Structural characterization of the O-specific polysaccharide from the lipopolysaccharide of the fish pathogen Aeromonas bestiarum strain P1S

Carbohydrate Research 346 (2011) 815–821

Contents lists available at ScienceDirect

Carbohydrate Research journal homepage: www.elsevier.com/locate/carres

Structural characterization of the O-specific polysaccharide from the lipopolysaccharide of the fish pathogen Aeromonas bestiarum strain P1S Anna Turska-Szewczuk a,⇑, Leszek Guz b, Buko Lindner c, Hubert Pietras a, Ryszard Russa a, Otto Holst d a

Department of Genetics and Microbiology, M. Curie-Sklodowska University, Akademicka 19, 20-033 Lublin, Poland Department of Fish Diseases, Agriculture University, Akademicka 12, 20-033 Lublin, Poland c Division of Immunochemistry, Research Center Borstel, Parkallee 10, 23845 Borstel, Germany d Division of Structural Biochemistry, Research Center Borstel, Parkallee 4a/c, 23845 Borstel, Germany b

a r t i c l e

i n f o

Article history: Received 28 October 2010 Received in revised form 31 January 2011 Accepted 1 February 2011 Available online 4 March 2011 Keywords: 3-Hydroxybutyric acid Aeromonas bestiarum O-Specific polysaccharide Lipopolysaccharide

a b s t r a c t The O-specific polysaccharide obtained by mild-acid degradation of lipopolysaccharide of Aeromonas bestiarum P1S was studied by sugar and methylation analyses along with 1H and 13C NMR spectroscopy. The sequence of the sugar residues was determined using 1H,1H NOESY and 1H,13C HMBC experiments. The Ospecific polysaccharide was found to be a high-molecular-mass polysaccharide composed of tetrasaccharide repeating units of the structure

Since small amounts of a terminal Quip3N residue were identified in methylation analysis, it was assumed that the elucidated structure also represented the biological repeating unit of the O-specific polysaccharide. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Gram-negative bacteria of the genus Aeromonas are widespread in various habitats. They have been isolated from sources such as food, drinking water, sewage, environmental water, and human clinical specimens.1 Strains of the species Aeromonas hydrophila and Aeromonas caviae have been recognized as causative agents of sporadic diarrhea, dysentery, and life-threatening extra-intestinal infections in immuno-compromised patients and children.2–4 The varied clinical picture of Aeromonas infections indicates a multifactor model of pathogenesis.5 Cell-surface components such as outer membrane proteins, lipopolysaccharide (LPS),6 the S-layer,7 polar flagella, and pili4 have been identified as putative virulence factors, which play an important role in adhesion of the bacteria to epithelial cells, determining virulence and serum resistance as well as biofilm formation.2,4 LPS, which is a highly immunoreactive cell surface compound, has been implicated in the pathogenesis of Aeromonas strains.8 Studies of antigen expression in biofilm cells of A. hydrophila have revealed that a complete LPS structure is important for the ability of Aeromonas species to maintain tetragonal S-layer on the cell surface. On the other hand, changes in the LPS molecule can result in the loss of appropriate conformation of ⇑ Corresponding author. Tel.: +48 81 537 59 78; fax: +48 81 537 59 59. E-mail address: [email protected] (A. Turska-Szewczuk). 0008-6215/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.carres.2011.02.003

the protein structure.9 Aeromonas bestiarum and other species belonging to the motile aeromonad group, such as A. hydrophila, Aeromonas sobria, Aeromonas veronii, Aeromonas allosaccharophila, and Aeromonas jandaei, have been reported as fish pathogens.10– 12 The clinical manifestations of fish diseases range from dermal and ophthalmic ulcerations to more severe symptoms including soft tissue infections, hemorrhagic septicemia, motile aeromonad septicemia (MAS), or motile aeromonad infection (MAI).2–6,10–12 Outbreaks usually occur when a fish’s immune system is depressed due to an environmental stress such as overcrowding, poor water quality, organic pollution, and hypoxia in conjunction with other diseases.13,14 Aeromonas salmonicida subsp. salmonicida, a non-motile aeromonad, is the etiological agent of a bacterial septicemia in salmonid fish, called furunculosis. Furunculosis is an important disease in wild and cultured stocks of salmonid and other fish species and can have significant negative economic impacts on aquaculture operations. Most Aeromonas species are opportunistic pathogens, entering through wounds or affecting only stressed fish. A. salmonicida subsp. salmonicida, however, is a specific pathogen capable of causing diseases in healthy salmonid fish at a very low level of infection.15,16 A comparative analysis of sequenced genomes of A. hydrophila strain ATCC 7966 and A. salmonicida strain A449 provided an opportunity to identify genes involved in host invasion and virulence. The studies revealed that genes determining the synthesis

