Structure and genetics of the O-antigen of Cronobacter turicensis G3882 from a new serotype, C. turicensis O2, and identification of a serotype-specific gene

Authors

  • Yamin Sun,

    1. The Key Laboratory of Molecular Microbiology and Technology of the Ministry of Education, Nankai University, Tianjin, China
    2. TEDA School of Biological Sciences and Biotechnology, Nankai University, TEDA, Tianjin, China
    Search for more papers by this author
  • Nikolay P. Arbatsky,

    1. N.D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, Russia
    Search for more papers by this author
  • Min Wang,

    1. The Key Laboratory of Molecular Microbiology and Technology of the Ministry of Education, Nankai University, Tianjin, China
    2. TEDA School of Biological Sciences and Biotechnology, Nankai University, TEDA, Tianjin, China
    Search for more papers by this author
  • Alexander S. Shashkov,

    1. N.D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, Russia
    Search for more papers by this author
  • Bin Liu,

    1. The Key Laboratory of Molecular Microbiology and Technology of the Ministry of Education, Nankai University, Tianjin, China
    2. TEDA School of Biological Sciences and Biotechnology, Nankai University, TEDA, Tianjin, China
    Search for more papers by this author
  • Lei Wang,

    Corresponding author
    1. TEDA School of Biological Sciences and Biotechnology, Nankai University, TEDA, Tianjin, China
    • The Key Laboratory of Molecular Microbiology and Technology of the Ministry of Education, Nankai University, Tianjin, China
    Search for more papers by this author
  • Yuriy A. Knirel

    1. N.D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, Moscow, Russia
    Search for more papers by this author

Correspondence: Lei Wang, TEDA School of Biological Sciences and Biotechnology, Nankai University, TEDA, #23 HongDa Street, Tianjin 300457, China. Tel.: +186 22 66229591; fax: +186 22 66229596; e-mail: wanglei@nankai.edu.cn

Abstract

Lipopolysaccharides on the cell surface of Gram-negative bacteria are an important factor in pathogenicity, and the O-specific polysaccharide chain (O-polysaccharide, O-antigen) defines the immunospecificity of different bacterial strains. Cronobacter turicensis is an emerging foodborne pathogen which causes severe invasive infections in neonates. In this study, a new O serotype, C. turicensis O2, was established, the structure and genetics of the O-antigen were investigated, and a serotype-specific gene was identified. Sugar and methylation analyses, and nuclear magnetic resonance spectroscopy indicated that the O-polysaccharide contains d-galactose (d-Gal), N-acetyl-d-glucosamine (d-GlcNAc), l-rhamnose (l-Rha) and 5,7-diacetamido-3,5,7,9-tetradeoxy-d-glycero-d-galacto-non-2-ulosonic acid (di-N-acetyllegionaminic acid, Leg5Ac7Ac). The structure of the tetrasaccharide repeat of the O-polysaccharide was established as:

display math

The O-antigen gene cluster of C. turicensis O2 was sequenced and compared with related gene clusters from available databases. Putative genes for the synthesis of l-Rha and Leg5Ac7Ac, and genes encoding sugar transferases and O-antigen processing genes (wzx and wzy) were found. The tentatively assigned functions of the O-antigen genes were in agreement with the structure of the O-polysaccharide. In addition, primers based on the wzy gene were shown to be specific for C. turicensis O2 in a screen against 145 strains.

Introduction

Cronobacter (formerly Enterobacter sakazakii) is a novel genus containing six genomospecies: Cronobacter sakazakii, Cronobacter malonaticus, Cronobacter turicensis, Cronobacter muytjensii, Cronobacter genomospecies 1 and Cronobacter dublinensis (Iversen et al., 2008). Cronobacter spp. are considered emerging opportunistic pathogens which can cause severe infections in neonates and lead to necrotizing enterocolitis, sepsis and neonatal meningitis. In addition, recent reports have highlighted the risk posed to immunocompromised adults with bacteremia and osteomyelitis by Cronobacter spp. (Corti et al., 2007).

Lipopolysaccharide (LPS) located on the cell surface is an important factor in the determination of bacterial pathogenicity, and its O-specific polysaccharide chain (OPS, O-antigen) defines immunospecificity in bacteria. The O-antigen is extremely variable, with variations in the types of sugars present, their arrangement within the oligosaccharide (OS) repeat (O-unit), and the linkages between the OS units. The O-antigen diversity is mainly due to variations in the cluster of genes for the synthesis of OPS, which are classified into three groups: (1) genes for the synthesis of nucleotide sugars; (2) genes for the transfer of sugars from their respective nucleotide donors to synthesize an O-unit; and (3) O-antigen processing genes for O-unit flippase (wzx) and O-antigen polymerase (wzy), which are responsible for the translocation of O-units across the membrane and their specific assembly, respectively. The O-antigen is subject to intense selection by the host immune system and other environmental factors, such as bacteriophages, which may account for the maintenance of diverse O-antigen forms within a species (Reeves, 1995).

