Characterization of the Bacteroides fragilis pathogenicity island in human blood culture isolates

ARTICLE IN PRESS

Anaerobe 12 (2006) 17–22 www.elsevier.com/locate/anaerobe

Clinical medicine$

Characterization of the Bacteroides fragilis pathogenicity island in human blood culture isolates M.C. Clarosa,, Z.C. Clarosa, D.W. Hechtb, D.M. Citronc, E.J.C. Goldsteinc, J. Silva Jrd, Y. Tang-Feldman, A.C. Rodloffa a

Institute of Medical Microbiology and Infectious Epidemiology, University of Leipzig, Germany b Loyola University, Chicago, USA c R.M. Alden Research Laboratory, Santa Monica, California, USA d Department of Internal Medicine, UC Davis, USA Received 17 November 2004; received in revised form 15 June 2005; accepted 15 June 2005 Available online 15 August 2005

Abstract Bacteroides fragilis is an important anaerobic pathogen accounting for up to 10% of bacteremias in adult patients. Enterotoxin producing B. fragilis (ETBF) strains have been identified as enteric pathogens of children and adults. In order to further characterize the B. fragilis pathogenicity island (BfPAI) and using PCR assays for bft- and mpII-metalloprotease genes, we determined the frequency of B. fragilis strains with pattern I (containing the BfPAI and its flanking region), pattern II (lacking both the BfPAI and the flanking region), and pattern III (lacking the BfPAI but containing the flanking region) in 63 blood culture isolates. The results were compared to 197 B. fragilis isolates from different clinical sources. We found 19% of blood culture isolates were pattern I (ETBF), 43% were pattern II (NTBF) and 38% were pattern III (NTBF). Comparatively, B. fragilis isolates from other clinical sources were 10% pattern I, 47% pattern II and 43% pattern III. This suggests that the pathogenicity island and the flanking elements may be general virulence factors of B. fragilis. r 2005 Elsevier Ltd. All rights reserved. Keywords: Enterotoxigenic Bacteroides fragilis; Bacteroides fragilis pathogenicity island; Sepsis

1. Introduction Bacteroides fragilis constitutes 1–2% of the normal gut flora [1] and is commonly associated with bacteremia, soft tissue infections, intra-abdominal infections and abscesses [2–5]. The species B. fragilis exhibits virulence factors that are unique among anaerobic bacteria, such as its complex carbohydrate capsule, which has been shown to cause abscess formation [6], as well as changes in the Corresponding author. Tel.: +41 61 486 1417; fax: +41 486 1549.

E-mail address: [email protected] (M.C. Claros). Paper from Anaerobe 2004. The 7th Biennial Congress of the Anaerobe Society of the Americas, Annapolis, Maryland, USA, 19–21 July 2004. $

1075-9964/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.anaerobe.2005.06.005

outer membrane protein in strains that show resistance to carbapenems [7]. Other virulence factors contributing to B. fragilis pathogenicity, infections and therapeutic failures remain to be determined. The isolation of B. fragilis from blood cultures has been associated with significant mortality [4,8] but the causative relationship of B. fragilis sepsis and mortality remains unclear [9]. Kato and co-workers suggested that toxigenic B. fragilis (ETBF) strains are more virulent than toxin negative strains in patients with sepsis [10]. In order to further characterize the pathogenicity factors in ETBF B. fragilis, we used PCR assays to detect genetic sequences of the BfPAI, the enterotoxin gene (bft), and a second metalloprotease gene (mpII) in blood stream isolates and compared the results to control isolates from different clinical sources. In

ARTICLE IN PRESS M.C. Claros et al. / Anaerobe 12 (2006) 17–22

18

addition, sequences of the
12-kb BfPAI flanking regions were analyzed and compared in ETBF and NTBF strains [11]. In order to determine the genetic sequences flanking the BfPAI of ETBF and NTBF strains, sequence analysis were performed on six selected strains, including BFT positive and negative control strains.

2. Materials and methods

2.2. PCR fingerprinting The PCR fingerprint assay was performed as described previously [13]. Briefly, a t-RNA gene primer was used as a single primer in the PCR assays in order to obtain fingerprints of all strains. The samples were amplified for 35 cycles as follows: 1 min at 94 1C, 1 min at 50 1C, and 2 min at 72 1C [13]. Patterns were compared visually and also by scanning. B. fragilis PCR groups I and II have been already described [13].