816

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

of LPS are located adjacent to clusters of genes encoding several types of adhesins (e.g., the surface layer protein VapA, flagella, and pili), which are important colonization factors.16 Moreover, a structural analysis of O-specific polysaccharides (OPSs) of A. hydrophila, A. salmonicida, and A. caviae strains,17–20 supported by serological studies could provide an explanation of the host– pathogen interactions contributing to the disease state.8,9 The recently determined OPS structure of A. bestiarum strain 207, which consists of branched pentasaccharide-repeating units containing four L-rhamnose and one D-glucosamine residues,21 complements the current knowledge of the compositional diversity of O-antigens among Aeromonas strains. This study reports the structural characterization of the OPS of A. bestiarum strain P1S isolated in the course of motile aeromonad septicemia11 in a Polish carp farm. Results of some taxonomic studies have revealed that diseases, and thus losses, in commercial aquacultures have mostly been caused by strains within this genomospecies.11,12 2. Results and discussion The LPS of A. bestiarum P1S was isolated by the hot phenol– water procedure22 from enzymatically-digested bacterial cells. Its SDS–PAGE profile (Fig. 1) revealed that the most prominent bands appeared in the region corresponding to the rough Ra chemotype of Salmonella LPS and were accompanied by much less intense slow-moving bands representing smooth (S-form) LPS. The OPS was released by mild-acid degradation of the LPS and isolated, in the void volume, by gel-permeation chromatography (GPC). The low yield of the OPS, making up only 8% of the total amount of the material eluted from the column, confirmed that the R-type LPS dominated in A. bestiarum P1S cells, and only a minor part of the material was of the S-type. GLC–MS analyses of alditol acetates obtained after complete acid hydrolysis of the OPS revealed ribose (Rib), 3-amino-3,6-dideoxyglucose (Qui3N), 2-amino-2,6-dideoxygalactose (FucN), and galactose (Gal) residues in a relative peak area ratio of 1:1.4:1.1:2. Qui3N and FucN were identified by comparing their retention times and mass spectra with those of authentic samples isolated from the O-specific polysaccharides of Escherichia coli O523 (a strain obtained from the Institute of Immunology and Experimental Therapy, Wroclaw, Poland) and Acinetobacter baumannii strain 34,24 respectively. The GLC–MS analysis

Figure 1. Silver-stained SDS–Tricine PAGE of the LPSs of A. bestiarum strain P1S (lane 2, 4 lg; lane 3, 2 lg) and Salmonella enterica sv. Typhimurium (Sigma) as reference (lane 1, 2 lg).

