truncation completely abolished enzymatic activity (Fig. 2D), indicating that the C-terminal domain is indispensable for NmB-polyST activity. The fusion protein MBP–NmB-polyST produces capsular polySia in vivo Our efforts to express recombinant NmB-polyST in the E. coli expression strain BL21 (DE3) clearly revealed beneficial effects of large N-terminal fusion parts on the expression of active, soluble enzyme. However, to analyse if polyST fusion proteins maintain enzymatic activity also in vivo, wild-type and MBP–NmB-polyST were subcloned into a neisserial expression vector and transformed into the polyST-deficient neisserial strain © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1262 F. Freiberger et al. Fig. 3. In vivo activity of NmB-polyST fused to maltose-binding protein (MBP). A. MBP–NmB-polyST and native NmB-polyST carrying no additional tags were cloned into a neisserial expression vector and transformed into the polyST-deficient Neisseria strain 2517. PolySia capsules of parental strain (mock) and transformants were analysed by whole-cell ELISA using mab 735. Equal loading of the wells with Neisseria was controlled using mab P1.2 for detection of the meningococcal major outer membrane protein PorA. Each value represents the average of three independent determinations with the standard deviation indicated. B. Lysates of the parental strain (mock) and transformants were additionally analysed by SDS-PAGE and Western blotting using mab 735 before and after treatment with polySia-degrading endoN (top). Equal sample loading was confirmed in a parallel Western blot immunostained with anti-PorA antibody P1.2 (bottom). 2517. Parental strain and transformants were analysed for capsular polySia in a quantitative whole-cell ELISA. Equal loading of the microtitre plates with bacteria was confirmed in a parallel ELISA directed against the meningococcal major outer membrane protein PorA. Interestingly, no difference in capsule expression was observed between strains complemented with wild-type or the MBP fusion construct (Fig. 3A). Moreover, Western blot analysis of the neisserial lysates revealed a similar polySia staining of wild-type and fusion protein that was not detectable in samples treated with endoN (Fig. 3B), which is a bacteriophage-derived enzyme that degrades polySia with high substrate specificity (Mühlenhoff et al., 2003; Stummeyer et al., 2005). This clearly demonstrates that MBP–NmB-polyST is enzymatically active in vivo. Purification of recombinant NmB-polyST As MBP–NmB-polyST was shown to be active in vitro and in vivo and because MBP could be directly utilized for affinity chromatography, purification was optimized for the MBP–NmB-polyST construct schematically depicted in Fig. 1A. The fusion construct additionally carries two short epitope tags (N-terminal Strep II-, C-terminal His-tag), which were used for detection of the protein throughout the purification. After overexpression in E. coli BL21(DE3) at 15°C, the recombinant MBP–NmB-polyST was purified in two consecutive steps by MBP affinity and size exclusion chromatography. Trials to also utilize the C-terminal His-tag for purification failed, indicating that the epitope is not accessible in the native enzyme. As shown in Fig. 4A, the applied purification procedure yielded a highly enriched MBP–NmB-polyST (protein of 100 kDa). Major bands of smaller molecular weight still visible in the Coomassie-stained SDS-PAGE of the purified pool were also detected in a Western blot directed against the StrepII-tag, but not in a blot stained with an anti-His-tag antibody. This indicates that these bands represent breakdown products of MBP–NmB-polyST caused by C-terminal degradation. Throughout the purification, polyST activity was monitored by the radiochemical polyST assay using colominic acid as the acceptor. Enzyme activity was detected in all MBP–NmB-polyST-containing fractions and specific activity increased about 20-fold from cell lysate to the elute obtained after gel filtration (see Fig. 4A, bottom). As depicted in Fig. 4B, the fusion protein eluted in a single peak at the upper limit of the applied Superdex 200 gel filtration column, which suggests association of MBP– NmB-polyST as hexameric or higher-order oligomers. However, the purified protein was soluble and remained completely in the supernatant after high-speed centrifugation (100 000 g, Fig. 4C). One litre of expression culture yielded 1.3 mg of the highly enriched MBP–NmBpolyST. The protein was concentrated to 2 mg ml-1 and stored at 4°C for 3–4 weeks without detectable loss of activity or further degradation. Soluble recombinant MBP–NmB-polyST produces polySia chains in a non-processive manner Capsular polysaccharide purified from serogroup B N. meningitidis was shown to exhibit a high degree of polymerization (DP) (Gotschlich et al., 1969; Frosch and Müller, 1993). This implies that NmB-polyST is able to produce large polymers when the enzyme is part of the © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1263 Fig. 4. Purification of NmB-polyST. A. The MBP fusion protein of NmB-polyST was expressed in E. coli BL21 (DE3) and purified in a sequence of MBP affinity and size exclusion chromatography. Protein containing fractions were analysed by Coomassie-stained SDS-PAGE (left) and Western blot directed against the N-terminal StrepII-tag (middle) and the C-terminal His-tag (right). Protein bands corresponding to MBP–NmB-polyST are indicated by an arrow. Enzyme activity was monitored by the radiochemical polyST assay and specific activities were calculated as indicated. B. Size exclusion chromatography of MBP–NmB-polyST. Elution positions of standard proteins [Thyroglobulin (669 kDa), b-Amylase (200 kDa)] are indicated. Enzymatic activity of the collected fractions (secondary y-axis, grey circles) was determined using the enzyme-linked polyST assay. C. High-speed centrifugation of purified NmB-polyST. Samples were centrifuged as indicated and equal amounts of supernatant (sn) and pellet (p) fraction, which was re-suspended in the same volume as the supernatant, were analysed by Coomassie-stained SDS-PAGE. native capsule biosynthesis complex associated with the inner bacterial membrane. To analyse products synthesized by purified NmB-polyST in more detail, pentameric (DP5) a-2,8-linked oligosialic acid was used as the primer for the recombinant enzyme and the polymerization reaction was started by addition of radiolabelled sugar donor substrate CMP-[14C]-Neu5Ac. Products were separated by high percentage polyacrylamide gel electrophoresis and visualized by autoradiography. Because polySia can be specifically degraded with bacteriophage endosialidases (endoN) (Mühlenhoff et al., 2003; Stummeyer et al., 2005), the nature of the synthesized products was controlled by treating one of two parallel samples with endoN prior to electrophoresis. As shown on the gel in Fig. 5A, which separates polymerization products from DP 10 to > 100, a large fraction of the synthesized polySia was longer than 100 residues per chain. All reaction products were specifically degraded by endoN. This demonstrates that the recombinant MBP–NmB-polyST is able to synthesize long polySia chains starting with short oligomeric acceptors in vitro. However, in addition to the highmolecular-weight polySia, also shorter oligosialic acid reaction products were detectable (Fig. 5A). To validate the presence of short reaction products and analyse the mode of chain elongation in more detail, we studied the influence of acceptor concentrations on chain length distribution. Therefore, DP5 concentrations were altered over three orders of magnitude starting with 1 mM, while the concentration of enzyme and donor substrate CMPNeu5Ac was kept constant. As shown in Fig. 5B, increasing acceptor concentrations augment the concentration of short- and medium-sized reaction products until – at the point of equimolar concentrations of DP5 and CMPNeu5Ac (lanes 4 and 5) – virtually all radiolabelled products remain below DP11 (Fig. 5B). An identical correlation of chain length distribution and acceptor concentration was obtained for trimeric (DP3) a-2,8-linked oligo sialic acid (Fig. S1). These data argue against processivity of NmB-PolyST in vitro, as in case of a highly processive mechanism synthesis of few long chains should be favoured over synthesis of many short chains. To evaluate these data in a second assay system, we used our recently developed fluorescence-based polyST assay (Vionnet and Vann, 2007) that utilizes the trisialylganglioside analogue GT3-FCHASE as artificial acceptor substrate. In a first experiment GT3-FCHASE was incubated with purified MBP–NmB-polyST and increasing concentrations of the donor substrate CMP-Neu5Ac (0, 5, 50, 500 mM). After an incubation time of 30 min the synthesized reaction products were analysed by ion exchange HPLC. As shown in Fig. 6, the chain length of synthesized polySia increased with increasing concentra- © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1264 F. Freiberger et al. Fig. 5. PolySia biosynthesis of purified NmB-polyST. A. Purified NmB-polyST was assayed for 30 min in the presence of 0.1 mM pentameric a2,8-linked sialic acid (DP5) and 1 mM CMP-[14C]-Neu5Ac. Subsequently, half of the sample was digested with polySia-specific endoN. Radiolabelled reaction products were separated by acrylamide gel electrophoresis (10%) and detected by autoradiography. The following dyes were used as standards and correspond to polySia chain length given in brackets: trypan blue (DP100), xylene cyanol (DP52), bromophenol blue (DP19), bromcresol purple (DP11). B. Dependence of chain length distribution on acceptor concentration. The assay was performed as described in (A) but DP5 concentrations were varied as indicated. Samples were separated on a 25% polyacrylamide gel. To display single oligomeric-reaction products at the highest acceptor concentration more clearly, less sample volume was applied in lane 4. tions of donor substrate and resulted in a complete conversion of the acceptor substrate to high-molecularweight polymer at a CMP-Neu5Ac concentration of 500 mM. This clearly demonstrates synthesis of long polySia chains. To verify the suggested