Up to now, seven C. sakazakii O serotypes (O1–O7) have been established (Sun et al., 2011), five of their O-antigen gene clusters have been characterized (Mullane et al., 2008; Arbatsky et al., 2010b, 2011; Jarvis et al., 2011; Shashkov et al., 2011), and the chemical structures of the O-antigen in five O serotypes and related strains have been elucidated (Szafranek et al., 2005; MacLean et al., 2009b; Arbatsky et al., 2010a,b; Czerwicka et al., 2010; MacLean et al., 2010; Arbatsky et al., 2011; Shashkov et al., 2011). In addition, one O serotype of both C. turicensis and C. muytjensii, and two O serotypes of C. malonaticus have been established (Jarvis et al., 2011), their O-antigen gene clusters have been characterized (Jarvis et al., 2011), and the chemical structures of one O-antigen from each species have been determined (MacLean et al., 2009a,c, 2011).

In this work, we report on the chemical structure and gene cluster of an O-antigen from a new C. turicensis O serotype, C. turicensis O2. A rarely occurring higher acidic sugar, di-N-acetyllegionaminic acid, was isolated by mild acid hydrolysis of LPS and identified as a component of the O-unit. The O-antigen gene cluster of C. turicensis O2 was sequenced and analyzed, and the putative functions of the genes in the OPS biosynthesis were inferred and found to be in agreement with the structure of the OPS. In addition, a specific gene for the C. turicensis O2 serotype was identified, and a specific primer pair for this gene was obtained.

Materials and methods

Bacterial strain, cultivation and LPS isolation

Strain G3882 was isolated from shrimp strips collected at the Hunan Entry-Exit Inspection and Quarantine Bureau of China. Chromogenic agar, 16S rRNA gene sequencing, and phenotypic characterization were used to identify the strain as C. turicensis. Restriction fragment length polymorphism mapping of the O-antigen gene clusters showed the gene cluster of G3882 is quite different from that of C. turicensis O1 (Jarvis et al., 2011), which suggested that G3882 is likely a new serotype (data not shown). The bacteria were grown in 8 L Luria–Bertani medium using a 10 L fermentor (Biostat C10; B. Braun Biotech International, Germany) under constant aeration at 37 °C and pH 7.0, and bacterial cells were washed and dried as previously described (Robbins & Uchida, 1962).

LPS was isolated using the phenol–water procedure (Westphal & Jann, 1965), followed by dialysis of the extract without layer separation and purification as previously described (Arbatsky et al., 2010a). The yield of LPS was ~ 8% of the dried cell mass.

Degradations of the LPS

An LPS sample (90 mg) was heated with 2% HOAc for 1.5 h at 100 °C, and a lipid precipitate was removed by centrifugation. Gel-permeation chromatography of the supernatant on a column (68 × 3.4 cm) of Sephadex G-50 Superfine (Amersham Biosciences, Sweden) in 0.1% HOAc monitored using a UV-detector (Uvicord, Sweden) at 206 nm afforded a low molecular-mass material (42 mg). The subsequent fractionation on a column (98 × 4 cm) of Sephadex G-25 under the same conditions gave a monosaccharide (Leg5Ac7Ac, 3.2 mg) and an OS (15.5 mg).

An LPS sample (115 mg) was heated with 12.5% aqueous ammonia for 16 h at 37 °C, a precipitate was removed by centrifugation, and the supernatant was fractionated by gel-permeation chromatography on Sephadex G-50 Superfine as above to yield an O-deacylated LPS preparation (math formula, 56 mg).

A sample of math formula (15 mg) was oxidized with 0.05 M NaIO4 (2 mL) at 20 °C for 64 h in the dark, reduced with NaBH4 at 20 °C for 16 h, an excess of NaBH4 was destroyed with concentrated HOAc, the solution was evaporated, 10% HOAc in methanol was added and evaporated (five times), and the residue was applied to a column (110 × 1.2 cm) of Sephadex G-25 in 0.1% HOAc. The isolated polymer was hydrolyzed with 2% HOAc (100 °C, 1 h) and OS (2.4 mg) was isolated by gel-permeation chromatography on the same column.

Monosaccharide analysis

An OS sample (1 mg) was hydrolyzed with 3 M CF3CO2H (120 °C, 3 h) and sugars were identified using gas-liquid chromatography (GLC) of the alditol acetates derived conventionally from the released monosaccharides. The absolute configurations of the monosaccharides were determined by GLC of the acetylated (S)-2-octyl glycosides (Leontein & Lönngren, 1993). GLC was performed on a Hewlett-Packard 5890 instrument equipped with a capillary HP-5ms column (25 m × 0.25 mm) using a temperature gradient of 160 °C (3 min) to 290 °C at 7 °C min−1.