also showed the presence of a small peak corresponding to a product of 3,6-dideoxy-3-(3-hydroxybutyramido)-glucose (Qui3NAcyl). The substitution of Qui3N with 3-hydroxybutyrate was further confirmed by NMR analysis. The absolute configurations of the monosaccharides25 determined by GLC of the acetylated (S)-2-butyl glycosides showed that Rib, Gal, Qui3N, and FucN had the D-configuration, and a GLC–MS analysis of trimethylsilylated (R)and (S)-2-octyl esters obtained by a modified method26 demonstrated the presence of (R)-3-hydroxybutyrate. The D-configuration of Qui3N was confirmed by GLC coinjection experiments, an analysis of glycosylation effects in a 13C NMR spectrum of the studied polysaccharide (see below), and by comparison with other published data.23,27,28 Methylation analysis of the OPS identified 1,2,4-tri-Oacetyl-3,5-di-O-methyl-ribitol, 1,4,5-tri-O-acetyl-2,3,6-tri-O-methylgalactitol, 1,2,5-tri-O-acetyl-4-O-methyl-3-(N-methylacetamido)-3, 6-dideoxyglucitol, and 1,3,5-tri-O-acetyl-4-O-methyl-2-(N-methylacetamido)-2,6-dideoxygalactitol, elucidating the glycosidic substitutions of ribofuranose (Ribf) at C-2, Galp at C-4, Quip3N at C-2, and FucpN at C-3. Moreover, among the methylation analysis products a small amount of 1,5-di-O-acetyl-2,4-di-O-methyl-3-(Nmethylacetamido)-3,6-dideoxyglucitol was detected, identifying a terminal Quip3N residue. The 13C NMR spectrum of the A. bestiarum P1S OPS (Fig. 2a) demonstrated that the polymer had a regular structure composed of tetrasaccharide repeating units. It contained signals for four anomeric carbons at d 107.65, 105.83, 104.78, and 96.85 (labeled A, D, C, and B, respectively) with an integral intensity ratio of 1.1:1:1:1. Signals were also found for two nitrogen-bearing carbons (FucN C2 and Qui3N C-3) at d 48.42 and 56.35, two methyl groups of 6deoxysugars (FucN and Qui3N) at d 16.19 and 17.85, one N-acetyl group (CH3 at d 23.16 and CO at d 175.06), one N-(3-hydroxybutyryl) group (CH3 at d 22.87, CO at d 175.19, as well as CH–OH and CH2 at d 65.6 and 45.94, respectively), and 15 other non-anomeric sugar ring carbons in the region d 61.76–84.28, some of which overlapped. The presence of a signal at d 83.02, which is typical of C-4 of a furanose ring, confirmed the occurrence of a ribofuranose residue.29,30 Consistent with this, the 1H NMR spectrum of the OPS (Fig. 2b) contained, inter alia, signals for four anomeric protons at d 5.62, 5.53, 4.89, and 4.42, labeled A through D, respectively. In the high field region of the spectrum, there were also signals originating from the methyl groups of FucN and Qui3N at d 1.25 and 1.31, respectively, one signal of an N-acetyl group at d 2.05 (singlet), one at d 1.23 (CH3), two at d 2.35 and 2.48 (CH2, both m), and one at d 4.21 (CH–OH) of N-(3-hydroxybutyryl), the chemical shifts of which were consistent with published data.28,31 The 1H and 13C NMR spectra of the OPS were assigned using 1 1 H, H, DQF-COSY, TOCSY, NOESY, 1H,13C HSQC, and 1H,13C HMBC experiments. All chemical shifts are summarized in Table 1. The TOCSY spectrum of the OPS revealed spin systems of b-Rib (A), a-FucN (B), b-Qui3N (C), and b-Gal (D), and the data of the COSY and NOESY (Fig. 3a) experiments enabled a differentiation among protons within each spin system. N-Acetamido sugars were identified by correlations of the protons H-2 at d 4.32 (for FucN) and H-3 at d 3.96 (for Qui3N) to the corresponding carbon-bearing nitrogen at d 48.42 and 56.35, respectively, as revealed by a 1H,13C HSQC experiment (Fig. 3b). The same spectrum also allowed the complete assignment of the chemical shifts of all corresponding proton–carbon resonances (Table 1). In particular, the characteristic downfield shift of the anomeric proton signal of residue A at dH 5.62 (s, J1,2 1 Hz) and the corresponding carbon resonance at dC 107.65 identified b-D-Ribf.29 In support of this, a correlation between H-1 and C-4 at dH/dC 5.62/83.02 was found in the 1H,13C HMBC spectrum, a finding previously observed for the furanosidic ring form.23,30 The relatively small J1,2 coupling constant value of