2.1. Bacterial strains, cultivation and identification

2.3. Specific PCR

Thirty-two clinical blood culture isolates pre-identified as B. fragilis were obtained from the R.M. Alden Research Laboratory (RMA), as well as 30 from the Loyola University. One strain, 3431, was isolated at the Institute of Microbiology, University of Leipzig. Control group: 197 strains, 112 of these consecutively collected at the University of Leipzig (1999–2001) and 85 non-consecutively collected (Germany and US) from human infection non-blood culture sources were used as a control group. Reference strains: B. fragilis ATCC 25285 (BFT negative strain), ATCC 43858, and ATCC 43859 (BFT positive strains) obtained from the American Type Culture Collection and the DNA homology group II control strain VPI 2393 obtained from the Virginia Polytechnical Institute (VPI) [12] were used as reference strains. Strains were grown on Columbia blood agar, supplemented with blood (10% sheep blood) (Oxoid, Basingstoke, UK). All strains were identified using the RapID Ana II System (Innovative Systems, CA) and an additional indole-test (Anaerobe Systems, CA).

The optimization of the detection of the pathogenicity genes was performed using modified Taguchi methods [14]. Primers were chosen from publications (Table 1). The PCR reactions for the detection of the bft- and mpII-genes [10,15,16] were performed using the following cycler profiles: 2 min 95 1C, one cycle, 2 min 95 1C, 1 min 65 1C, 1 min 72 1C, 45 cycles, and 2.5 min 72 1C, 1 cycle. The amplification reaction of the BfPAI flanking regions was as follows: 5 min 94 1C, one cycle, 1 min 94 1C, 2 min 66 1C, 1 min 72 1C, 30 cycles and 5 min 72 1C, one cycle [11]. 2.4. Direct sequencing of PCR products Direct sequencing of PCR products was initiated using appropriate primers (Table 1) in a dye-terminator cycle-sequencing reaction (Perkin Elmer ABI PRISM, Applied Biosystems, Warrington, United Kingdom) and the sequence analysis carried out on an ABI 373 Stretch automatic sequencer (Applied Biosystems, Warrington, United Kingdom) (performed at the University of

Table 1 Primers for the specific PCR and sequencing Primer

Fragment name

Fragment length

T3Ba RS-3b RS-4 GBF 101c GBF 110 LO1d RO1 P1T7e P1T7-1 P1T3 P1T3-1 P1T7P1T3

T3B bft

367 bp

bft

358 bp

mpII

350 bp

T7-T71

1.3kb

T31-T3

1.1kb

P1T7-P1T3

1.6kb

a

Claros et al. [13]. Shetab et al. [15]. c Kato et al. [10]. d Moncrief et al. [16]. e Franco et al. [11]. b

Primer sequence (50 - 30 ) AGGTCGCGGGTTCGAATCC TGAAGTTAGTGCCCAGATGAGG GCTCAGCGCCCAGTATATGACC AGCCGAAGACGGTGTATGT CCCACTGGCTTCAAAATCCGAAGC CCACCGTGCCAATGTCAGATA CTGAAGAACGAGGCGGTATC GCTGGTAGACTACCTGAGTAAGGAGTC GCTTCCGTACCCAGGTATCTCTCCATA TTCAACCTGATCGATCCGGAAGATCCG GGTAGTGCTTATGTCCCTGCAACCCTA GCTGGTAGACTACCTGAGTAAGGAGTC TTCAACCTGATCGATCCGGAAGATCCG

ARTICLE IN PRESS M.C. Claros et al. / Anaerobe 12 (2006) 17–22

Leipzig, IZKF core unit laboratory). All the sequences were subject to a database search using BLASTs (Basic Local Alignment Search Tool) program. BLASTN was used for the comparison of each nucleotide query sequence against the nucleotide sequence database [17]. Sequence alignments ClustalW was used as the program for multiple alignment. It provides an integrated environment for performing multiple sequence and profile alignments and analyzing the results. This program is utilized from within either BioEdit or UK HGMP [18].