Nuclear magnetic resonance (NMR) spectroscopy

Samples were freeze-dried twice from 99.9% D2O and dissolved in 99.95% D2O. 1H and 13C NMR spectra were recorded on a Bruker Avance II 600 spectrometer (Germany) at 30 °C (for mono- and oligo-saccharides) or 65 °C (for math formula) using internal sodium 3-(trimethylsilyl)propanoate-2,2,3,3-d4H 0, δC −1.4) as reference. Two-dimensional NMR spectra were obtained using standard pulse sequences, and the Bruker TopSpin 2.1 program was employed to acquire and process the NMR data. A mixing time of 60 ms and a spin-lock time of 200 ms were used in TOCSY and ROESY experiments, respectively. Other NMR experimental parameters were set essentially as described (Hanniffy et al., 1999).

Construction of the DNaseI shotgun bank

Chromosomal DNA was prepared as described (Bastin & Reeves, 1995). Long-range PCR was performed with the Expand Long Template PCR system (Roche Applied Science, Basel, Switzerland) using the primers wl-10324 (5′-GCACTGGTAGCTATTGAGCCAGGGGCGGTAGCAT-3′) and wl-2211 (5′-ACTGCCATACCGACGACGCCGATCTGTTGCTTGG-3′), based on the sequences of the JUMPStart site and gnd gene, respectively, which flank the O-antigen gene cluster in Cturicensis. PCR was performed over 32 cycles of denaturation at 94 °C for 30 s, annealing at 61 °C for 45 s, and extension at 68 °C for 15 min. The PCR products were digested with DNase I, and the resulting DNA fragments were cloned into pGEM-T Easy (Promega) to produce a shotgun bank as previously described (Wang & Reeves, 1998).

Sequencing and analysis

Sequencing was carried out using an ABI 3730 automated DNA sequencer (Applied Biosystems, Foster City, CA). Sequencing data were assembled using the staden Package software (Staden, 1996) and annotated with the Artemis program (Rutherford et al., 2000). blast analysis was used to search available databases such as GenBank (Benson et al., 2000). The Clusters of Orthologous Groups (Tatusov et al., 1997) and Pfam (Finn et al., 2006) protein motif databases were used to search for conserved protein domains. tmhmm 2.0 (http://www.cbs.dtu.dk/services/TMHMM) was used to identify potential transmembrane segments.

PCR specificity test

The primers wl-44241 (5-TTTCTTGTTATTGCCTGTGT-3) and wl-44242 (5-AACAAAATCAGCGAGACTAA-3) were designed based on the sequence of the wzy gene using primer premier 5.0 software program (Premier Biosoft International, Palo Alto, CA), and used to amplify the DNA templates of the strains listed in Table 1. PCR reactions were performed in a total volume of 25 μL containing 1× PCR buffer (50 mM KCl, 10 mM Tris-HCl, pH 8.3), 2.5 mM MgCl2, 400 μM dNTPs, 0.15 μM each primer, 2.5 U Taq DNA polymerase (TaKaRa Biotechnology Co. Ltd, Dalian, China) and 50–100 ng DNA template. Thermal PCR conditions were as follows: an initial denaturation step at 95 °C for 5 min, 30 cycles of 94 °C for 30 s, 50 °C for 30 s and 72 °C for 1 min, and a final extension at 72 °C for 5 min. The reaction mixture (5 μL) was subjected to agarose gel electrophoresis to examine the amplified products.

Table 1. Sources of the bacterial strains used in this study
BacteriaNo. of strains and sourceTotal no.
  1. a

    American Type Culture Collection (ATCC).

  2. b

    Czech Collection of Microorganisms (CCM), Czech Republic.

  3. c

    Environmental isolates from food by Tianjin Entry-Exit Inspection and Quarantine Bureau, China.

  4. d

    Environmental isolates from food by Liaoning Entry-Exit Inspection and Quarantine Bureau, China.

  5. e

    Chinese Academy of Inspection and Quarantine, Beijing, China.

  6. f

    Environmental isolates from food by Jilin Entry-Exit Inspection and Quarantine Bureau, China.

  7. g

    Environmental isolates from food by Neimenggu Entry-Exit Inspection and Quarantine Bureau, China.

  8. h

    Environmental isolates from food by Shenyang Entry-Exit Inspection and Quarantine Bureau, China.

  9. i

    Environmental isolates from food by Xinjiang Entry-Exit Inspection and Quarantine Bureau, China.

  10. j

    Environmental isolates from food by Guangdong Entry-Exit Inspection and Quarantine Bureau, China.

  11. k

    Environmental isolates from food by Hunan Entry-Exit Inspection and Quarantine Bureau, China.

  12. l

    Environmental isolates from food by Beijing Entry-Exit Inspection and Quarantine Bureau, China.

  13. m

    Environmental isolates from food by Hubei Entry-Exit Inspection and Quarantine Bureau, China.

  14. n

    China General Microbiological Culture Collection Center (CGMCC), China.

  15. o

    National Center for Medical Culture Collection (CMCC), China.