817

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

3.6 Hz and the corresponding 1JC-1,H-1 coupling (177 Hz, Fig. 4) showed that FucN (unit B) was a-linked, whereas the b-configuration of both, Qui3N and Gal (C and D) was established by J1,2 (7.7 Hz) and 1JC-1,H-1 (161 Hz) coupling constant values. The latter conclusion was confirmed by NOE contacts between the H-1,H-3 and H-1,H-5 pairs of protons of the b-linked sugar residues observed in the NOESY spectrum (Fig. 3a). In the TOCSY spectrum, the spin system of Qui3N was unique due to correlations of each, H-1 to H-6 with all other protons of this residue, as well as to the large J2,3, J3,4, and J4,5 coupling constant values (9.8 Hz). The spin systems of signals B and D, attributed to Gal and FucNAc, were characterized by cross-peaks of H-1 with H-2, H-3, and H-4, as observed in the TOCSY spectrum. All other assignments of the chemical shifts of the galacto sugar residues were completed by the NOE connectivities and COSY cross-peaks. Low-field positions of the signals for C-2 of Ribf (d 84.28), C-3 of FucN (d 78.33), C-2 of Qui3N (d 75.06), and C-4 of Gal (d 76.85), as compared with the chemical shifts of the corresponding nonsubstituted monosaccharides, elucidated the glycosylation pattern of the sugar residues.23,27–30 The sequence of the sugar residues in the repeating unit was determined by 1H,1H NOESY and 1H,13C HMBC experiments. In the NOESY spectrum of the OPS (Fig. 3a), the following strong NOE contacts were observed: b-Rib H-1 (A), b-Gal H-4 (D, d 5.62/ 3.98); b-Gal H-1 (D), a-FucN H-3 (B, d 4.42/3.79); a-FucN H-1 (B), b-Qui3N H-2 (C, d 5.53/3.48), and b-Qui3N H-1 (C), b-Rib H-2 (A, d 4.89/4.20). These data were confirmed by a 2D H-detected heteronuclear multiple-bond (1H,13C HMBC) experiment (Fig. 4), which showed the following inter-residue cross-peaks: b-Rib H-1 (A), b-Gal C-4 (D, d 5.62/76.85); b-Gal H-1 (D), a-FucN C-3 (B, d 4.42/78.33); a-FucN H-1 (B), b-Qui3N C-2 (C, d 5.53/75.06), and b-Qui3N H-1 (C), b-Rib C-2 (A, d 4.89/84.28). The site of attachment of the 3-hydroxybutyryl group (acyl) at the amino group of Qui3N (Qui3NAcyl) was confirmed by the correlation of its C-1 with H-3 of Qui3N at dC/dH 175.19/3.96 (1H,13C HMBC spectrum). The acetyl group was amide-bound to FucN, which was confirmed by the correlation of its C-1 with H-2 of the carbon-bearing nitrogen at dC/dH 175.06/4.32. ESI FT-ICR MS analysis enabled in-source fragmentation of the native O-chain polysaccharide, and the obtained data were consistent with those presented above. The charge deconvoluted ESI FTICR MS spectrum (negative-ion mode) of the OPS fraction showed mass peaks at 2867.19, 3579.48, 4291.77, 5004.11, 5716.35, 6428.72, and 7141.01 u, which differed by Dm1 = 712.3 u (n = 5– 10, Fig. 5), a value corresponding to the calculated molecular mass of one chemical repeating unit composed of Gal-FucNAc-Qui3NAcyl-Rib. These peaks originated from the cleavage of two furanosidic linkages between Ribf and Galp residues at both the reducing and the terminal end of OPS. In addition, the spectrum contained a second series of mass peaks at 3230.39, 3942.67, 4654.94, 5367.26, and 6079.55 u (n + Dm2), which were assigned to the mass of repeating unit oligomers with an additional ‘‘cap’’ fragment of 363.15 u (Dm2). The sequences of mass peaks at

Figure 2. 13C NMR (90 MHz) (a) and 1H NMR spectra (600 MHz) (b) of the OPS of A. bestiarum strain P1S. Capital letters and Arabic numerals refer to atoms in the sugar residues denoted as shown in Table 1. NAc, N-acetyl group (dC 23.16; dH 2.05); IS, acetone as internal standard (dC 31.07; dH 2.225); Hb, 3-hydroxybutyrate. Spectra were recorded at 40 °C in D2O as a solvent.

Table 1 1 H and 13C NMR chemical shifts of the constituents of the repeating unit of A. bestiarum P1S OPS (d in ppm) H-1

C-1

H-2

C-2

H-3

C-3

H-4

C-4

H-5a, H-5b

C-5

?2)-b-D-Ribf-(1?

Data A

5.62

107.65

4.20

84.28

4.21

70.28

3.96

83.02

3.63, 3.85

63.30

?3)-a-D-FucpNAc-(1?

B

5.53

96.85

4.32

48.42

3.79

78.33

3.97

72.07

3.96

67.50

1.25

16.19

?2)-b-D-Quip3NAcyl-(1?

C

4.89

104.78

3.48

75.06

3.96

56.35

3.20

73.87

3.59

74.02

1.31

17.85

?4)-b-D-Galp-(1?

D

4.42

105.83

3.47

71.47

3.73

73.48

3.98

76.85

3.69

75.16

3.71–3.73

61.76

175.06 175.19

2.05 2.35; 2.48

23.16 45.94

4.21

65.60

1.23

22.87

CH3CON (R)-3-Hydroxybutyrate

H-6

C-6

818

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

Figure 3. Parts of a 2D 1H,1H NOESY spectrum (a) and 1H,13C HSQC spectrum (b) of the OPS of A. bestiarum strain P1S. The map (a) shows NOE contacts between anomeric protons and protons at the glycosidic linkages (underlined). Some other H/H correlations are also depicted. The corresponding parts of the 1H and 13C NMR spectra are shown along the horizontal and vertical axes, respectively (b). Capital letters and Arabic numerals refer to atoms in the sugar residues denoted as shown in Table 1.