3. Results 3.1. Bloodstream isolates ETBF. Of the 63 blood culture isolates 12 (19%) strains were determined to be ETBF. Amplification of bft and mpII in ETBF strains. In these 12 strains, the bft gene fragments (358 bp [10]; 367 bp [15]) were amplified. All the strains also carried the mpII-gene (350 bp sequence [16]). Amplification of BfPAI flanking regions. All ETBF strains (12) carried the 1.3 and 1.1 kb flanking sequences of the BfPAI, specific for pattern I (Pattern I-ETBF: containing the BfPAI and its flanking region, pattern IINTBF lacking the BfPAI as well as the flanking region, pattern III-NTBF containing the flanking region but lacking the BfPAI) as devised by Franco and co-workers [11]. Of the 51 NTBF strains, 24 (38%) did not carry the BfPAI but the 1.6 kb flanking region of the BfPAI. The type strain B. fragilis ATCC 25285 grouped among these. All these strains could be grouped into pattern III (Fig. 1). Patterns IV and V were not distinguished from pattern III by PCR; therefore pattern III contains strains which might belong to pattern IV/V by probe analysis. Twenty-seven strains contained neither the pathogenicity island nor the BfPAI flanking regions and belonged to pattern II [11].

Fig. 1. Amplification of a 1.6 kb long DNA fragment which can only be found in certain NTBF strains. Lanes 1–2, 5–9 and 13–17 ETBF strains LOY 6253, 6332, 6408, 6974, 7869, RMA 5081, 5447, 5835, 6470, 6791, 8769, 10085; Lanes 3–4 and 10 clinical NTBF strains Loy 6293, 6299, 6404; Lane 11 NTBF ATCC 25285 (positive for the 1.6 kb fragment); Lane 12 ETBF ATCC 43858 (negative for the 1.6 kb fragment), Lane 18 negative control without DNA, Lane 19 123 bp ladder.

19

PCR fingerprinting of B. fragilis isolates. Group I: The 12 ETBF isolates belonged to the PCR fingerprint group I that resembles the B. fragilis type strain ATCC 25285. All ETBF isolates demonstrated similar PCR fingerprint profiles as the non-ETBF isolates in group I [13]. Group II: Seven out of 63 strains (11%) belonged to the PCR group II. Off these, two strains carried the 1.6 kb flanking region of the BfPAI. This element carries the integration site for the BfPAI [11]. Sequencing of flanking regions. The flanking elements of six strains with specific characteristics, including three ETBF (pattern I) and one NTBF (pattern III) isolates as well as the two reference strains ATCC 25285 and ATCC 43859, were chosen to be sequenced in order to detect possible similarities and differences among them (Table 2). Those sequences were aligned to several sequences of the pathogenicity island [19]. There were certain point mutations obtained in fragment T7–T71 within the NTBF pattern III strains 3431 and ATCC 25285 (Fig. 2). Also, within the flanking element T71–T7 of the ETBF strain 6974 four nucleotide exchanges were determined (Fig. 3). In addition, a point mutation in fragment T3–T31 of NTBF pattern III strains 3431 and ATCC 25285 compared to ETBF strains as well as a greater exchange among the two clinical ETBF isolates 10085 and 6974 were demonstrated (Fig. 4). 3.2. Control group The control group consisted of 197 B. fragilis isolates from skin- and soft tissue infections. Of this group, 20 out of 197 strains (10%) were ETBF (pattern I). Of the NTBF, 92 out of 197 (47%) strains belonged to the NTBF (pattern II), and 85 strains (43%) belonged to the NTBF (pattern III), carrying the 1.6 kb fragment. Of these NTBF III strains, two belonged to the PCR group II.

4. Discussion B. fragilis was recognized as an emerging etiologic agent of diarrhea in animals and humans in 1984 [20]. ETBF strains of B. fragilis were shown to produce a 20 kDa zinc-dependent metallo-protease-toxin (termed B. fragilis toxin or BFT) that mediates the cleavage of Ecadherin, resulting in an altered morphology of certain human intestinal carcinoma cell lines, fluid accumulation in ligated lamb ileal loops, and intestinal epithelial cell proliferation [20]. Three isotypes of the BFT are encoded by distinct bft loci contained within a 6-kb chromosomal region unique to ETBF strains termed the B. fragilis pathogenicity island (BfPAI) [11]. ETBF was also found to be the cause of diarrheal diseases in young children (41 year of age) and adult patients [15,21–29].