non-processive elongation mode of NmB-polyST (Fig. 5B and Fig. S1), we performed a time-course experiment under reaction conditions that, as shown above, result in synthesis of long polySia chains (Fig. 7). If the reaction was stopped after 2 min predominantly short oligosialic acid-containing products were found (Fig. 7A) that were further elongated as the polymerization progressed (Fig. 7B and C) to finally yield high-molecular-weight polymer (Fig. 7D). This is reflected by a shift of the detected product peaks to longer retention times (Fig. 7). The occurrence of intermediates in the polymerization reaction argues, in agreement with the data obtained from the radioactive assay system, for non-processive chain elongation by MBP–NmB-polyST and indicates that the enzyme dissociates from the product after each addition of sialic acid. Fig. 6. Dependence of GT3-FCHASE extension by purified NmB-PST on CMP-Neu5Ac concentration. Purified NmB-polyST was incubated with GT3-FCHASE and CMP-Neu5Ac for 5 min as described in Experimental procedures. The respective CMP-Neu5Ac concentrations are indicated. The reaction mixtures were then adjusted to 25% ethanol. The supernatants were applied to a DNA Pac PA-100 column and chromatographed with a gradient of NaNO3 according to Inoue et al. (2001) and Inoue and Inoue (2003). © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1265 Fig. 7. Time-course of elongation of GT3-FCHASE by NmB-polyST. Purified NmB-polyST (37 mg ml-1) was incubated with GT3-FCHASE and 0.5 mM CMP-Neu5Ac for (A) 2 min, (B) 5 min, (C) 10 min and (D) 30 min prior to adjusting the reaction mixtures to 25% ethanol. The supernatants were applied to a DNA Pac PA-100 column and chromatographed with a gradient of NaNO3 according to Inoue et al. (2001) and Inoue and Inoue (2003). Identification of two conserved motifs in bacterial sialyl- and polysialyltransferases To identify amino acid residues critical for (poly)sialyltransferase activity, we searched for common sequence motifs in the available bacterial sialyltransferase sequences. By iterative steps of pairwise and multiple alignments combined with visual inspection of the sequences, we identified two short motifs (D/E-D/E-G and HP). Both motifs are present in a range of enzymes with otherwise little identity (Fig. 8) that, based on functional and sequence properties, had been allocated to different CAZy families (GT-38 and GT-52) and to pfam family 05855. While GT-38 contains bacterial polysialyltransferases, GT-52 includes bacterial LOS sialyltransferases and pfam 05855 groups sialyltransferases similar to the sialyltransferase Lst of Haemophilus ducreyi (Bozue et al., 1999). The D/E-D/E-G and the HP motif identified in this study are conserved in all members of the three families and, interestingly, are also found in the bacterial LOS sialyltransferases grouped in GT-80. However, the relation to GT-80 family members is not easily seen in an alignment, as the stretch of sequence between the D/E-D/E-G and HP motif is approximately 50 amino acids longer than in the other families. D/E-D/E-G and HP motifs are crucial for NmB-polyST activity To investigate the functional relevance of the identified motifs, point mutations were introduced into NmB-polyST by site-directed mutagenesis. The histidine and proline of the HP motif as well as the glycine and its adjacent glutamate residue of the D/E-D/E-G motif were individually changed to alanine, resulting in the mutants H278A, P279A, E153A and G154A. Mutants were expressed in E. coli BL21(DE3) and lysates were analysed for protein expression and activity by Western blot and the radiochemical polyST activity assay, respectively. As shown in Fig. 9A, all mutants were expressed at the level of the wild-type protein and the ratio between soluble and insoluble protein was in all cases similar to wild type. In contrast, enzyme activity was dramatically impaired in all mutants (Fig. 9B). While mutations in the D/E-D/E-G motif (E153A and G154A) fully inactivated the protein, the H278A and P279A mutants of the HP motif maintained © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1266 F. Freiberger et al. Fig. 8. Conserved motifs in bacterial sialyl- and polysialyltransferases. Sequence alignment of bacterial sialyltransferases identifying two short motifs conserved in CAZy families GT-38 (bacterial polysialyltransferases), GT-52 (bacterial LOS-sialyltransferases) and in pfam family 05855 (similar to H. ducreyi sialyltransferase Lst). To improve clarity of the illustration only three representatives of pfam 05855 are shown. PolyST_NmB (AAA20478), polyST_EcK92 (AAA24215), polyST_EcK1 (CAA43053), 2,3ST_cpsK_Sa (EAO062164), 2,3-ST_Hd (AAD28703), SiaA_Lst_Hi (AAL38659), PM0508_Lst_Pm (AAK02592), LsgB_Hi (AAX88755), Cps8K_Sa (AAR29926), 2,3_ST_Nm (AAC44544), lst_Ng (AAY41933), lst_Ap (AAS66624), Psyc_0663_Pa (AAZ18522). Alignments were made using Multalign (Corpet, 1988). © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1267 Fig. 9. Analysis of NmB-polyST mutants in vivo and in vitro. A. Single-point mutations were introduced into the D/E-D/E-G and HP motif of NmB-polyST. Wild-type and mutant enzymes were expressed in E. coli and soluble (s) and insoluble (i) fractions of the bacterial lysates were analysed for expression by SDS-PAGE and Western blotting using anti-T7-tag antibody. B. PolyST activity was determined in protein lysates with similar expression levels by the radiochemical assay. Relative activities were calculated compared with the wild-type enzyme (100%). C. Mutant strains of N. meningitidis were generated by replacing the native polyST gene with mutant polySTs carrying the respective point mutations. Capsule expression was analysed with the polySia-specific antibody 735 by slide agglutination. D. Capsular polySia of mutant strains with residual capsule expression was quantified by ELISA using mab 735. Equal loading of the wells with NmB was controlled using mab P1.7 for detection of the meningococcal major outer membrane protein PorA. Each value represents the average of three independent determinations with the standard deviation indicated. residual activity (below 10% of wild type). However, when both residues of the HP motif were changed to alanine simultaneously (H278A/P279A) enzyme activity was abolished. To confirm the functional relevance of the D/E-D/E-G and HP motifs in vivo, mutant meningococcal strains were generated by homologous recombination and capsular phenotypes of the mutant strains were analysed by slide agglutination using the polySia-specific antibody 735. In agreement with the in vitro studies, Neisseria strains exhibiting the E153A, G154A or the double mutation H278A/P279A did not synthesize polySia capsules while capsules were still produced in strains carrying the single mutations H278A and P279A (Fig. 9C). However, a whole-cell ELISA used to quantitatively compare polySia synthesis in wild-type bacteria and mutants clearly demonstrated impairment of capsule biosynthesis in mutants with residual in vivo activity. While the PorA control was similar for wild type and mutants, the polySia signal was significantly reduced to 30% and 60% for the mutants carrying the amino acid exchanges H278A and P279A respectively. This demonstrates that mutations within the HP motif not only decrease enzymatic activity in vitro, but also reduce capsule production in vivo. An enzyme-linked assay system for the functional characterization of purified polysialyltransferases To compare wild-type and mutant polySTs with partial activity (particularly mutants H278A and P279A) we decided to perform kinetic studies. As a first step towards this goal, the glycosyltransferase assay developed by Gosselin et al. (1994) was adapted to assay polyST activity. This spectrophotometric assay links the release of CMP by polyST during polySia synthesis to NADH oxidation and thus enables continuous monitoring of the enzyme reaction. PolyST activity towards short oligosialic acid acceptors (DP2 to DP5) was measured at constant CMP-Neu5Ac and enzyme concentrations. As shown in Fig. 10A, efficient chain elongation required oligomers of at least DP3. These findings are in agreement with the acceptor dependence obtained when polyST-containing © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1268 F. Freiberger et al. vided a useful tool to carry out kinetic analysis of polySTs and greatly facilitated the acquisition of data which, until now, have required single-point measurements and elaborate detection methods. The HP motif is involved in CMP-Neu5Ac binding Fig. 10. Enzymatic characterization of NmB-polyST. A. To analyse acceptor specificity, oligomeric a2,8-linked sialic acids with a degree of polymerization (DP) ranging from DP2 to DP6 (0.5 mM) were assayed at constant enzyme (4 mM) and CMP-Neu5Ac concentrations (1 mM) using the continuous spectrophotometric assay. PolyST activity is detected as decreasing absorption at 340 nm (NADH oxidation). B. To determine Km and Vmax of NmB-polyST for the donor substrate CMP-Neu5Ac, measurements were performed at 30°C and 0.28 mg ml-1 colominic acid. The resulting substrate velocity curve and Lineweaver–Burk plot are depicted. Kinetic parameters were obtained by non-linear regression in Prism (GraphPad Software). Because NmB-polySTs carrying the point mutations H278A and P279A retained residual activity in vitro, kinetic properties of these mutants could be determined. Both mutants were expressed as MBP fusion proteins, purified and tested in the enzyme-linked polyST assay. As listed in Table 1, Vmax values for CMP-Neu5Ac were decreased by a factor of 4 (P279A) and 6 (H278A) with respect to the wild-type enzyme and, in both proteins, Km values for CMP-Neu5Ac were increased three- to fivefold. These data suggest that the HP motif of NmB-polyST is involved in binding of the donor substrate CMP-Neu5Ac. To further analyse the effect of both mutations on acceptor binding, Km values for colominic acid were determined. Interestingly, Michaelis constants were not significantly influenced by either mutation indicating that: (i) the HP motif does not vitally participate in acceptor binding, and (ii) the introduced point mutations did not cause major