Cronobacter sakazakii 3a, 2b, 39c, 3d, 4e, 2f, 2g, 3h, 4i, 4j, 5k, 48l119
C. sakazakii O11a, 5c, 2d, 2e, 1f, 1g, 1h, 2i, 2k, 32l49
C. sakazakii O22a, 2b, 16c, 1e, 4j, 2k, 7l34
C. sakazakii O37c, 1k8
C. sakazakii O43c, 1f, 4l8
C. sakazakii O53c, 1i4
C. sakazakii O61c, 1i, 6l8
C. sakazakii O74c, 1d, 1e, 1g, 1h8
Cronobacter turicensis 2e, 3i, 1m6
Cronobacter malonaticus 2e2
Cronobacter muytjensii 1a1
Cronobacter dublinensis 1b, 1e, 3i, 2l7
Enterobacter cloacae 1a, 1n2
Enterobacter aerogenes 2n2
Escherichia coli 1a, 1o2
Salmonella enterica 1a, 1o2
Shigella 1a, 1l2
Total 145

Nucleotide sequence accession numbers

The DNA sequence of the O-antigen gene cluster of C. turicensis O2 has been deposited in GenBank under the accession number JQ354993.

Results and discussion

Structure of the O-polysaccharide

The LPS was obtained from dried cells of C. turicensis O2 by extraction with hot aqueous phenol. To release the polysaccharide chain, mild acid degradation of the LPS was performed but no polysaccharide was isolated, as it was cleaved at the acid-labile glycosidic linkage of a nonulosonic acid present in each O-unit (see below). As a result, an OS and a monosaccharide were obtained, which were separated by gel-permeation chromatography on Sephadex G-25.

The 1H and 13C NMR spectra of the monosaccharide were essentially identical to those of the authentic sample of 5,7-diacetamido-3,5,7,9-tetradeoxy-d-glycero-d-galacto-non-2-ulosonic acid (di-N-acetyllegionaminic acid, Leg5Ac7Ac), which had been synthesized earlier (Tsvetkov et al., 2001; Knirel et al., 2003). In particular, it could be distinguished from another naturally occurring isomer, di-N-acetyl-8-epilegionaminic acid (8eLeg5Ac7Ac), by the C-6 and C-8 chemical shifts (compare δC-6 70.7 and δC-8 67.7 for the major β-anomer of the isolated monosaccharide with the values δ 70.9 and 67.5 for β-Leg5Ac7Ac, δ 72.9 and 69.3 for β-8eLeg5Ac7Ac, respectively) (Knirel et al., 2003).

The NMR spectra showed that Leg5Ac7Ac is also present in the OS. Full acid hydrolysis of the OS followed by GLC of the acetylated alditols revealed rhamnose (Rha), galactose (Gal), and 2-amino-2-deoxyglucose (GlcN). No free Leg was detected in the hydrolysate, as monosaccharides of this class are known to be highly acid-labile and are destroyed under conditions of full acid hydrolysis. Determination of the absolute configurations by GLC of the acetylated (S)-2-octyl glycosides showed that Gal and GlcNAc have the d configuration and Rha has the l configuration. The absolute configuration of Leg5Ac7Ac was confirmed by analysis of 13C NMR chemical shifts of the OS taking into account the known regularities (Knirel et al., 2003) (see below).

The 13C NMR spectrum of the OS (Fig. 1) contained four signals for the anomeric carbons at δ 96.6–103.2, including one signal for a quaternary carbon (C-2 of Leg) at δ 97.0. The spectrum also showed signals for two CH3–C groups (C-6 of Rha and C-9 of Leg) at δ 18.0 and 20.4, one C–CH2–C group (C-3 of Leg) at δ 40.6, two HOCH2–C groups (C-6 of Gal and GlcN) at δ 61.9 and 62.0, three nitrogen-bearing carbons (C-2 of GlcN, C-5 and C-7 of Leg) at δ 52.0–57.1, one HO2C–C group (C-1 of Leg) at δ 174.3 and three N-acetyl groups (Me at δ 23.0–23.6 and CO at δ 175.0–175.7). The 1H NMR spectrum of the OS contained inter alia signals for three anomeric protons at δ 4.55 (J1,2 = 8.2 Hz), 4.86 (J1,2 < 2 Hz) and 5.12 (J1,2 = 3.8 Hz), two CH3–CH groups (H-6 of Rha and H-9 of Leg) at δ 1.13 and 1.33 (d), one CH–CH2–C group at δ 2.46 (dd, Leg H-3eq) and 1.90 (t, Leg H-3ax) and three N-acetyl groups at δ 1.95–2.01.

Figure 1.

13C NMR spectrum of the OS from LPS of Cronobacter turicensis O2. Numbers refer to the carbons in the sugar residues denoted as follows: G, Gal; GN, GlcNAc; L, Leg5Ac7Ac; R, Rha.

These data indicated that the OS is a tetrasaccharide containing one residue each of d-Gal, d-GlcNAc, l-Rha and Leg5Ac7Ac.