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

819

Figure 4. Part of a 2D 1H,13C HMBC spectrum of the OPS of A. bestiarum strain P1S. The map shows correlations between anomeric protons and carbons at the glycosidic linkages (boldface). Some other H/C correlations are also depicted (italic). Arabic numerals refer to H/C pairs in sugar residues denoted by capital letters as shown in Table 1. The coupling constant values 1JC-1,H-1 presenting a- or b-anomeric configurations of the sugar residues are given at the bottom of the spectrum. 1JC-1,H-1 for A (179 Hz), B (177 Hz), C (160 Hz), and D (161 Hz).

Figure 5. Part of a charge deconvoluted ESI FT-ICR MS (negative ion mode) of the OPS of A. bestiarum strain P1S. The spectrum shows mass peaks for two types of oligomers, the first one originating from the cleavage of two linkages between Ribf and Galp residues at both ends of OPS (bold numerals) and the second (italic numerals) originating from the cleavage of the furanosidic linkage at the reducing end of OPS alone. Dm1 = 712.3 u corresponds to the molecular mass of the OPS repeating unit; Dm2 = 363.15 u corresponds to the terminal fragment composed of Qui3NAcyl and Rib; n = 4,5,6,7,8,9,10 is the number of chemical repeating units in an OPS fragment.

3579.48 and 3942.67 u, corresponding to the molecular mass of five chemical repeating units (M) and five units with an additional terminal disaccharide (M + Dm2), are shown below. The patterns

of chemical ‘‘{ }’’and biological repeating units ‘‘[ ]’’ were marked with brackets; the capital letters G, F, R, and Q correspond to Gal, FucN, Rib, and Qui3NAcyl, respectively.

820

M

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

fG-F-Q-Rgn¼5

M þ Dm2 ½Q -R-fG-F-½Q-Rgn¼5 The 363.15 Da (M + Dm2) mass increment, corresponding to the Qui3NAcyl and Rib disaccharide located at the non-reducing end of the OPS, resulted from degradation of one, instead of two, furanosidic linkages between Ribf and Galp residues during mild acid hydrolysis. A previous report shows32 that under these hydrolytic conditions, furanosidic linkages within LPS are easily degraded. Hydrolysis at the reducing end of the OPS produced only small amounts of heptoses originating from the core region. The identification of a disaccharide containing a Qui3NAcyl-Rib residue at the terminal position of OPS was consistent with methylation analysis data and confirmed the composition of the biological repeating unit, in which Qui3N had been thought to represent the non-reducing sugar residue of the OPS. Consequently, the reducing sugar in the O-unit was a FucNAc residue, whose transfer to an undecaprenol carrier could have initiated biosynthesis of the OPS. Based on all data obtained, it was concluded that the biological repeating unit of the OPS isolated from A. bestiarum P1S had the linear structure:

can act as new immunoactive determinants and thus result in increased chemical and serological diversity of bacterial antigens. Although it is not clearly evident which structural determinants are the most important for virulence, it has been found that some serotypes are far more frequently associated with certain infections.23,35 In general, studies on the structures of bacterial O-antigens have been of importance to the understanding of the role of those molecules in serological specificity and pathogenesis. 3. Experimental 3.1. Bacterial strain and LPS isolation and degradation The A. bestiarum strain designated as P1S was isolated from an outbreak of motile aeromonad septicemia (MAS) in a carp farm in Poland. It belongs to a group of strains virulent for carp, as previously reported.12 For LPS analysis, A. bestiarum strain P1S was obtained from the collection of the Department of Fish Diseases, National Veterinary Research Institute (Pulawy, Poland). The bacteria were cultivated in tryptic soy broth (TSB) at 28 °C for 72 h. The cells were harvested by low speed centrifugation (8000g, 20 min). The recovered bacterial cell pellet was washed twice with