structural changes as the acceptor binding site appeared to be largely unaffected. Remarkably, the recently solved crystal structure of P. multocida sialyltransferase PmST1 (Ni et al., 2006) revealed a similar CMP-Neu5Ac binding function of the HP motif in the GT-80 sialyltransferase family. The enzyme consists of two Rossmann domains that form a deep cleft in which the active site is located and has been crystallized in the presence and absence of donor and acceptor substrates. The active site of PmST1 with bound donor analogue CMP-3F(a)Neu5Ac and acceptor lactose (Ni et al., 2007) is shown in Fig. 11 and illustrates that the histidine residue of the HP motif (H311 Table 1. Kinetic parameters of wild-type and mutant NmB-polyST. CMP-Neu5Aca membrane preparations of E. coli K1 or K92 were assayed in radiochemical or HPLC-based test systems (Steenbergen and Vimr, 1990; Ferrero et al., 1991; Chao et al., 1999). Subsequently, kinetic parameters of MBP–NmB-polyST were determined for the donor substrate CMP-Neu5Ac (Fig. 10B, Table 1). The calculated Km value of 0.42 mM is fivefold higher than that obtained for the membrane bound enzyme from E. coli K1 polyST (Kundig et al., 1971; Vijay and Troy, 1975). However, because apart from enzyme source and preparation but also assay and buffer conditions differ between earlier and current studies, further interpretation of the observed variations in Km appears unreasonable. In summary, the enzyme-linked assay pro- Colominic acidb NmB-polyST Vmax (mmol min-1 mg-1) Km (mM) Kmc (mM) Wild type H278A P279A 25.8 ⫾ 3.6 4.0 ⫾ 1.1 6.0 ⫾ 1.0 0.42 ⫾ 0.03 1.37 ⫾ 0.12 2.15 ⫾ 0.51 0.63 ⫾ 0.10 0.40 ⫾ 0.14 0.51 ⫾ 0.13 a. Determined at constant acceptor concentration of 0.28 mg ml-1 colominic acid. b. Determined at constant donor concentration of 1 mM CMPNeu5Ac. c. Kinetic constants were calculated based on an average chain length of 32 residues for colominic acid as estimated using the thiobarbituric acid assay procedure according to Skoza and Mohos (1976). Kinetic values were obtained using the continuous polysialyltransferase assay and are presented as the means ⫾ SE of three independent determinations. © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1269 Fig. 11. Location of the conserved HP and D/E-D/E-G motifs in the P. multocida sialyltransferase PmST1. A. Surface representation of PmST1 (PDB entry 2EX1). The motifs are coloured in red (D/E-D/E-G) and blue (HP) while bound CMP is shown in yellow stick representation. B. Active site view of PmST1. Amino acids that are part of the two motifs are labelled and depicted in green stick representation. Hydrogen bonds are shown as dotted line while the bound donor analogue CMP-3F(a)-Neu5Ac is depicted in yellow and the bound acceptor lactose is shown in cyan. Figures were generated with Pymol (http://www.pymol.org). in PmST1) is directly involved in CMP-Neu5Ac binding. It forms hydrogen bonds with the phosphate group of CMP and with the carboxylate function of the sialic acid moiety. Moreover, this structure shows that also the D/E-D/E-G motif is located directly at the active site cleft of PmST1 (DDG) and is involved in binding of donor and acceptor substrates. The second aspartic acid residue of the DDG sequence forms hydrogen bonds to the acceptor lactose and to the hydroxyl group in position C4 of the sialic acid moiety. In combination, the structural and biochemical data provide strong evidence that the D/E-D/E-G and HP sequences are crucial for CMP-Neu5Ac binding and enzyme catalysis, in bacterial sialyl- and polysialyltransferases that harbour these motifs. Discussion In this study, we functionally characterized the polysialyltransferase responsible for capsule biosynthesis in N. meningitidis serogroup B. Although polySTs are inter- esting therapeutic targets to combat these pathogens, no data on isolated proteins and virtually no information on structure–function relationships, which are crucial for the rational design of inhibitors, have been reported. Attempts to purify or solubilize membrane-associated polySTs have failed or resulted in inactivation of the enzymes, so that membrane association was proposed to be essential for polyST activity (Steenbergen and Vimr, 2003; Vionnet et al., 2006). With the aim of overcoming the lack in structure– function information, we started this study by screening for production systems that enable expression and purification of active NmB-polyST. Soluble expression of active protein was significantly improved after addition of large N-terminal fusion parts and allowed purification of NmBpolyST fused to MBP. Purified MBP–NmB-polyST was enzymatically active and, as described for membrane bound bacterial polySTs of E. coli K1 and K92 (Steenbergen and Vimr, 1990; Ferrero et al., 1991; Chao et al., 1999), able to synthesize long polySia chains from oligomeric primers of at least DP3. Consequently, our data provide clear evidence that membrane association is not a prerequisite for NmB-polyST activity. Furthermore, we demonstrate that the purified enzyme elongates polySia chains in a non-processive manner. This is in contrast to in vivo studies carried out in E. coli K1 that suggest a processive mode of polySia biosynthesis (Steenbergen and Vimr, 2003). PolySTs were proposed to be part of a large capsule biosynthesis complex, in which biosynthesis and translocation of polySia across the inner and outer bacterial membranes are tightly linked (Steenbergen and Vimr, 2003) and the functional E. coli K92 polysialyltransferase complex was found to be larger than a monomer (Vionnet et al., 2006). Thus, potential interaction partners of polyST could increase the efficiency of polySia biosynthesis and thereby increase processivity. We (W.F.V. and J.V.) recently demonstrated that also membrane-bound E. coli K92 polyST is a non-processive enzyme in vitro (Vionnet and Vann, 2007), but the utilized membranes lacked other gene products of the K92 capsule biosynthesis cluster as well. However, even polySia biosynthesis assayed with intact membrane preparations of E. coli K1, which most likely include all relevant interaction partners of polyST, was found to be less efficient with exogenously added sialyloligomers than with endogenous or lipid bound acceptors (Cho and Troy, 1994; Chao et al., 1999; McGowen et al., 2001). This may argue for a more effective binding of the endogenous acceptor and hence result in increased processivity. In conclusion, the finding that recombinant soluble NmB-polyST is non-processive does not exclude processivity of the enzyme in the living system. The investigation of functional properties of recombinant NmB-polyST in complex with other factors of the capsular biosynthetic machinery is therefore an impor- © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1270 F. Freiberger et al. tant aim in future studies. Interestingly in this regard, MBP–NmB-polyST expressed in a polyST-deficient Neisseria strain is capable to complement the defect (see Fig. 3). The fusion protein may thus provide an interesting tool for studies aimed at identifying polyST-mediated protein–protein interactions in vivo. Mapping of the minimal catalytic domain of NmBpolyST has shown that the C-terminal domain, though not conserved in the homologous E. coli enzymes, is essential for catalytic activity. All C-terminally truncated NmBpolySTs were completely inactive. Studies are underway to analyse if potential variations in folding, in the oligomerization status or in substrate binding properties are responsible for inactivation of the truncated polySTs (Fig. 2). To characterize the catalytic domain of NmB-polyST in more detail, we aimed at identifying key residues for polyST function. So far no functional residues have been described for bacterial polysialyltransferases and no sequence relationships to other bacterial sialyltransferases have been reported. In contrast to the eukaryotic sialyltransferases, which are grouped into a single CAZy family (GT-29) and harbour four highly conserved and well-described sialylmotifs (Drickamer, 1993; Livingston and Paulson, 1993; Geremia et al., 1997; Jeanneau et al., 2004), the less homologous bacterial sialyltransferases are found in four different CAZy families. GT-38 contains bacterial polysialyltransferases, while GT-42, GT-52 and GT-80 include bacterial LOS sialyltransferases. Heretofore, no conserved sequence features relating members of different CAZy families have been described. In the current study, we combined sequence alignments and visual inspections and identified two short motifs in bacterial sialyltransferases, the D/E-D/E-G motif and the HP motif. Both motifs are conserved throughout the CAZy families GT-38 and GT-52 and GT-80, with the only exception of a H. ducreyi sialyltransferase (AAP9506, HA instead of the HP), and are in addition found in pfam family 05855. The latter contains bacterial sialyltransferases similar to the Lst of H. ducreyi that have not yet been included in the CAZy database. Our mutagenesis studies of NmB-polyST demonstrate that the D/E-D/E-G motif is essential for polyST activity. Single-amino-acid substitutions within this motif (E153A and G154A) completely abolished enzyme activity in vitro and in vivo. In contrast, the simultaneous mutation of H278 and P279 of the HP motif to alanine is required to destroy polyST activity. Single mutations within this motif did not obstruct polyST activity but resulted in reduced capsule production and lowered catalytic efficiency of the enzyme as shown by in vivo studies and kinetic analysis of the purified mutants. Most interestingly, the HP motif appears to be involved in binding of the donor substrate as reflected by increased Km values for CMP-Neu5Ac in the H278A and P279A mutants. In contrast, binding of the acceptor was not influenced. The determined Michaelis constants for colominic acid were not significantly affected which furthermore argues against severe misfolding of these mutants. Support for the functional relevance of the D/E-D/E-G and the HP motif in bacterial sialyltransferases is furthermore provided by the recently solved crystal structure of the P. multocida sialyltransferase PmST1 (Ni et al., 2006). PmST1 is a member of CAZy family