The NMR spectra of the OS were assigned using two-dimensional 1H, 1H COSY, TOCSY, ROESY and 1H, 13C HSQC experiments, and spin-systems for α-Galp, β-GlcpNAc, β-Rhap and β-Legp5Ac7Ac were identified (Table 2). Low-field positions of the signals for C-3 of Rha and C-4 of GlcNAc and Leg5Ac7Ac at δ 77.7–78.6, as compared with their positions in the corresponding non-substituted monosaccharides (Lipkind et al., 1988; Knirel et al., 2003), revealed the substitution pattern in the OS. The C-2–C-5 chemical shifts of Gal, which occupies the non-reducing end of the OS, were close to those of α-galactopyranose (Lipkind et al., 1988). The ROESY spectrum of the OS (Fig. 2) showed the following interresidue correlations for anomeric protons: Gal H-1/Rha H-2 and H-3 at δ 5.12/4.27 and 3.69, Rha H-1/GlcNAc H-4 at δ 4.86/3.64 and GlcNAc H-1/Leg5Ac7Ac H-4 at δ 4.57/4.00. These data define the monosaccharide sequence, Leg5Ac7Ac being at the reducing end of the OS. Therefore, the OS has the following structure:

display math
Figure 2.

Part of a two-dimensional ROESY spectrum of the OS from LPS of Cronobacter turicensis O2. The corresponding parts of the 1H NMR spectrum are shown along the axes. Numbers refer to the protons in the sugar residues denoted as follows: G, Gal; GN, GlcNAc; L, Leg5Ac7Ac; R, Rha.

Table 2. 1H and 13C NMR chemical shifts of the OS and math formula from Cronobacter turicensis O2 (δ, p.p.m.)
Sugar residueC-1 H-1C-2 H-2C-3 H-3 (3eq, 3ax)C-4 H-4C-5 H-5C-6H-6 (6a, 6b)C-7H-7C-8H-8C-9
  1. Signals for the N-acetyl groups are at δH 1.95–2.01; δC 23.0–23.6 (CH3), 175.0–175.7 (CO).

OS
α-d-Galp-(1→

96.6

5.12

69.5

3.85

70.6

3.95

70.4

4.01

72.0

4.18

62.0

3.72, 3.70

   
→3)-β-l-Rhap-(1→

101.8

4.86

68.4

4.27

78.6

3.69

71.6

3.46

73.4

3.43

18.0

1.33

   
→4)-β-d-GlcpNAc-(1→

103.2

4.57

57.1

3.62

74.8

3.67

78.2

3.64

75.7

3.49

61.9

3.93, 3.81

   
→4)-β-Legp5Ac7Ac174.397.0

40.6

2.46, 1.90

77.7

4.00

52.0

3.81

71.1

4.27

54.3

3.84

67.5

3.84

20.4

1.13

math formula
→4)-α-Legp5Ac7Ac-(2→174.3101.0

41.3

3.01, 1.80

79.0

3.66

51.4

3.80

72.9

3.88

55.0

3.86

68.7

3.99

19.5

1.18

→3)-α-d-Galp-(1→

96.5

5.18

67.8

3.93

73.8

4.46

69.4

4.03

72.2

4.16

62.2

3.72, 3.70

   
→3)-β-l-Rhap-(1→

101.7

4.87

68.3

4.27

78.5

3.69

71.7

3.50

73.4

3.44

18.1

1.35

   
→4)-β-d-GlcpNAc-(1→

102.9

4.57

57.2

3.65

74.9

3.69

78.5

3.71

75.7

3.52

62.2

3.95, 3.87

   

As in the free form, in the OS Leg5Ac7Ac exists mainly in the β-pyranose form with the equatorial carboxyl group. A relatively large α-effect on C-4 (8.8 p.p.m.), a small β-effect on C-3 (−0.3 p.p.m.) and a negative β-effect on C-5 (−2.1 p.p.m.) of Leg5Ac7Ac caused by its 4-glycosylation with β-d-GlcpNAc proved that Leg5Ac7Ac has the d-glycero-d-galacto configuration rather than the l-glycero-l-galacto configuration (Shashkov et al., 1988; Knirel et al., 2003).

Most likely, the OS represents a chemical repeating unit of the OPS, which was released from the LPS by selective cleavage of the glycosidic linkage of Leg5Ac7Ac. To confirm this and to establish the mode of the linkage between the O-units, the LPS was O-deacylated by treatment with aqueous ammonia and the product (math formula) was isolated by gel-permeation chromatography on Sephadex G-50.