ARTICLE IN PRESS M.C. Claros et al. / Anaerobe 12 (2006) 17–22

20 Table 2 Data on sequenced strains Accession Numbers

Strain number

DNA-Fragment

Origin

AJ AJ AJ AJ

850062 850066 850072 850076

10085

T31-T3 T3-T31 T71-T7 T7-T71

R.M. Alden Research Laba

AJ AJ AJ AJ

850065 850071 850075 850080

7869

T31-T3 T3-T31 T71-T7 T7-T71

Loyola Universityb

AJ850064 AJ 850070 AJ 850074 AJ 850081

6974

T31-T3 T3-T31 T71-T7 T7-T71

Loyola Universityb

AJ 850068 AJ 850078

3431

T3-T31 T7-T71

Leipzig Universityc

AJ AJ AJ AJ

850063 850069 850073 850079

43859

T31-T3 T3-T31 T71-T7 T7-T71

ATCC

AJ 850067 AJ 850077

25285

T3-T31 T7-T71

ATCC

a Blood culture isolate from the MSD intra-abdominal infection study, 1999. The isolate was carbapenem sensible (MIC o0.06). From the same specimen there was also grown Bacteroides thetaiotaomicron. b No clinical information. c Male patient, Department of Surgery, obstructive arterial disease, post-amputation sepsis, pneumonia.

Fig. 2. Sequence analysis of the T7–T71 flanking element demonstrating point mutations in positions (33 and 86) of the NTBF strains (B. fragilis type strain ATCC 25285 NTBF pattern III and clinical isolate 3431) compared to the ETBF strains (control strain ATCC 43859 and clinical isolate 10085). The sequences were aligned to the sequence of Franco et al., 1999 [11].

Fig. 3. A part of sequence analysis of the clinical isolate 6974 (ETBF) using the pair of primers T71–T7, demonstrating differences in the nucleotide positions 86–100 and 117. The sequences were aligned to the sequence of Franco et al., 1999 [11].

Fig. 4. Demonstrating mutations (in fragments T3–T31) of NTBF pattern III strains compared to ETBF strains (position 296) and a greater exchange of nucleotides in strains 10085 and 6974 (positions 42–43 and 54–58). The sequences were aligned to the sequence of Franco et al. [11].

The prevalence of ETBF varies between 9% and 25% of all B. fragilis isolates (mostly skin- and soft tissue infections), according to author and geographical origin [15,29]. Our previous study [30] demonstrated that the incidence of ETBF in skin- and soft tissue infections was 9% in Leipzig (Germany) and 14% in strains isolated in the Los Angeles area (US). In this study, the percentage of ETBF isolates from blood culture isolates (19%) collected at two institutions within the US was higher than from other types of

ARTICLE IN PRESS M.C. Claros et al. / Anaerobe 12 (2006) 17–22

specimens (the frequency of isolation was not significantly different in blood culture isolates). In our control group 197 strains obtained from different clinical sites and geographical origins we only observed 10% ETBF. This correlates with the results obtained by Kato et al. in Japan who found that a greater percentage (28%) of blood isolates were ETBF compared to non-blood culture isolates (14%) [10]. In addition, three major patterns of B. fragilis strains were found based on the presence of BfPAI and its flanking region: Pattern I (ETBF), pattern II (NTBF) and III (NTBF) with different flanking regions [11]. Interestingly, the B. fragilis type strain (ATCC 25285), contained the 1.6 kb BfPAI flanking region and therefore, belonged to pattern III (NTBF). It was demonstrated that 57% of the blood stream isolates carried genetic elements integrated into the chromosome (either BfPAI or the flanking regions only). Franco [19] proposed that foreign elements were acquired by NTBF (pattern II) strains through horizontal gene transfer. Sequence analysis was performed to determine changes in the BfPAI flanking region on ETBF and NTBF strains. Substitutions were observed between ETBF and NTBF (pattern III) strains as well as among clinical ETBF isolates. The above sequences were aligned with published sequences [16,19] (nucleotide sequence accession numbers: AF038459, AY 372755). There were single point mutations in the flanking sequences of the ETBF compared to NTBF (pattern III) strains. Interestingly, an accumulation of point mutations in the ETBF strain LOY6974 was located, proposing a possible variability in the amino acid sequences. Two of the NTBF group III strains belonged to the corresponding PCR fingerprint group II. These strains showed intermediate or high resistance to betalactam antibiotics including carbapenems (unpublished data) and carried foreign sequences, demonstrating a high potential of pathogenicity in bloodstream isolates. Since ETBF strains are more prominent in bacteremia than in other types of infections, the enterotoxin produced by B. fragilis strains may have virulence features that go beyond its enterotoxigenic effects and that BfPAI and the flanking regions may be general virulence factors.