GT-80 and belongs to the glycosyltransferase-B structural superfamily. Interestingly, this fold was also predicted for the bacterial polysialyltransferases of GT-38 including NmB-polyST (Breton et al., 2006). PmST1 harbours both motifs (D/E-D/E-G and HP) close to its active site cleft, whereby the histidine residue of the HP motif (H311) forms a hydrogen-bond to the phosphate group of bound CMP. Mutation of this residue to alanine resulted in a 30-fold reduction of activity and a twofold increase in Km for CMP-Neu5Ac and it was suggested that H311 stabilizes the CMP leaving group in PmST1 catalysis (Ni et al., 2007). As the H278A mutation had similar effects on the kinetic properties of NmBpolyST, this residue may have a related function in polyST catalysis. So far no data on mutagenesis of the equivalent proline residue (P312) of PmST1 are available. However, mutation of this residue to alanine is likely to influence the structure of the loop harbouring the HP motif and may thereby displace the histidine residue. A similar effect may explain the reduced activity of the P279A mutation in NmB-polyST. Strikingly, residues of the D/E-D/E-G motif were also found to be essential for PmST1 function. Although the first aspartic acid residue (D140) of the motif points towards the protein core and is most likely not involved in substrate interactions, the second aspartic acid (D141) protrudes into the active site cleft of PmST1 (Fig. 11). Moreover, this residue was shown to interact with the PmST1 acceptor lactose and suggested to act as general base in PmST1 catalysis. Mutation of D141 to alanine virtually inactivated the enzyme (20 000-fold reduction of activity) (Ni et al., 2007). This is again in perfect agreement with the inactive E153A mutant of NmB-polyST and may suggest an analogous function of this residue in polyST catalysis. Indirect evidence that the protein stretch containing the conserved D/E-D/E-G motif is located close to the active site also in CAZy GT-52 family members was provided by Wakarchuk et al. (2001). They showed that the neisserial a2,3-sialyltransferase Lst switches to a bifunctional a2,3/6-sialyltransferase mode upon mutation of the single-residue G168 and predicted this residue to be positioned in an acceptor-binding cavity. Interestingly, G168 is found only two residues upstream of the D/E-D/E-G motif (residues 164–166) in this enzyme. Finally it should be mentioned that structural information is available for a second sialyltransferase the cst-II from C. jejuni (Chiu et al., 2004). However, this enzyme is a © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1271 member of CAZy family GT-42, which is the only bacterial sialyltransferase family so far that does not harbour HPand D/E-D/E-G motifs. In conclusion, the establishment of efficient expression, purification and assay procedures for NmB-polyST allowed us to identify and characterize key groups for polyST function. Alignments with bacterial sialyltransferases revealed two functional motifs, the D/E-D/E-G and the HP motif, that are conserved in otherwise unrelated bacterial sialyltransferases of CAZy families GT-38, GT-52 and GT-80 as well as in pfam 05855. The functional importance of both motifs for enzyme catalysis and/or CMP-Neu5Ac binding was demonstrated by mutational analysis of NmB-polyST and is emphasized by structural and biochemical data available for the P. multocida sialyltransferase PmST1. Our data therefore allow hypothesizing that basic features of substrate binding and enzyme catalysis are conserved in a wide range of bacterial sialyltransferases and improve the basis for design of sialyltransferase-specific drugs. Experimental procedures Materials Colominic acid and CMP-Neu5Ac were purchased from Sigma. Oligomeric sialic acids were from Nacalai Tesque. The pMAL-c Vector was from New England Biolabs, and pET vectors were purchased from Novagen. subcloned from pET43a-Strep-NmB-polyST using the NheI/ NotI sites resulting in pMBP-Strep-NmB-polyST that encodes for the polyST fused with an N-terminal MBP/Strep-tag and a C-terminal His6-tag. Generation of truncated proteins N-terminally truncated NmB-polySTs were generated by PCR using the forward primers IO07 (5′-GCA TGG ATC CAC ATC TCC ATT TTA TCT TAC-3′) for D23 NmB-polyST, IO08 (5′GCA TGG ATC CAA CAA TTT ATT TGT CAT ATC TA-3′) for D33 NmB-polyST and IO09 (5′-GCA TGG ATC CTT ATA TAC TTC TAA AAA CTT AAA A-3′) for D64 NmB-polyST and the reverse primer KS41 (5′-GCA GGC GGC CGC TCT ATC TCT ACC AAT TCT-3′). C-terminally truncated NmB-polySTs were generated by PCR using the reverse primers FF01 (5′-GCA TGC GGC CGC ATC TTT ACT ATG AAA GTC-3′) for NmB-polyST D22, FF02 (5′-GCA TGC GGC CGC CCC TAA TAA GGT AAT ATT G-3′) for NmB-polyST D60, KS335 (5′-G CAG GCG GCC GC TTC AAA TGT TTC TTC TGT TTT AAA-3′) for NmB-polyST D94, KS334 (5′-GC AG GCG GCC GC AAA TGT TTC TTC TGT TTT AAA GA-3′) for NmBpolyST D95 and FF02 (5′-GCA GGC GGC CGC TTC TTC TGT TTT AAA GAG AG-3′) for NmB-polyST D97 and the forward primer KS333 (5′-G CAT GGA TCC CTA AAG AAA ATAAAAAAA GCT CTT-3′). BamHI and NotI sites (underlined) in forward and reverse primers, respectively, were used for subcloning of the PCR products into the BamHI/NotI sites of pET43a-Strep. The resulting constructs encode for proteins with an N-terminal NusA-/Strep-tag and a C-terminal His6-tag. The identity of all constructs was confirmed by sequencing. Site-directed mutagenesis Cloning of NmB-polyST expression vectors NmB-polyST was amplified by PCR using plasmid pUE3 (Frosch et al., 1991) as template and the primer pair KS23 (5′-G CAT GGA TCC CTA AAG AAA ATA AAA AAA GCT-3′) and KS41 (5′-GCA GGC GGC CGC TCT ATC TCT ACC AAT TCT-3′) containing BamHI and NotI sites (underlined) respectively. The PCR product was subcloned into the respective sites of pET23a to result in an N-terminally T7- and C-terminally His6-tagged construct (pET23a-NmB-polyST). To generate an expression construct including an N-terminal NusA fusion part followed by a StrepII-tag, adapters Strepa (5′-CTA GTG CTA GCT GGA GCC ACC CGC AGT TCG AAA AAG GCG CCC TGG TTC CGC GTG-3′) and Strepas (5′GAT CCA CGC GGA ACC AGG GCG CCT TTT TCG AAC TGC GGG TGG CTC CAG CTA GCA-3′) were inserted by adapterligation into the BamHI and SpeI restriction sites of pET43a generating pET43a-Strep. The NmB-polyST gene was subcloned from pET23a-NmB-polyST using restriction sites BamHI and NotI to generate the expression vector pET43a-Strep-NmB-polyST that encodes for the polyST fused with an N-terminal NusA/Strep-tag and a C-terminal His6-tag. To obtain the vector pMBP-Strep, the MBP was amplified by PCR with primers FF03 (5′-GAT ATT CAT ATG AAA ACT GAA GAA GGT AAA CT-3′) and FF04 (5′-CAT ATA CTA GTC CTA CCC TCG ATG GAT CC-3′) from the vector pMAL-c and subsequently subcloned into the NdeI and SpeI restriction sites of pET43a-Strep. Finally, NmB-polyST was Single-point mutations of NmB-polyST were obtained by QuickChange site-directed mutagenesis (Stratagene) following the manufacturer’s instructions using the plasmid pET23a-NmB-polyST as template. Mutated polySTs were subsequently subcloned into the BamHI and NotI sites of pET23a resulting in expression of N-terminally T7-tagged proteins. The identity of all constructs was confirmed by sequencing. Mutagenic primers are given below with the mutated base triplets underlined: pET23a-G154A: KS52 (5′-G ACT CAT TTA ATT GAT GAA GCG ACT GGA ACA TAT GCT CC-3′) and KS53 (5′-GG AGC ATA TGT TCC AGT CGC TTC ATC AAT TAA ATG AGT C-3′); pET23a-E153A: KS54 (5′-T ACG ACT CAT TTA AT T GAT GCA GGG ACT GGA ACA TAT GC-3′) and KS55 (5′-GC ATA TGT TCC AGT CCC TGC ATC AAT TAA ATG AGT CGT A-3′); pET23a-H278A: KS56 (5′-ATT AAA GGA AAG ATA TTT ATT AAA CTA GCC CCA AAA GAG ATG GGC AAC AAC-3′) and KS57 (5′-GTT GTT GCC CAT CTC TTT TGG GGC TAG TTT AAT AAA TAT CTT TCC TTT AAT-3′); pET23a-P279A: KS58 (5′-GGA AAG ATA TTT ATT AAA CTA CAC GCA AAA GAG ATG GGC AAC AAC TA-3′) and KS59 (5′-TA GTT GTT GCC CAT CTC TTT TGC GTG TAG TTT AAT AAA TAT CTT TCC-3′). Mutants H278A and P279A were additionally introduced into the pMBP-Strep vector. Therefore, the polyST inserts of pET23a-H278A and pET23a-P279A were ligated into pET43a-Strep using restriction sites BamHI and NotI and subsequently subcloned into pMBP-Strep using the NheI and NotI sites. © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1272 F. Freiberger et al. Expression of recombinant NmB-PolyST To optimize protein production, bacteria were either cultivated at 30°C (only polySTs fused to NusA) or at 15°C. For production at 30°C, freshly transformed E. coli BL21(DE3) were grown in PowerBroth medium (Athena ES) containing 200 mg l-1 carbenicillin at 30°C and 225 r.p.m. At an optical density of OD600 = 1.8 expression was induced by adding 1 mM IPTG. Bacteria were harvested 3 h after induction by centrifugation (6000 g for 15 min, 4°C). For 15°C production, bacteria were grown at 30°C to an optical density of OD600 = 0.9. Cultures were then rapidly cooled (ice bath) and further grown at 15°C until OD600 = 1.8 was reached. Transgene expression was induced by the addition of 1 mM IPTG and cells were harvested 24 h after induction. All pellets were washed once with PBS and stored at -20°C. Separation of soluble and insoluble fractions of NmB-PolyST To analyse NmB-PolyST expression and enzymatic activity in bacterial lysates, cells were re-suspended in 50 mM Tris-HCl pH 8.0, 40 mM MgCl2 and lysed by sonication. Soluble and insoluble fractions were obtained following centrifugation (16 000 g, 15 min, 4°C). The insoluble fraction (pellet) was re-suspended in 50 mM Tris-HCl pH 8.0, 40 mM MgCl2 in a volume equal to that of the soluble fraction. Purification of recombinant MBP–NmB-PST fusion protein Bacterial pellets from 0.5 l of cultures were re-suspended in binding buffer (20 mM Tris; pH 7.4; 1 mM EDTA; 1 mM DTT; 25 mM NaCl) including protease inhibitors (40 mg ml-1 Bestatin, 1 mg ml-1 Pepstatin and 1 mM PMSF) to give a final volume of 20 ml. Cells were disrupted by sonication and samples were centrifuged (30 min; 16 000 g, 4°C) and filtered (Sartorius Minisart 0.8 mm). For affinity absorption of MBP–NmB-polyST, pre-swollen amylose resin (New England Biolabs) was added to the cleared supernatant and incubated