The one- and two-dimensional 1H and 13C NMR spectra showed that the math formula has the same monosaccharide composition as the OS. The 1H, 13C HSQC spectrum of the math formula (Fig. 3) revealed differences in the anomeric configuration of Leg5Ac7Ac and the position of substitution of Gal. The difference of 1.21 p.p.m. between H-3eq and H-3ax chemical shifts and the chemical shift of δ 72.9 for C-6 were characteristic for the α configuration of Leg5Ac7Ac (Knirel et al., 2003). A significant down-field displacement of the signals for both H-3 and C-3 of α-Galp in the math formula to δH 4.46 and δC 73.8, as compared with the values δH 3.95 and δC 70.6 in the OS, respectively, demonstrated substitution of the Gal residue at position 3. This conclusion was confirmed by Smith degradation of the math formula, which did not destroy any monosaccharide, and upon mild acid hydrolysis the Smith degradation product afforded the same OS as the initial LPS.

Figure 3.

Part of a two-dimensional 1H, 13C HSQC spectrum of the math formula from Cronobacter turicensis O2. The corresponding parts of the 1H and 13C NMR spectra are shown along the horizontal and vertical axes, respectively. Numbers refer to the H/C pairs of the sugar residues denoted as follows: G, Gal; GN, GlcNAc; L, Leg5Ac7Ac; R, Rha.

Based on the data obtained, it was concluded that the OPS of C. turicensis O2 has the following structure:

display math

A partial release of Leg5Ac7Ac as the monosaccharide upon mild acid hydrolysis of the LPS suggests that it occupies the non-reducing end of the OPS. Hence, the structure shown above represents the biological O-unit of the OPS, which is assembled and then polymerized in a O-antigen polymerase (Wzy)-dependent manner (Valvano, 2011).

To our knowledge, the structure established is unique among the known bacterial polysaccharide structures. It includes Leg5Ac7Ac, which is an uncommon component of bacterial polysaccharide and is found in no non-bacterial natural carbohydrates. Earlier, this and other N-acyl derivatives of legionaminic acid have been identified as components of bacterial homo- and hetero-polysaccharides (Knirel et al., 2003, 2011), including OPS of C. turicensis HPB 3287 (MacLean et al., 2011), core OS of LPS, and O-glycan chains of bacterial flagellin (Knirel et al., 2011).

Characterization of the O-antigen gene cluster

A sequence of 16 346 bp from the JUMPStart site to the gnd gene was obtained and 16 open reading frames (ORFs), excluding gnd, were identified (Fig. 4). Functions were tentatively assigned to all of the ORFs, based on their similarity to related genes in the available databases and taking into account the C. turicensis O2 O-antigen structure (Table 3).The O-antigen gene cluster of C. turicensis O2 is quite different from that of C. turicensis O1 in terms of length and gene organization (sequences of 12 462 bp and 12 ORFs were obtained in C. turicensis O1). Genes of C. turicensis O1 and O2 are highly divergent, with the DNA identity levels ranging from 8.5% to 37.9% and protein identity levels from 1.3% to 20.2%, indicating they are two distinct serotypes.

Figure 4.

Organization of the O-antigen gene cluster of Cronobacter turicensis O2. Glycosyltransferase genes and O-antigen processing genes are shown in black and gray, respectively.