→2)-β-D-Quip3NAcyl-(1→2)-β-D-Ribf-(1→4)-β-D-Galp-(1→3)-α-D-FucpNAc-(1→ C

A

D

B

The presence of a terminal Qui3N residue in the O-antigen repeating unit also made it possible to identify the linking sugar which binds the OPS to the core region as a FucNAc residue. Although the FucNAc was a-linked in the repeats, a b-linkage between the O-chain and the core may be suggested, consistent with earlier reports on other LPS.31 For instance, in the OPSs of Hafnia alvei 1185, 1199, and 1205 strains, the first O-units were linked to the core by a b-D-GlcpNAc residue even though they had the a-configuration in the following O-units.33 These findings suggest a similarity between the synthesis of A. bestiarum P1S OPS and the Wzy-dependent pathway synthesis of heteropolymeric O-antigens from Salmonella enterica serovars Typhimurium and Anatum. Structural analyses of OPS of several Aeromonas strains revealed their heteropolymeric character. For example, the OPS of A. salmonicida strain A449 was composed of a Rha-ManNAc backbone with a glucose residue and an O-acetyl group as substituents.18 The OPS of A. caviae strains 11212 and ATCC 15468 contained deoxyhexoses, hexoses, and aminohexoses as compounds building pentasaccharide and tetrasaccharide repeating units, respectively.19,20 Similarly, an A. hydrophila O:34 OPS repeating unit was built up of 6deoxy-L-talose, mannose, and GalNAc residues.17 It is worth noting that the structure of the A. bestiarum P1S OPS is different from those published for other Aeromonas strains, but resembles, to some extent, that of E. coli O5.23 The similarities include the presence of b-D-Ribf, b-D-Quip3N, and b-D-Galp while the differences between these OPSs include the presence of a-FucpNAc instead of a-GalpNAc, and the different position of the linkages of Ribf and Quip3N. Moreover, while the amino group of Quip3N identified in E. coli O5 OPS is acetylated, the one studied here carries a 3-hydroxybutyryl group. Various derivatives of uncommon monosaccharides, such as 3-acetamido-3,6-dideoxy-D-glucose and 4-acetamido-4,6-dideoxy-D-glucose N-acylated with 3-hydroxybutyryl groups, have been reported as components of bacterial polysaccharides isolated from cells of Pseudomonas fluorescens strain IMV 247 and Pseudoalteromonas haloplanktis strain KMM 223.34 Atypical sugar residues and other, randomly distributed noncarbohydrate substituents present in O-specific polysaccharides

0.5 M saline and once more with distilled water. Bacterial cells (7.5 g dry weight) were digested with lysozyme, RNAse, and DNAse (24 h, 1 mg/g), and then with Proteinase K (36 h, 1 mg/g) in 50 mM phosphate buffer (pH 7.0) containing 5 mM MgCl2. The suspension was dialyzed against distilled water and freeze-dried. The digested cells were extracted three times with aq 45% phenol at 68 °C,22 and the separated layers were dialyzed against tap and distilled water. The LPS (317 mg) recovered only from the water phase was purified by ultracentrifugation at 105,000g and freeze-dried to give a yield of 4.2% of bacterial dry weight. The OPS was obtained by mild acid hydrolysis of the LPS (120 mg) with 2% acetic acid at 100 °C for 3 h, followed by separation of the high-molecular-mass fraction by GPC on a Sephadex G-50 column (1.8  80 cm) using 1% acetic acid as an eluent and monitoring with a Knauer differential refractometer. The yield of the OPS fraction was 8% of the LPS weight. 3.2. Chemical analyses The OPS of A. bestiarum strain P1S was hydrolyzed with 2 M CF3CO2H (120 °C, 2 h). Monosaccharides were identified by GLC– MS of alditol acetates36 on a Hewlett–Packard HP5890A-HP5971 instrument equipped with a capillary column (HP-5MS, 30 m  0.25 mm), applying a temperature gradient of 150 °C (5 min) to 310 °C at 5 °C min1. The absolute configuration of monosaccharides was determined by GLC of acetylated (R)- or (S)-2-butyl glycosides using authentic sugars as standards.25 The absolute configuration of 3-hydroxybutyric acid was determined according to the procedure of Kenne et al.,26 modified as follows: after hydrolysis of the OPS in 2 M CF3CO2H (120 °C, 4 h), the product was extracted with EtOAc (3  3 mL), evaporated under nitrogen, and subjected to solvolysis with 2 M HCl in (R)-2-octanol. The reaction was carried out at 85 °C for 12 h. The mixture was then concentrated to dryness and trimethylsilylated derivatives were obtained. The product had the same retention time as the authentic octyl (R)-3-trimethylsilyloxybutyrate, but was well-separated from the corresponding (S) derivative. GLC–MS data of the product and the derivative of the standard, were identical.