at 4°C for 1 h. Subsequently, the incubation mixture was transferred to a column and washed with 12 volumes of binding buffer at a flow rate of 0.5 ml min-1. Bound protein was eluted with elution buffer (binding buffer containing 10 mM maltose). Fractions containing the fusion protein were pooled and passed through a desalting column (High Prep 26/10) equilibrated in 50 mM NaH2PO4 pH 8.0 and finally concentrated to 2 mg ml-1 using Amicon Ultra centrifugal devices (Millipore). To further enrich the MBP–NmB-polyST fusion protein, a gel filtration chromatography step was applied. Samples were filtered (Millipore Ultrafree MC 0.2 mm) and loaded on a Superdex 200 10/300 GL column (GE Healthcare). Proteins were eluted at a flow rate of 0.5 ml min-1 with 50 mM NaH2PO4 pH 8.0 buffer and fractions of 0.5 ml were collected. Obtained protein samples were stable for more than 3 weeks if stored at 4°C. Western blot analysis, proteins were blotted onto nitrocellulose (Whatman). Proteins containing an N-terminal StrepII-tag were detected by StrepTactin-alkaline phosphatase-conjugate (StrepTactin-AP; IBA) according to the manufacturer’s guidelines. His-tagged proteins were detected with 1 mg ml-1 penta-His antibody (Qiagen) followed by goat anti-mouse IgG-AP (Dianova). For quantification by infrared fluorescence detection, samples and standard proteins were blotted onto PVDF membranes (Millipore). Histagged proteins were detected with 1 mg ml-1 penta-His antibody (Qiagen) followed by 50 ng ml-1 goat anti-mouse IR680 antibody (LI-COR) and quantified according to the recommendations of the Odyssey infrared imaging system (LI-COR). Radiochemical polyST assays PolyST activity in bacterial lysates was analysed as described previously (Weisgerber et al., 1991). Briefly, 10 ml of lysate was mixed with 10 ml of TMD buffer (40 mM MgCl2, 50 mM, Tris/HCl pH 8.0) and 2 ml of colominic acid (100 mg ml-1) as acceptor and reactions were started by adding 2 ml of CMP-[14C]-Neu5Ac (13 mM, 1.55 mCi mmol-1). Samples were incubated at 37°C and 5 ml of aliquots were spotted on Whatman 3MM CHR paper after the respective reaction time. Following descending paper chromatography, the chromatographically immobile 14Clabelled polyST reaction products were quantified by scintillation counting. To analyse the products of purified NmBpolyST in more detail, reaction mixtures were analysed in the TBE-buffered (90 mM Tris, 90 mM borate, 2 mM EDTA, pH 8.3) electrophoresis system described for analysis of acidic capsular polysaccharides (Pelkonen et al., 1988). Enzyme reactions were carried out in TMD in a total volume of 24 ml as described above, using 10 mg of enzyme and 0.001–1 mM a2,8-linked sialic acid (DP5) as acceptor. Control samples were subsequently treated with 1 mg of polySia-degrading endoN (Stummeyer et al., 2005) for 20 min at 37°C. Equal volumes of sample buffer (2 M sucrose in TBE) were added and samples were electrophoresed at 4°C and 200 V overnight. To visualize 14C-labelled reaction products, gels were vacuum-dried immediately after electrophoresis and exposed to an imaging film (BioMax, Kodak). Fluorescent polyST assays PolyST activity of purified NmB-polyST was also monitored using the fluorescent acceptor GT3-FCHASE as acceptor (Vionnet and Vann, 2007). Briefly, GT3-FCHASE (0.23 mM), CMP-Neu5Ac (50–500 mM) and purified NmB-polyST (30– 180 mg ml-1) were incubated in TMD buffer (40 mM MgCl2, 50 mM, Tris/HCl pH 8.0) at 37°C. At the indicated time points (2–30 min) reactions were stopped by adjusting to 25% ethanol and samples were further analysed by HPLC as described (Vionnet and Vann, 2007). SDS-PAGE and immunoblotting Continuous spectrophotometric polysialyltransferase assay SDS-PAGE was performed under reducing conditions using 2.5% (v/v) b-mercaptoethanol and 1.5% (w/v) SDS. For For rapid characterization of purified polysialyltransferases, the glycosyltransferase testing system described by Gosselin © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1273 et al. (1994) was adapted to polysialyltransferases. All measurements were carried out in 96-half area well plates (Greiner Bio-one) in a total volume of 106.5 ml. In detail, a master solution was prepared containing the linking enzymes pyruvate kinase (16.5 U ml-1, Sigma), lactate dehydrogenase (23.5 U ml-1, Sigma) and nucleotide monophosphate kinase (0.5 U ml-1, Roche) in reaction buffer [1.5 mM ATP (Sigma), 1 mM PEP (Fluka), 0.13 mM NADH (Roche), 14 mM MgSO4, 56 mM KCl in 100 mM Tris pH 7.5]. As acceptor substrate either colominic acid or sialyloligomers were added in various concentrations. After addition of the donor substrate CMPNeu5Ac (62 mM to 2 mM) samples were first monitored at OD 340 nm until a stable baseline was reached (3–5 min). This was essential to metabolize free CMP, which is always present in CMP-Neu5Ac preparations due to hydrolysis of the substrate. Finally, PST was added (10–20 mg ml-1) and reactions were followed until the total amount of NADH was metabolized. Generation and analysis of mutant Neisseria strains For mutagenesis of N. meningitidis strains, the plasmids pET23a-G154A, pET23a-E153A, pET23a-H278A and pET23a-P279A needed to be modified to allow regular homologous recombination into the meningococcal chromosome and selection for the mutants. The siaD downstream region was amplified from serogroup B strain MC58 with primers GH149 (5′-GCG CGC CTC GAG AAT ACT ATG ACT TCT GA TCT CC-3′) and GH150 (5′-GCG CGC CTC GAG CGA GTA ATT TGA CAA TAG AGC G-3′) and integrated into each plasmid downstream of the mutated siaD gene using the respective XhoI sites (underlined). The resulting plasmids were linearized with NotI, blunt ended with T4 DNA polymerase and ligated with the kanamycin resistance cassette excised from pUC4K (GE Healthcare) by HincII. The final plasmids harbouring the kanamycin resistance cassette between the siaD gene and the siaD downstream region were used to transform serogroup B meningococcal strain MC58. Transformants were selected on GC agar supplemented with 100 mg ml-1 kanamycin. Recombination of the mutagenized motifs into the meningococcal siaD gene was verified by sequencing the PCR product obtained with primer pair GH157 (5′-CA GGC CAC TAC TCC TAT C TG-3′)/Kana2 (5′-GAT TTT GAG ACA CAA CGT GG-3′) with either primers GH157 and GH160 (5′-AGG TTC ATT AAT AAC TAC CAG C-3′) (D/E-D/E-G motif) or primers UE8a (5′-AA CGC TAC CCC ATT TCA-3′) and GH160 (HP motif). Furthermore, regular homologous recombination was confirmed by Southern blot hybridizations with the siaD gene and the kanamycin resistance gene used as a probe respectively. The meningococcal capsule phenotype was analysed with mab 735 by slide agglutination as described previously (Vogel et al., 2001). Quantitative analysis of capsule expression was performed by whole-cell ELISA as described (Vogel et al., 2001). Briefly, microtitre plates were pre-coated with poly D-lysine (25 mg ml-1 in PBS) for 1 h at room temperature. After three washing steps with PBS, bacterial suspensions (20 ml well-1, OD600 = 0.10 in PBS) were applied for 2 h and cross-linked to poly D-lysine by adding glutaraldehyde (0.05% in PBS) for 10 min. Plates were washed three times with PBS and non-specific binding sites were saturated by incubation with 1% BSA in PBS for 1 h. After three washing steps capsular polySia was detected by immunostaining using the polySia-specific mab 735. The amount of bacteria bound to each well was controlled with a parallel set of microplates using mab P1.7 directed against the PorA antigen of the meningococcal strain MC58 for detection. Subsequently, plates were incubated with peroxidase coupled secondary antibody (Dianova) and analysed by colour reaction. In vivo analysis of MBP–NmB-polyST Neisserial expression constructs were generated for in vivo comparison of wild type and MBP–NmB-polyST. The MBP– polyST insert of the E. coli expression construct pMBP-StrepNmB-polyST was amplified with primers HC574 (5′-GCG CGC TCT AGA GAA GGA GAT ATA CAT ATG AAA AC-3′) and HC572 (5′-GCG CGC GAT ATC TTA GTG GTG GTG GTG GTG G-3′) and integrated between the SpeI and EcoRV sites of the neisserial expression vector pAP1 (Lappann et al., 2006) resulting in pAP1-MBP–NmB-polyST. For comparison also the polyST gene without further tags was amplified with primers HC573 (5′-GCG CGC GAT ATC AGA GAT ACA ATA ATG CTA AAG AAA ATA AAA AAA GC-3′) and HC572 and integrated into the EcoRV site of pAP1 resulting in pAP1-NmBpolyST. Further in vivo studies were performed using strain 2517, an unencapsulated polyST knockout mutant of the meningococcal serogroup C strain 2120 (Ram et al., 2003) that was transformed with the resulting plasmids. Wild type and transformants were analysed for capsule expression by whole-cell ELISA as described above. Bacterial loading was controlled using anti-PorA antibody P1.2. Additionally, neisserial lysates were analysed by Western blot analysis. Bacteria were pelleted from 2 ml of suspension cultures (OD600 = 1.1), re-suspended in 100 ml of Lämmli sample buffer [100 mM Tris pH 6.8, 1.7% SDS, 16.5% glycerol (v/v), 2.5% 2-mercaptoethanol, bromophenol blue] and lysed by sonication. The lysates were centrifuged (16.000 g, 10 min, 4°C) and supernatants were divided in two aliquots. One aliquot was subsequently digested with 1 mg of endoN (Stummeyer et al., 2005) and incubated for 15 min at room temperature followed by a 15 min incubation at 37°C. All samples were incubated for 10 min at 60°C prior to electrophoresis. SDS-PAGE and Western blot were performed as described above. Acknowledgements We thank Andrea Bethe and Gabi Heinze for excellent technical assistance and Timothy Keys for critically reading the manuscript. We also wish to thank Warren Wakarchuk and Michel Gilbert for the kind gift of GT3-FCHASE. This work was supported by grants from the Deutsche Forschungsgemeinschaft (to R.G.-S., M.M. and U.V.) and the European Community (6th Framework programme PROMEMORIA, to R.G.