Table 3. Characteristics of the ORFs in the O-antigen gene cluster of Cronobacter turicensis O2
ORF no.Gene namePosition of gene (aa length)G + C content (%)Conserved domain(s) (Pfam no./E value)Similar protein(s), strain(s) (GenBank accession no.)% Identity/% similarity (no. of aa overlap)Putative function of protein
1 rmlB 167:1249 (360)50.50NAD dependent epimerase/dehydratase family (PF01370/9e−81)dTDP-d-glucose-4,6-dehydratase, Cronobacter sakazakii (ABX51885.1)97/99 (352)dTDP-d-glucose 4,6-dehydratase
2 rmlD 1252:2148 (298)49.94RmlD substrate binding domain (PF04321/3.6e−110)dTDP-6-deoxy-l-mannose dehydrogenase, Cronobacter sakazakii (ABX51886.1)97/99 (293)dTDP-6-deoxy-l-mannose dehydrogenase
3 rmlA 2099:3076 (325)46.21Nucleotidyl transferase (PF00483/1.9e−73)Glucose 1-phosphate thymidylyltransferase, Cronobacter sakazakii (ABX51887.1)96/98 (282)Glucose 1-phosphate thymidylyltransferase
4 rmlC 3080:3637 (185)40.32dTDP-4-dehydrorhamnose 3,5-epimerase (PF00908/1.5e−78)dTDP-4-dehydrorhamnose 3,5-epimerase, Escherichia coli (AAZ85717.1)76/85 (134)dTDP-4-dehydrorhamnose 3,5-epimerase
5 elg1 3630:4820 (396)37.36Polysaccharide biosynthesis protein (PF02719/2.8e−42)UDP-N-acetylglucosamine 4,6-dehydratase, Vibrio fischeri ES114 (YP_203523.1)75/83 (298)UDP-N-acetylglucosamine 4,6-dehydratase
6 elg2 4831:5979 (382)41.34DegT/DnrJ/EryC1/StrS aminotransferase family (PF01041/1.1e−80)UDP-bacillosamine synthetase, Vibrio parahaemolyticus 16 (ZP_05118541.1)72/87 (274)Aminotransferase
7 elg3 5992:7146 (384)43.20UDP-N-acetylglucosamine 2-epimerase (PF02350/7e−93)Putative NeuC protein, Vibrio parahaemolyticus RIMD 2210633 (NP_796578.1)69/83 (209)UDP-N-acetylglucosamine 2-epimerase
8 elg4 7151:8227 (358)43.36NeuB family (PF03102/1.3e−97)Putative N-acetylneuraminic acid synthetase, Vibrio parahaemolyticus RIMD 2210633 (NP_796579.1)73/83 (258)N-Acetylneuraminic acid synthetase
9 elg5 8196:8831 (211)42.13Bacterial transferase hexapeptide (three repeats) (PF00132/0.014)Acetyltransferase, Aliivibrio salmonicida LFI1238 (YP_002261713.1)47/64 (95)Acetyltransferase
10 elg6 8875:9930 (351)38.35Nucleotidyl transferase (PF00483/2.4e−33)Mannose 1-phosphate guanyltransferase, Vibrio fischeri ES114 (YP_203528.1)64/79 (226)Guanyltransferase
11 elg7 9930:10622 (230)37.51Cytidylyltransferase (PF02348/2.2e−38)CMP-N-acetylneuraminic acid synthetase, Vibrio splendidus 12B01 (ZP_00989910.1)50/64 (114)CMP-N-acetylneuraminic acid synthetase
12 wzx 10635:11876 (413)35.26Polysaccharide biosynthesis protein (PF01943/1.7e−20)Wzx, Proteus mirabilis (ADL32319.1)42/65 (173)Flippase
13 wepA 11888:12868 (326)32.92Glycosyltransferase family 52 (PF07922/4.5e−30)Glycosyltransferase WbwA, Escherichia coli (AAK64371.1)33/52 (105)Glycosyltransferase
14 wepB 12861:14021 (387)35.16Glycosyl transferases group 1 (PF00534/9.5e−22)Glycosyltransferase, Algoriphagus sp. PR1 (ZP_01718721.1)28/44 (90)Glycosyltransferase
15 wzy 14014:15099 (361)30.93Protein of unknown function (DUF571) (PF04600/0.0021)Hypothetical protein HMPREF0104_02464, Bacteroides sp. 3_1_19 (ZP_06986237.1)29/49 (68)Polymerase
16 wepC 15096:16151 (351)37.21Glycosyltransferases group 1 (PF00534/8.5e−08)Glycosyltransferases WcvF, Vibrio vulnificus (ADO64244.1)59/77 (208)Glycosyltransferase

Sugar biosynthetic pathway genes

Biosynthesis of the cytidine monophosphate (CMP)-activated derivative of Leg5Ac7Ac in Legionella pneumophila and Campylobacter jejuni starts from either UDP-α-d-GlcNAc or GDP-α-d-GlcNAc (Glaze et al., 2008; Schoenhofen et al., 2009). A similar biosynthetic pathway has been proposed for Leg5Ac7Ala in Escherichia coli O161 (Li et al., 2010) and analysis of the O-antigen gene cluster suggested that an essentially similar pathway is used by C. turicensis O2. Indeed, seven ORFs, orfs 5–11, were tentatively identified as genes for the synthesis of CMP-Leg5Ac7Ac.

Orf5 shares 75% identity with the UDP-N-acetylglucosamine 4,6-dehydratase of Vibrio fischeri; hence, a 4,6-dehydratase function, which converts a nucleoside diphosphate (NDP) derivative of α-d-GlcNAc into NDP-2-acetamido-2,6-dideoxy-α-d-xylo-hexos-4-ulose, was assigned to Orf5. Orf6 shares 68% identity with Lea2 of E. coli O161 and was therefore proposed to be an aminotransferase, whose product is NDP-2-acetamido-4-amino-2,4,6-trideoxy-α-d-glucopyranose. Orf9 is 47% identical to the acetyltransferase of Aliivibrio salmonicida LFI1238 and was proposed to catalyze N-acylation of the Orf6 product into NDP-2,4-acetamido-2,4,6-trideoxy-α-d-glucopyranose. Orf7 shares 69% identity with the NeuC protein of Vibrio parahaemolyticus RIMD 2210633. Searching against the Pfam databases showed that Orf7 is related to UDP-N-acetylglucosamine 2-epimerase (PF02350; 7e−93). Therefore, Orf7 was suggested to catalyze the conversion of the Orf9 product to its C-2 epimer, NDP-2,4-acetamido-2,4,6-trideoxy-α-d-mannopyranose. Orf10 shares 64% identity with mannose 1-phosphate guanyltransferase of V. fischeri ES114. Searching against the Pfam databases showed that Orf10 is related to nucleotidyl transferase (PF00483; 2.4e−33). Therefore, Orf10 was proposed to be a nucleotidase responsible for the removal of the NDP group from the Orf7 product. Orf8 shares 50% identity with Lea4 of E. coli O161 and was proposed to catalyze the condensation of the Orf7 product 2,4-acetamido-2,4,6-trideoxy-d-mannose with phosphoenolpyruvate to give Leg5Ac7Ac. Orf11 shares 50% identity with the CMP-N-acetylneuraminic acid synthetase of Vibrio splendidus 12B01. Hence, Orf11 was suggested to catalyze the synthesis of CMP-Leg5Ac7Ac from Leg5Ac7c. The biosynthesis pathway of CMP- Leg5Ac7Ac is hypothesized, but not confirmed, numbers rather than letters for the 4th character in the names were finally adopted. Therefore, orfs 5–11 were named elg1, elg2, elg3, elg4, elg5, elg6, and elg7, respectively.