A. Turska-Szewczuk et al. / Carbohydrate Research 346 (2011) 815–821

Methylation analysis of the OPS was performed according to the Hakomori method.37 The permethylated OPS was subjected to hydrolysis in 2 M CF3CO2H (120 °C, 2 h), N-acetylation, and reduction with NaBD4. Partially methylated alditols were converted into acetate derivatives and analyzed by GLC–MS as above. 3.3. NMR spectroscopy 1D 1H NMR and 2D NMR experiments were recorded in a D2O solution at 300 K using a Bruker Avance DRX-600 MHz spectrometer and standard Bruker software. Chemical shifts were reported relative to internal acetone as reference (dH 2.225 ppm, dC 31.07 ppm). The following standard homo- and heteronuclear correlated two-dimensional techniques were used for general assignments: DQF-COSY, TOCSY, NOESY, 1H,13C HSQC, and 1H,13C HMBC. The 1JC,H anomeric coupling constant values were measured in the 1H,13C HMBC experiment carried out with a low-pass filter set to a low frequency value as described previously.38 1D 13C NMR was measured in D2O at 300 K using a Bruker DPX360 MHz spectrometer. 3.4. Electrospray ionization Fourier-transform ion cyclotron resonance mass spectrometry (ESI FT-ICR MS) All experiments were performed in the negative ion mode by using a hybrid Apex Qe FT-ICR MS instrument (Bruker Daltonics, Billerica, MA, USA), equipped with a 7 T superconducting magnet and an Apollo dual ion source. Data were recorded in the broadband mode with a 512 k data sampling rate. The mass scale was calibrated externally by using compounds of known structure. For the negative ion mode, samples (10 ng/lL) were dissolved in a 50:50:0.001 (v/v/v) mixture of 2-propanol, water, and triethylamine. Samples were sprayed at a flow rate of 2 lL/min. The capillary entrance voltage was set to 3.8 kV, and the drying gas temperature was set to 150 °C. The spectra, which showed several charge states for each component, were charge deconvoluted, and the mass numbers given refer to the mono-isotopic molecular masses. 3.5. SDS–PAGE LPS preparations were separated in 12.5% SDS–Tricine polyacrylamide electrophoresis gel,39 and bands were visualized by silver staining after oxidation with periodate according to the method of Tsai and Frasch.40 Acknowledgments This work was financially supported by the Grant of the ViceRector of M. Curie-Sklodowska University in Lublin, Poland—the Polish Ministry of Science and Higher Education research fund, (Grant No. BW-01-1000-14-09). We thank Herman Moll and Regina Engel for their help with GLC–MS, Heiko Käßner for recording the NMR spectra, and Katarzyna A. Duda for her interest and valuable advice during this study (all at Research Center Borstel, Germany).