-S.). References Andreishcheva, E.N., and Vann, W.F. (2006) Gene products required for de novo synthesis of polysialic acid in Escherichia coli K1. J Bacteriol 188: 1786–1797. © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 1274 F. Freiberger et al. Bhattacharjee, A.K., Jennings, H.J., Kenny, C.P., Martin, A., and Smith, I.C. (1975) Structural determination of the sialic acid polysaccharide antigens of Neisseria meningitidis serogroups B and C with carbon 13 nuclear magnetic resonance. J Biol Chem 250: 1926–1932. Bozue, J.A., Tullius, M.V., Wang, J., Gibson, B.W., and Munson, R.S., Jr (1999) Haemophilus ducreyi produces a novel sialyltransferase. Identification of the sialyltransferase gene and construction of mutants deficient in the production of the sialic acid-containing glycoform of the lipooligosaccharide. J Biol Chem 274: 4106–4114. Breton, C., Snajdrova, L., Jeanneau, C., Koca, J., and Imberty, A. (2006) Structures and mechanisms of glycosyltransferases. Glycobiology 16: 29R–37R. Chao, C.F., Chuang, H.C., Chiou, S.T., and Liu, T.Y. (1999) On the biosynthesis of alternating alpha-2,9/alpha-2,8 heteropolymer of sialic acid catalyzed by the sialyltransferase of Escherichia coli Bos-12. J Biol Chem 274: 18206– 18212. Chiu, C.P., Watts, A.G., Lairson, L.L., Gilbert, M., Lim, D., Wakarchuk, W.W., et al. (2004) Structural analysis of the sialyltransferase CstII from Campylobacter jejuni in complex with a substrate analog. Nat Struct Mol Biol 11: 163–170. Cho, J.W., and Troy, F.A. (1994) Polysialic acid engineering: synthesis of polysialylated neoglycosphingolipids by using the polysialyltransferase from neuroinvasive Escherichia coli K1. Proc Natl Acad Sci USA 91: 11427–11431. Claus, H., Vogel, U., Mühlenhoff, M., Gerardy-Schahn, R., and Frosch, M. (1997) Molecular divergence of the sia locus in different serogroups of Neisseria meningitidis expressing polysialic acid capsules. Mol Gen Genet 257: 28–34. Corpet, F. (1988) Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16: 10881–10890. Coutinho, P.M., Deleury, E., Davies, G.J., and Henrissat, B. (2003) An evolving hierarchical family classification for glycosyltransferases. J Mol Biol 328: 307–317. van Deuren, M., Brandtzaeg, P., and van der Meer, J.W. (2000) Update on meningococcal disease with emphasis on pathogenesis and clinical management. Clin Microbiol Rev 13: 144–166. Drickamer, K. (1993) A conserved disulphide bond in sialyltransferases. Glycobiology 3: 2–3. Edwards, U., Müller, A., Hammerschmidt, S., GerardySchahn, R., and Frosch, M. (1994) Molecular analysis of the biosynthesis pathway of the alpha-2,8 polysialic acid capsule by Neisseria meningitidis serogroup B. Mol Microbiol 14: 141–149. Egan, W., Liu, T.Y., Dorow, D., Cohen, J.S., Robbins, J.D., Gotschlich, E.C., and Robbins, J.B. (1977) Structural studies on the sialic acid polysaccharide antigen of Escherichia coli strain Bos-12. Biochemistry 16: 3687– 3692. Ferrero, M.A., Luengo, J.M., and Reglero, A. (1991) H.p.l.c. of oligo(sialic acids). Application to the determination of the minimal chain length serving as exogenous acceptor in the enzymic synthesis of colominic acid. Biochem J 280 (Part 3): 575–579. Frosch, M., and Müller, A. (1993) Phospholipid substitution of capsular polysaccharides and mechanisms of capsule for- mation in Neisseria meningitidis. Mol Microbiol 8: 483– 493. Frosch, M., Edwards, U., Bousset, K., Krausse, B., and Weisgerber, C. (1991) Evidence for a common molecular origin of the capsule gene loci in Gram-negative bacteria expressing group II capsular polysaccharides. Mol Microbiol 5: 1251–1263. Geremia, R.A., Harduin-Lepers, A., and Delannoy, P. (1997) Identification of two novel conserved amino acid residues in eukaryotic sialyltransferases: implications for their mechanism of action. Glycobiology 7: v–vii. Gosselin, S., Alhussaini, M., Streiff, M.B., Takabayashi, K., and Palcic, M.M. (1994) A continuous spectrophotometric assay for glycosyltransferases. Anal Biochem 220: 92–97. Gotschlich, E.C., Liu, T.Y., and Artenstein, M.S. (1969) Human immunity to the meningococcus. 3. Preparation and immunochemical properties of the group A, group B, and group C meningococcal polysaccharides. J Exp Med 129: 1349–1365. Hammerschmidt, S., Birkholz, C., Zähringer, U., Robertson, B.D., van Putten, J., Ebeling, O., and Frosch, M. (1994) Contribution of genes from the capsule gene complex (cps) to lipooligosaccharide biosynthesis and serum resistance in Neisseria meningitidis. Mol Microbiol 11: 885–896. Inoue, S., and Inoue, Y. (2003) Ultrasensitive analysis of sialic acids and oligo/polysialic acids by fluorometric highperformance liquid chromatography. Methods Enzymol 362: 543–560. Inoue, S., Lin, S.L., Lee, Y.C., and Inoue, Y. (2001) An ultrasensitive chemical method for polysialic acid analysis. Glycobiology 11: 759–767. Jeanneau, C., Chazalet, V., Auge, C., Soumpasis, D.M., Harduin-Lepers, A., Delannoy, P., et al. (2004) Structure– function analysis of the human sialyltransferase ST3Gal I: role of n-glycosylation and a novel conserved sialylmotif. J Biol Chem 279: 13461–13468. Kundig, J.D., Aminoff, D., and Roseman, S. (1971) The sialic acids. XII. Synthesis of colominic acid by a sialyltransferase from Escherichia coli K-235. J Biol Chem 246: 2543–2550. Lappann, M., Haagensen, J.A., Claus, H., Vogel, U., and Molin, S. (2006) Meningococcal biofilm formation: structure, development and phenotypes in a standardized continuous flow system. Mol Microbiol 62: 1292–1309. Livingston, B.D., and Paulson, J.C. (1993) Polymerase chain reaction cloning of a developmentally regulated member of the sialyltransferase gene family. J Biol Chem 268: 11504– 11507. McGowen, M.M., Vionnet, J., and Vann, W.F. (2001) Elongation of alternating alpha 2,8/2,9 polysialic acid by the Escherichia coli K92 polysialyltransferase. Glycobiology 11: 613–620. Mackinnon, F.G., Borrow, R., Gorringe, A.R., Fox, A.J., Jones, D.M., and Robinson, A. (1993) Demonstration of lipooligosaccharide immunotype and capsule as virulence factors for Neisseria meningitidis using an infant mouse intranasal infection model. Microb Pathog 15: 359–366. Masson, L., and Holbein, B.E. (1983) Physiology of sialic acid capsular polysaccharide synthesis in serogroup B Neisseria meningitidis. J Bacteriol 154: 728–736. Mühlenhoff, M., Stummeyer, K., Grove, M., Sauerborn, M., © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Functional motifs in bacterial sialyltransferases 1275 and Gerardy-Schahn, R. (2003) Proteolytic processing and oligomerization of bacteriophage-derived endosialidases. J Biol Chem 278: 12634–12644. Ni, L., Sun, M., Yu, H., Chokhawala, H., Chen, X., and Fisher, A.J. (2006) Cytidine 5′-monophosphate (CMP)-induced structural changes in a multifunctional sialyltransferase from Pasteurella multocida. Biochemistry 45: 2139–2148. Ni, L., Chokhawala, H.A., Cao, H., Henning, R., Ng, L., Huang, S., et al. (2007) Crystal structures of Pasteurella multocida sialyltransferase complexes with acceptor and donor analogues reveal substrate binding sites and catalytic mechanism. Biochemistry 46: 6288–6298. Pelkonen, S., Hayrinen, J., and Finne, J. (1988) Polyacrylamide gel electrophoresis of the capsular polysaccharides of Escherichia coli K1 and other bacteria. J Bacteriol 170: 2646–2653. Ram, S., Cox, A.D., Wright, J.C., Vogel, U., Getzlaff, S., Boden, R., et al. (2003) Neisserial lipooligosaccharide is a target for complement component C4b. Inner core phosphoethanolamine residues define C4b linkage specificity. J Biol Chem 278: 50853–50862. Shen, G.J., Datta, A.K., Izumi, M., Koeller, K.M., and Wong, C.H. (1999) Expression of alpha2,8/2,9polysialyltransferase from Escherichia coli K92. Characterization of the enzyme and its reaction products. J Biol Chem 274: 35139–35146. Skoza, L., and Mohos, S. (1976) Stable thiobarbituric acid chromophore with dimethyl sulphoxide. Application to sialic acid assay in analytical de-O-acetylation. Biochem J 159: 457–462. Steenbergen, S.M., and Vimr, E.R. (1990) Mechanism of polysialic acid chain elongation in Escherichia coli K1. Mol Microbiol 4: 603–611. Steenbergen, S.M., and Vimr, E.R. (2003) Functional relationships of the sialyltransferases involved in expression of the polysialic acid capsules of Escherichia coli K1 and K92 and Neisseria meningitidis groups B or C. J Biol Chem 278: 15349–15359. Stummeyer, K., Dickmanns, A., Mühlenhoff, M., GerardySchahn, R., and Ficner, R. (2005) Crystal structure of the polysialic acid-degrading endosialidase of bacteriophage K1F. Nat Struct Mol Biol 12: 90–96. Swartley, J.S., Marfin, A.A., Edupuganti, S., Liu, L.J., Cieslak, P., Perkins, B., et al. (1997) Capsule switching of Neisseria meningitidis. Proc Natl Acad Sci USA 94: 271–276. Troy, F.A. (1992) Polysialylation: from bacteria to brains. Glycobiology 2: 5–23. Vijay, I.K., and Troy, F.A. (1975) Properties of membraneassociated sialyltransferase of Escherichia coli. J Biol Chem 250: 164–170. Vimr, E.R., Bergstrom, R., Steenbergen, S.M., Boulnois, G., and Roberts, I. (1992) Homology among Escherichia coli K1 and K92 polysialytransferases. J Bacteriol 174: 5127– 5131. Vionnet, J., and Vann, W.F. (2007) Successive glycosyltransfer of sialic acid by Escherichia coli K92 polysialyltransferase in elongation of oligosialic acceptors. Glycobiology 17: 735–743. Vionnet, J., Kempner, E.S., and Vann, W.F. (2006) Functional molecular mass of Escherichia coli K92 polysialyltransferase as determined by radiation target analysis. Biochemistry 45: 13511–13516. Vogel, U., Claus, H., and Frosch, M. (2001) Capsular operons. In Meningococcal Disease: Methods and Protocols. Pollard, A.J., and Maiden, M.C. (eds). Totowa, NJ: Humana Press, pp. 187–201. Vogel, U., and Frosch, M. (1999) Mechanisms of neisserial serum resistance. Mol Microbiol 32: 1133–1139. Vogel, U., Weinberger, A., Frank, R., Müller, A., Kohl, J., Atkinson, J.P., and Frosch, M. (1997) Complement factor C3 deposition and serum resistance in isogenic capsule and lipooligosaccharide sialic acid mutants of serogroup B Neisseria meningitidis. Infect Immun 65: 4022–4029. Wakarchuk, W.W., Watson, D., St Michael, F., Li, J., Wu, Y., Brisson, J.R., et al. (2001) Dependence of the bi-functional nature of a sialyltransferase from Neisseria meningitidis on a single amino acid substitution. J Biol Chem 276: 12785– 12790. Weisgerber, C., Hansen, A., and Frosch, M. (1991) Complete nucleotide and deduced protein sequence of CMP-NeuAc: poly-alpha-2,8 sialosyl sialyltransferase of Escherichia coli K1. Glycobiology 1: 357–365. Whitfield, C. (2006) Biosynthesis and assembly of capsular polysaccharides in Escherichia coli. Annu Rev Biochem 75: 39–68. Supplementary material The following supplementary material is available for this article: Fig. S1. Elongation of trimeric a-2,8-linked sialic acid (DP3) by purified NmB-polyST. This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/ j.1365-2958.2007.05862.x (This link will take you to the article abstract). Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. © 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 65, 1258–1275 Characterization and Acceptor Preference of a Soluble Meningococcal Group C Polysialyltransferase Dwight C. Peterson, Gayathri Arakere, Justine Vionnet, Pumtiwitt C. McCarthy and Willie F. Vann J. Bacteriol. 2011, 193(7):1576. DOI: 10.1128/JB.00924-10. Published Ahead of Print 28 January 2011. These include: SUPPLEMENTAL MATERIAL REFERENCES CONTENT ALERTS Supplemental material This article cites 34 articles, 15 of which can be accessed free at: http://jb.asm.org/content/193/7/1576#ref-list-1 Receive: RSS Feeds, eTOCs, free email alerts (when new articles cite this article), more» Information about commercial reprint orders: http://journals.asm.org/site/misc/reprints.xhtml To subscribe to to another ASM Journal go to: http://journals.asm.org/site/subscriptions/ Downloaded from http://jb.asm.org/ on November 3, 2014 by MORGAN STATE UNIV Updated information and services can be found at: http://jb.asm.org/content/193/7/1576 JOURNAL OF BACTERIOLOGY, Apr. 2011, p. 1576–1582 0021-9193/11/$12.00 doi:10.1128/JB.00924-10 Copyright © 2011, American Society for Microbiology. All Rights Reserved. Vol. 193, No. 7 Characterization and Acceptor Preference of a Soluble Meningococcal Group C Polysialyltransferase䌤† Dwight C. Peterson,1 Gayathri Arakere,2 Justine Vionnet,1 Pumtiwitt C. McCarthy,1 and Willie F. Vann1* Laboratory of Bacterial Polysaccharides, Center for Biologics Evaluation and Research, Bethesda, Maryland 20892,1 and Sir Dorabji Tata Centre for Research in Tropical Diseases, Indian Institute of Science, Bangalore, India 5600122 Vaccines against Neisseria meningitidis group C are based on its ␣-2,9-linked polysialic acid capsular polysaccharide. This polysialic acid expressed on the surface of N. meningitidis and in the absence of specific antibody serves to evade host defense mechanisms. The polysialyltransferase (PST) that forms the group C polysialic acid (NmC PST) is located in the cytoplasmic membrane. Until recently, detailed characterization of bacterial polysialyltransferases has been hampered by a lack of availability of soluble enzyme preparations. We have constructed chimeras of the group C polysialyltransferase that catalyzes the formation ␣-2,9polysialic acid as a soluble enzyme. We used site-directed mutagenesis to determine the region of the enzyme necessary for synthesis of the ␣-2,9 linkage. A chimera of NmB and NmC PSTs containing only amino acids 1 to 107 of the NmB polysialyltransferase catalyzed the synthesis of ␣-2,8-polysialic acid. The NmC polysialyltransferase requires an exogenous acceptor for catalytic activity. While it requires a minimum of a disialylated oligosaccharide to catalyze transfer, it can form high-molecular-weight ␣-2,9-polysialic acid in a nonprocessive fashion when initiated with an ␣-2,8-polysialic acid acceptor. De novo synthesis in vivo requires an endogenous acceptor. We attempted to reconstitute de novo activity of the soluble group C polysialyltransferase with membrane components. We found that an acapsular mutant with a defect in the polysialyltransferase produces outer membrane vesicles containing an acceptor for the ␣-2,9-polysialyltransferase. This acceptor is an amphipathic molecule and can be elongated to produce polysialic acid that is reactive with group C-specific antibody. demiological tools. For instance, PCR assays based on the polysialyltransferase (PST) genes are routinely used for the detection and identification of serogroups (15). The polysialic acids are polymerized by a single polysialyltransferase in the case of each serogroup. The group B polysialyltransferase (NmB PST) is encoded by synD, and the group C PST is encoded by synE. Several investigators have cloned and begun to characterize the group B polysialyltransferase (9, 34). Both enzymes are membrane associated and are presumed to transfer sialic acid from the donor, CMP-NeuNAc, to the growing chain on the internal face of the cytoplasmic membrane of the bacterium (19, 21, 33). Much of the work on the biosynthesis of polysialic acid capsules has been done on the Escherichia coli K1 and K92 polysialyltransferases (21, 22). Like the meningococcal enzymes, the E. coli polysialyltransferases are associated with the cytoplasmic membrane and transfer sialic acid to the nonreducing end of the acceptor chain. Neither E. coli nor meningococcus can initiate synthesis de novo; they require an oligosialic acid or an endogenous acceptor (9, 17). The endogenous acceptor for these enzymes has yet to be identified. The meningococcal polysialyltransferase has a longer carboxyl-terminal sequence than the E. coli enzymes. The bacterial polysialyltransferases do not share motifs or sequence homologies with other sialyltransferases. E. coli and meningococcal polysialyltransferases belong to the CAZy glycosyltransferase family GT-38 (6). Until recently the characterization of the bacterial polysialyltransferases has been limited to studies with membrane Neisseria meningitidis groups B and C are the most common causes of meningococcal meningitis in adolescents and adults in Canada, Europe, and the United States. In the United States, 95% to 97% of cases of meningococcal disease are sporadic; however, since 1991, the frequency of localized outbreaks has increased (12, 13). Most of these outbreaks have been caused by serogroup C. Several vaccines based on the meningococcal capsular polysaccharides have been licensed. A tetravalent vaccine consisting of meningococcal groups A, C, Y, and W-135 has also been licensed. Subsequently, a conjugate vaccine of the same serogroup polysaccharide was licensed in the United States (3, 28). In addition there are two meningococcal group C conjugate vaccines licensed in Europe. These meningococcal capsular polysaccharides are polysialic acids. The N. meningitidis group B polysaccharide is an ␣-2,8linked polysialic acid, while the group C polysaccharide is an ␣-2,9-linked polysialic acid (see Fig. S1 in the supplemental material). The gene clusters responsible for the synthesis of these polysialic acids have been identified and characterized (5, 7, 10, 20, 24, 27). The glycosyltransferase genes of meningococcal gene clusters have been useful targets for the development of epi* Corresponding author. Mailing address: Laboratory of Bacterial Polysaccharides, Center for Biologics Evaluation and Research, Building 29, Room 103, 8800 Rockville Pike, Bethesda, MD 20892. Phone: (301) 496-2008. Fax: (301) 402-2776. E-mail: wvann@helix.nih.gov. † Supplemental material for this article may be found at http://jb .asm.org/. 䌤 Published ahead of print on 28 January 2011. 1576 Downloaded from http://jb.asm.org/ on November 3, 2014 by MORGAN STATE UNIV Received 6 August 2010/Accepted 17 January 2011 VOL. 193, 2011 MENINGOCOCCAL GROUP C POLYSIALYLTRANSFERASE 1577 TABLE 1. Plasmid constructs Plasmid Gene(s) synE-nusA pWV235 synE-SUMO pWV241 synE pWV237 synD pWV238 synD pWV245 synD-synE pWV246 synE-Halo tag pWV248 synE pWV249 synD-synE pWV250 synD-synE CPST LIC for CPST LIC rev CPST-SUMO Met for CPST-SUMO rev NmCPST fwd pENTR TOPO NmCPST BamHI rev NmB directional for NmB rev HindIII NmB directional for NmB rev HindIII DSnmB FOR DSnmB REV NmC PST Flexivector forward NmC PST Flexivector reverse NmCPST fwd pENTR TOPO MengC PST rev primer B322 primer C322 primer 151 primer C 151 primer D QCL NmC322 sense primer QCL NmC322 antisense primer fragments of cells harboring the polysialyltransferase genes or in vivo experiments (17, 23, 27, 29, 30). Soluble enzyme was not available for structure-function studies due to resistance of the membrane-associated enzymes to extraction in active form with detergents (17). The expression of some soluble membrane proteins has been achieved without detergents by fusion to proteins that promote solubilization. Recently, Freiberger et al. and Willis et al. (9, 34) demonstrated the ability to produce soluble group B polysialyltransferase as a chimera of synD and the malE or nusA gene. In our study, we constructed several soluble chimeras of the group C polysialyltransferase. The chimeric enzymes were expressed in E. coli and purified. The activity of the purified enzymes clearly demonstrated that only a single protein is required for elongation of polysialic acid acceptors. MATERIALS AND METHODS DNA manipulations. Recombinant DNA techniques were carried out using standard methods and commercially available materials. PCR amplifications were performed using either Taq Ready-to-Go beads (GE Healthcare) or a proofreading DNA polymerase, Phusion HF, purchased from New England BioLabs. Transformants were screened by either restriction digestion of plasmid minipreps or directly by PCR. Freshly picked colonies for PCR screening were boiled in diethyl pyrocarbonate water for 5 min and centrifuged. The supernatant was mixed with 1 ␮l of appropriate primers and Ready-to-Go PCR beads and then amplified in a thermocycler and analyzed on agarose gels. Construction of expression plasmids. Plasmids and primers used for our constructions are described in Table 1. pWV234 NusA-PST. The pWV234 NusA-PST plasmid encodes an aminoterminal NusA fusion with the NmC PST (synE). The synE gene was amplified from chromosomal DNA isolated from N. meningitidis group C strain P2181 with the forward primer CPST LIC FOR and the reverse primer CPST LIC REV and Taq Ready-to-Go PCR beads. The amplified fragment was purified and ligated into the expression vector pET44 Ek/LIC containing the nusA gene (Novagen) as described in the manufacturer’s instructions. pWV235 SUMO-PST. For WV235 SUMO-PST, the synE gene was amplified as described above by using plasmid pWV234 as template with forward