Orfs 1–4 shared high-level identities with many known RmlB, RmlD, RmlA, and RmlC proteins (76–97%). The rmlBDAC gene set responsible for the synthesis of dTDP-l-Rha has been well characterized (Giraud et al., 1999) and orfs 1–4 were therefore identified as rmlB, rmlD, rmlA, and rmlC, respectively.

No genes responsible for Gal and GlcNAc biosynthesis were identified in the gene cluster; therefore, it is assumed that these genes are located outside of the O-antigen gene cluster, as in E. coli (Samuel & Reeves, 2003).

Sugar transferase genes

Glycosyltransferases are specific for sugar donors and sugar acceptors and the linkage that they form. O-unit synthesis in many bacteria is initiated by transfer of N-acetylglucosamine 1-phosphate to an undecaprenyl phosphate carrier, in a reaction catalyzed by WecA (Valvano, 2011). It was proved that the wecA gene is located outside of the O-antigen gene cluster in Cronobacter by searching the genome of C. sakazakii BAA 894 (Locus_tag: ESA_03774). As four sugars are present in the O-unit, three glycosyltransferase genes were expected in the O-antigen gene cluster of C. turicensis O2.

Orf13 is 33% identical to glycosyltransferase WcvF of Vibrio vulnificus and belongs to the glycosyltransferase family 52 (PF07922; 9.5e−22). Orf14 is 28% identical to the glycosyltransferase of Algoriphagus sp. PR1 and belongs to the glycosyltransferase group 1 family (PF00534; 9.5e−22). Orf16 is 59% identical to the glycosyltransferase of V. vulnificus and also belongs to the glycosyltransferase group 1 family (PF00534; 8.5e−08). Based on these findings, orf13, orf14 and orf16 were assigned the functions of glycosyltransferase genes and named wepA, wepB, and wepC, respectively.

O-antigen processing genes

Wzx and Wzy are typical inner membrane proteins, which usually share little sequence identity with their homologues. Orf12 and Orf15 were the only proteins with predicted transmembrane segments. Orf12 has 12 predicted transmembrane segments, which is a typical topological characteristic of Wzx proteins (Liu et al., 1996). Orf12 also shares 65% similarity with the putative Wzx protein of Proteus mirabilis. Orf15 has nine predicted transmembrane segments and a large periplasmic loop of 48 amino acid residues, which is a typical topological characteristic of Wzy proteins (Daniels et al., 1998). Therefore, orf12 and orf15 were proposed to be the O-unit flippase gene (wzx) and the O-antigen polymerase gene (wzy), respectively, and were named accordingly.

Primer specificity

A specific primer pair for C. turicensis was designed on the sequence of the wzy gene, and was used to screen the DNA of 145 strains including 119 C. sakazakii, two C. malonaticus, one C. muytjensii, seven C. dublinensis and six C. turicensis strains, and 10 stains of closely related species (Table 1). Only the DNA of C. turicensis O2 (g3882) generated the expected PCR products, and no PCR products were detected in any other isolate (Supporting Information, Fig. S1). Therefore, the targeted wzy gene was proven to be specific for C. turicensis O2. The high specificity of the designed primer pair has the potential to be used for the identification and detection of C. turicensis O2.

Acknowledgements

This work was supported by the Federal Targeted Program for Research and Development in Priority Areas of Russia's Science and Technology Complex for 2007–2013 (State contract No. 16.552.11.7050), the Russian Foundation for Basic Research (Project No. 10-04-00598), the Chinese National Science Fund for Distinguished Young Scholars (No. 30788001), the National Natural Science Foundation of China (Nos. 31000044, 30670038, 30771175 and 30900041), the National High Technology Research and Development Program of China (863 Program) (No. 2009AA06Z403), the Fundamental Research Funds for the Central Universities (No. 65010721), and the National Key Program for Infectious Diseases of China (No. 2009ZX10004-108). We thank Xinghang Jiang (China Agricultural University) for technical assistance with the manuscript.

Authors' contribution

Y.S. and N.P.A. contributed equally to this work.

Ancillary