821

References 1. Araujo, R. M.; Arribas, R. M.; Pares, R. J. Appl. Bacteriol. 1991, 71, 182–186. 2. Palumbo, S. A.; Bencivengo, M. M.; Del Corral, F.; Williams, A. C.; Buchanan, R. L. J. Clin. Microbiol. 1989, 27, 854–859. 3. Altwegg, M. Infection 1985, 13, 228–230. 4. Rabaan, A. A.; Gryllos, I.; Tomas, J. M.; Shaw, J. G. Infect. Immun. 2001, 69, 4257– 4267. 5. Cahill, M. M. J. Appl. Microbiol. 1990, 69, 1–16. 6. Merino, S.; Rubires, X.; Aguillar, A.; Guillot, J. F.; Tomas, J. M. Microb. Pathog. 1996, 20, 325–333. 7. Cipriano, R.; Bertolini, J. J. Wildl. Dis. 1988, 24, 672–678. 8. Dooley, J. S. G.; Lallier, R.; Shaw, D. H.; Trust, T. J. J. Bacteriol. 1985, 164, 263– 269. 9. Asha, A.; Nayak, D. K.; Shankar, K. M.; Mohan, C. V. Fish Shellfish Immunol. 2004, 16, 429–436. 10. Ali, A.; Carnahan, A. M.; Altwegg, M.; Luthy-Hottenstein, J.; Joseph, S. W. Med. Microbiol. Lett. 1996, 5, 156–165. 11. Kozinska, A.; Figueras, M. J.; Chacon, M. R.; Soler, L. J. Appl. Microbiol. 2002, 93, 1034–1041. 12. Kozinska, A.; Guz, L. Fish Shellfish Immunol. 2004, 16, 437–445. 13. Stevenson, R. M. W. In Fish Vaccination; Ellis, A. E., Ed.; Academic Press: New York, 1988; pp 112–123. 14. Zhang, Y. L.; Ong, C. T.; Leung, K. Y. Microbiology 2000, 146, 999–1009. 15. Daly, J. G.; Kew, A. K.; Moore, A. R.; Olivier, G. Microb. Pathog. 1996, 21, 447– 461. 16. Reith, M. E.; Singh, R. K.; Curtis, B.; Boyd, J. M.; Bouevitch, A.; Kimball, J.; Munholland, J.; Murphy, C.; Sarty, D.; Williams, J.; Nash, J. H. E.; Johnson, S. C.; Brown, L. L. BMC Genomics 2008, 9, 427–442. 17. Knirel, Y. A.; Shashkov, A. S.; Senchenkova, S. N.; Merino, S.; Tomas, J. M. Carbohydr. Res. 2002, 337, 1381–1386. 18. Wang, Z.; Liu, X.; Dacanay, A.; Harrison, B. A.; Fast, M.; Colquhoun, D. J.; Lund, V.; Brown, L. L.; Li, J.; Altman, E. Fish Shellfish Immunol. 2007, 23, 1095–1106. 19. Linnenborg, M.; Wildmalm, G.; Rahman, M. M.; Jansson, P.-E.; Holme, T.; Qadri, F.; Albert, M. J. Carbohydr. Res. 1996, 291, 165–174. 20. Wang, Z.; Liu, X.; Li, J.; Altman, E. Carbohydr. Res. 2008, 343, 483–488. 21. Turska-Szewczuk, A.; Kozinska, A.; Russa, R.; Holst, O. Carbohydr. Res. 2010, 345, 680–684. 22. Westphal, O.; Jann, K. Methods Carbohydr. Chem. 1965, 5, 83–91. 23. MacLean, L. L.; Perry, M. B. Biochem. Cell Biol. 1997, 75, 199–205. 24. Vinogradov, E. V.; Pantophlet, R.; Dijkshoorn, L.; Brade, L.; Holst, O.; Brade, H. Eur. J. Biochem. 1996, 239, 602–610. 25. Leontein, K.; Lindberg, B.; Lönngren, J. Carbohydr. Res. 1978, 62, 359–362. 26. Kenne, L.; Lindberg, B.; Rahman, M. M.; Mosihuzzaman, M. Carbohydr. Res. 1993, 243, 131–138. 27. Lipkind, G. M.; Shashkov, A. S.; Knirel, Y. A.; Vinogradov, E. V.; Kochetkov, N. K. Carbohydr. Res. 1988, 175, 59–75. 28. Shashkov, A. S.; Paramonov, N. A.; Veremeychenko, S. P.; Grosskurth, H.; Zdorovenko, G. M.; Knirel, Y. A.; Kochetkov, N. K. Carbohydr. Res. 1998, 306, 297–303. 29. Shashkov, A. S.; Vinogradov, E. V.; Knirel, Y. A.; Nifant’ev, N. E.; Kochetkov, N. K.; Dabrowski, J.; Kholodkova, E. V.; Stanislavsky, E. S. Carbohydr. Res. 1993, 241, 177–188. 30. Kocharova, N. A.; Ovchinnikova, O. G.; Torzewska, A.; Shashkov, A. S.; Knirel, Y. A.; Rozalski, A. Carbohydr. Res. 2006, 341, 786–790. 31. Müller-Loennies, S.; Holst, O. In Comprehensive Glycoscience. From Chemistry to Systems Biology; Kamerling, J. P., Boons, G.-J., Lee, Y. C., Suzuki, A., Taniguchi, N., Voragen, A. G. J., Eds.; Elsevier Ltd.: Oxford, 2007; Vol. 1, pp 123–179. 32. Berst, M.; Hellerqvist, C. G.; Lindberg, B.; Lüderitz, O.; Svensson, S.; Westphal, O. Eur. J. Biochem. 1969, 11, 353–359. 33. Katzenellenbogen, E.; Kocharova, N. A.; Zatonsky, G. V.; Shashkov, A. S.; Bogulska, M.; Knirel, Y. A. FEMS Immunol. Med. Microbiol. 2005, 45, 269–278. 34. Hanniffy, O. M.; Shashkov, A. S.; Senchenkova, S. N.; Tomshich, S. V.; Komandrova, N. A.; Romanenko, L. A.; Knirel, Y. A.; Savage, A. V. Carbohydr. Res. 1998, 307, 291–298. 35. Valvano, M. A. Can. J. Microbiol. 1992, 38, 711–719. 36. Russa, R.; Urbanik-Sypniewska, T.; Lindström, K.; Mayer, H. Arch. Microbiol. 1995, 163, 345–351. 37. Hakomori, S. J. Biochem. (Tokyo) 1964, 55, 205–208. 38. Leone, S.; Molinaro, A.; Lindner, B.; Romano, I.; Nicolaus, B.; Parrilli, M.; Lanzetta, R.; Holst, O. Glycobiology 2006, 16, 766–775. 39. Lesse, A. J.; Campagnari, A. A.; Bittner, W. E.; Apicella, M. A. J. Immunol. Methods 1990, 126, 109–117. 40. Tsai, C. M.; Frasch, C. E. Anal. Biochem. 1982, 119, 115–119.