primer CPST-SUMO Met FOR and reverse primer CPST-SUMO REV. The amplified Sequence GACGACGACAAGATGTTGCAGAAAATAAGA GAGGAGAAGCCCGGTTTAATTGGTTACAAAGGC ATGTTGCAGAAAATAAGAAAA TTATTGGTTACAAAGGCTATA CACCATGTTGCAGAAAATAAGAAAAGCT TTTGGATCCTTATTGGTTACAAAGGCTATATTT CACCATGCTAAAGAAAATAAAAAAAGC AAGCTTTTATCTATCTCTACCAATTCT CACCATGCTAAAGAAAATAAAAAAAGC AAGCTTTTATCTATCTCTACCAATTCT GAGGAAAACCTGTATTTTCAGGGCT TAAATGAGTCGTAATATTCTTTTTTTTTGC AGGAGCGATCGCCATGTTGCAGAAAATAAGAAA AACTGTTTAAACTTATTGGTTACAAAGGCTAT CACCATGTTGCAGAAAATAAGAAAAGCT TTATTGGTTACAAAGGC GCGGGTCTCTTTTTTAGGAGTTATATTATT GCGGGTCTCAAAAAATTCCTATATCCAGC GCGGGTCTCCATTCTACGGAAAACCTGTA CCCAGTGCTGCAATGATACCGCGAGA CCTAATAATATAACTCCTAAAAAATTCCTATATATCCA GCGTGGATATAAGA TCTTATATCCACGCTGGATATATAGGAATTTTTTAGG AGTTATATTATTAGG fragment was ligated in the pETSUMO TOPO plasmid as recommended by Invitrogen. pWV237. For pWV237, N. meningitidis group B synD was amplified from chromosomal DNA isolated from strain H355 using primers NmB rev HindIII and NmB directional for. The resulting fragment was ligated into the pET151 TOPO vector from Invitrogen. pWV241. For pWV241, the Gateway system of Invitrogen was used to construct a chimera of the maltose binding protein (MBP) gene, malE, and the NmC PST gene, synE. The synE gene was introduced into an entry vector as follows. The synE gene was amplified from pWV234 with the forward primer NmC PST FWD pENTR TOPO, the reverse primer NmC PST BamHI REV, and the thermostable polymerase reagent Phusion HF master mix from Finnzymes. The purified fragment was ligated into the Invitrogen pENTR/TEV/D-TOPO plasmid as recommended by the manufacturer. pWN602 malE Gateway destination vector. The plasmid pN-WWW was a gift of Warrren Wakarchuk, National Research Council, Canada, and is based on pCWOri⫹ from the same source. pN-WWW was digested with NdeI and HindIII and converted to blunt ends with mung bean nuclease and calf intestine alkaline phosphatase. The Gateway cassette B was inserted into the blunt end site as described by Invitrogen. The resulting plasmid, pWN602, contained the Gateway B cassette at the 3⬘ end of the malE gene. pWV243 MBP-PST. For pWV243 MBP-PST, the syn E (NmC PST) gene in pWV241 was used as the insert gene for the LR clonase (Invitrogen) reaction. pWN602 was used as the destination vector for pWV243. The protein expressed by this plasmid has an MBP at the amino-terminal end of NmC PST separated by a tobacco etch virus (TEV) protease cleavage site. pWV246. For pWV246, the HaloTag fusion plasmid was constructed by inserting the synE gene into the expression plasmid pFN22K HaloTag Flexivector (Promega). The synE gene was amplified from the plasmid pWV243 with the Phusion Hi-Fidelity enzyme mix (New England BioLabs) using the primers NmC PST Flexivector forward and NmC PST Flexivector reverse. pWV239. For pWV239, the NmB PST gene, synD, was cloned into the malEcontaining Gateway vector pWN602 in a fashion similar to that described for pWV243 above. The synD gene was first amplified using primers NmB directional for and NmB rev HindIII and then inserted in pENTR/TEV-TOPO. The resulting plasmid, designated pWV238, was combined with pWN602 in an in vitro recombination reaction mixture to yield pWV239. The resulting plasmid expresses a MalE-PST chimera with MalE and the amino terminus of NmCPST separated by a TEV protease cleavage site. pWV248. For pWV248, the NmC PST synE gene was amplified with pWV234 as PCR template and NmCPST fwd pENTR TOPO and MengC PST rev as Downloaded from http://jb.asm.org/ on November 3, 2014 by MORGAN STATE UNIV pWV234 Primer 1578 PETERSON ET AL. transformants were used to inoculate 2 ⫻ 1.5 liters of LB-kanamycin and grown to an A600 of 0.7, then induced with 1 mM IPTG for 3 h. The cells were harvested and frozen in 25 ml 50 mM Tris, 25 mM MgCl2 (pH 8) overnight. The cell suspension was lysed in a French pressure cell, and cellular debris was removed by centrifugation at 10,000 ⫻ g for 15 min. The lysate (25 ml) was labeled with 250 ␮l of TAMR-HaloTag substrate (10 to 20 ␮g/ml) as described by Promega. The membranes were separated from the soluble fraction by ultracentrifugation at 114,000 ⫻ g. The membrane pellet was resuspended in 4 ml cryoprotective buffer. Gel filtration of HaloTag-NmC PST. The soluble fraction described above was adjusted to 25% saturated ammonium sulfate and stirred at 4°C for 1 h. The pellet was collected by centrifuging at 15,000 rpm for 30 min and dissolved in 3 ml of 50 mM Tris, 25 mM MgCl2 (pH 8.0). Most of the enzyme activity in the soluble fraction was precipitated at 25% (NH4)2SO4 saturation. The enzyme fraction was analyzed on a Superose 12 fast-protein liquid chromatography (FPLC) column (GE) on a Dionex Summit high-performance liquid chromatograph (HPLC) equipped with a Hitachi model L-7485 fluorescence detector. TAMR-labeled fractions were detected at an excitation wavelength of 555 nm and emission at 585 nm. Peak fractions were concentrated on Nanosep microconcentrators and assayed for polysialyltransferase activity as described previously (17). Assay for polysialyltransferase. Polysialyltransferase activity was measured using a paper chromatography assay described previously (17). The linkage of the product of the chimera plasmids pWV245 and pWV250 was determined as follows. In a typical assay, the membrane fraction was incubated in a solution containing 22 mM Tris, 11 mM MgCl2, pH 8, 10 ␮g polysaccharide acceptor (NmC polysaccharide, colominic acid, or K92 polysaccharide) in 110 ␮l and 3 ␮l 100-␮Ci/ml 14C-labeled CMP-sialic acid (American Radiolabeled Chemicals). After a 1-h incubation, ␣-2,8-endoneuraminidase was added, and incubation at 37°C continued for 3 h, prior to analysis by paper chromatography. Control reaction mixtures were incubated for 3 h without endoneuraminidase. NmC PST competition assay with unlabeled acceptors. Lactosyl-boron dipyrromethene label (BODIPY) was sialylated with the bifunctional sialyltransferase CST II in a large-scale reaction volume (500 ␮l), and oligosialylated products were separated by anion exchange HPLC as described before (16). Product peaks were collected and desalted using Sep Pak chromatography (Waters). The extent of sialylation was confirmed by matrix-assisted laser desorption ionization–mass spectrometery (data not shown). Trisialylated lactosyl-BODIPY (2.5 ␮M) was incubated with NmC PST (30 ␮g/ml) in the presence of CMP-NeuNAc (5 mM) and MnCl (40 mM) in 200 mM sodium cacodylate buffer (pH 8.0) at 37°C for 40 min. For competition experiments, the reaction was run in the presence of sialic acid trimer, hexamer, or GD3 (Calbiochem) at a concentration of 5 ␮M. All reactions were quenched by the addition of ethanol to 25%. Products were dried by using a Speedvac, brought up to 50 ␮l with H2O, and subjected to HPLC analysis as previously described (18). Preparation of outer membrane vesicles. Outer membrane vesicles (OMV) were prepared from cultures of mutant N. meningitidis group B strains according to the procedure described by Frasch and Peppler (8). Cultures were grown overnight on tryptic soy broth supplemented with yeast extract as described by Arakere et al. (2). N. meningitidis strain M7 lacking the ability to synthesize sialic acid and a synD mutant lacking an active polysialyltransferase were obtained from David Stephens, Emory University (14). Detection of polysialic acid formation by ELISA. NmC PST (13 ␮l) was incubated with 10 ␮l OMV and CMP-NeuNAc (33, 330, 1,000, or 2,000 ␮M) in a total of 30 ␮l as described previously (17) for 1 h at 37°C. The assay mixture was then diluted to 400 ␮l with enzyme-linked immunosorbent assay (ELISA) coating buffer (2), which was then used to coat Immulon 1B plates (100 ␮l/well) at room temperature overnight. The plates were washed and then incubated with mouse monoclonal IgG antibody Mcl2075 (against OAc⫹) or Mcl2016 (against OAc⫺) group C polysaccharide at 1:400 dilution (kindly donated by Marjorie Shapiro, FDA) (11). Plates were developed with goat anti-mouse IgG–alkaline phosphatase conjugate (Sigma) diluted 1:2,000. RESULTS AND DISCUSSION Distribution and purification of PST activity. The polysialyltransferases of Neisseria meningitidis groups B and C are encoded by synD and synE, respectively. Experiments by Freiberger et al. (9) demonstrated that a soluble and active group B ␣-2,8polysialyltransferase could be expressed by some chimeras of synD. We prepared several chimeras of synE that were capable Downloaded from http://jb.asm.org/ on November 3, 2014 by MORGAN STATE UNIV primers. The resulting NmC PST synE was ligated into pET151TOPO as described by the manufacturer. Mutagenesis. The domain swap mutants of NmB and NmC PST were constructed by two methods. The plasmid pWV245 was constructed with the QuikChange II XL kit (Stratagene). The sequence encoding the amino terminus of SynD (nucleotides 1 to 439) was amplified using the primers DSnmB FOR and DSnmB REV with the NmB PST plasmid pWV239 as a template. The resulting fragment was purified and used as a mutagenic primer with the QuikChange II XL kit to mutate the NmC PST expression plasmid pWV243. This procedure requires significant areas of homology in the regions to be swapped. Because of this requirement for homology, the QuikChange method was not suitable for swapping the desired shorter region of synD. A modification of the method of Stemmer and Morris (25) was used to construct the other synD-synE chimera, although inverse PCR was not performed (4). Briefly, nucleotides 1 to 322 were amplified with a primer that included nucleotides 322 of synD in a Bsa1 cleavage site and a pET151 vector-based primer, including its Bsa1 site from pWV237. Similarly, nucleotides 322 to 1479 of synE were amplified with a primer that included nucleotides 322 of synE in a Bsa1 cleavage site and a synE primer that included its Bsa1 site from pWV248. The fragments were isolated, cut with Bsa1, and ligated together. The resulting hybrid fragment was ligated into the Bsa1 sites of pWV248 to yield plasmid pWV249. Plasmid pWV249 was sequenced and found to have a deletion error in the region of one of the primers used in its construction. The error was corrected using a QuikChange Lightning mutagenesis kit to give...