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

References [1] Finegold SM. Anaerobic bacteria in human disease. New York: Academic Press; 1977. [2] Brook I, Frazier EH. Aerobic and anaerobic microbiology in intra-abdominal infections associated with diverticulitis. J Med Microbiol 2000;49:827–30. [3] Goldstein EJC, Citron DM. Annual incidence, epidemiology, and comparative in vitro susceptibilities to cefoxitin, cefotetan, cefmetazole, and ceftizocime of recent community-acquired

[20]

[21]

[22]

21

isolates of the Bacteroides fragilis group. J Clin Microbiol 1988;26:2361–6. Nguyen MH, Yu VL, Morris AJ, McDermott L, Wagener MW, Harrell L, Snydman DR. Antimirobial resistance and clinical outcome of Bacteroides bacteremia: findings of a multicenter prospective observational trial. Clin Inf Dis 2000;30:870–6. Javaloyas M, Garcia-Somoza D, Gudiol F. Epidemiology and prognosis of bacteremia: a 10-year study in a community hospital. Scand J Infect Dis 2002;34:436–41. Kasper DL, Onderdonk AB, Crabb J, Bartlett JG. Protective efficacy of immunization with capsular antigen against experimental infection with Bacteroides fragilis in an intraabdominal abscess model. J Clin Invest 1979;69:9–16. Edwards R, Greenwood D. Distinctive outer membrane protein and lipopolysaccharide composition of Bacteroides fragilis strains that produce metallo-betalactamase. Anaerobe 1997;3:233–6. Saito T, Senda K, Takakura S, Fujihara N, Kudo T, Linuma Y, Fujita N, et al. Anaerobic bacteremia: the yield of positive anaerobic blood cultures: patient characteristics and potential risk factors. Clin Chem Lab Med 2003;41:293–7. Bryan CS, Renolds KL, Kirhart B, Brown JJ. Bacteroides bacteremia. Analysis of 142 episodes from one metropolitan area. Arch Surg 1984;119:894–8. Kato N, Kato H, Watanabe K, Ueno K. Association of enterotoxigenic Bacteroides fragilis with bacteremia. Clin Infect Dis 1996;23(S1):S832. Franco AA, Cheng RK, Chung GT, Wu S, Oh OB, Sears CL. Molecular evolution of the pathogenicity island of enterotoxigenic Bacteroides fragilis strains. J Bacteriol 1999;181:6623–33. Johnson JL. Taxonomy of the Bacteroides. I. Deoxyribonucleic acid homologies among Bacteroides fragilis and other saccharolytic Bacteroides species. Int J System Bacteriol 1978;28:245–56. Claros MC, Scho¨nian G, Gra¨ser Y, Montag Th, Rodloff AC, Citron DM, Goldstein EJC. Identification and strain differentiation of ‘Bacteroides fragilis group’ species and Prevotella bivia by PCR fingerprinting. Anaerobe 1995;1:209–17. Cobb BD, Clarkson JM. A simple procedure for optimizing the polymerase chain reaction (PCR) using modified Taguchi methods. Nucl Acid Res 1994;22:3801–5. Shetab R, Cohen SH, Prindiville T, Tang YJ, Cantrell M, Rahmani D, Silva Jr J. Detection of Bacteroides fragilis enterotoxin gene by PCR. J Clin Microbiol 1998;36:1729–32. Moncrief JS, Duncan AJ, Wringht RL, Barroso LA, Wilkins TD. Molecular characterization of the fragilysin pathogenicity islet of enterotoxigenic Bacteroides fragilis. Infect Immun 1998;66: 1735–9. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. Journal of Molecular Biology 1990;215:403–10. Thompson JD, Higgins DG, Gibson TJ. CLUSTALW improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl Acids Res 1994;22:4673–80. Franco AA. The Bacteroides fragilis pathogenicity island is contained in a putative novel conjugative transposon. J Bacteriol 2004;186:6077–92. Myers LL, Firehammer BD, Shoop DS, Border MM. Bacteroides fragilis: a possible cause of acute diarrheal diseases in newborn lambs. Infect Immun 1984;44:241–4. Pantosti A, Piersimoni C, Perissi G. Detection of Bacteroides fragilis enterotoxin in the feces of a child with diarrhea. Clin Infect Dis 1994;19:809–10. Sack RB, Albert MJ, Alam K, Neogi PKB, Akbar MS. Isolation of enterotoxigenic Bacteroides fragilis from Bangladeshi children with diarrhea: a case-control study. J Clin Microbiol 1994; 32:960–3.

ARTICLE IN PRESS 22

M.C. Claros et al. / Anaerobe 12 (2006) 17–22

[23] San Joaquin VH, Griffis JC, Lee C, Sears CL. Association of Bacteroides fragilis with childhood diarrhea. Scand J Infect Dis 1995;27:211–5. [24] Sack RB, Myers LL, Almeido-Hill J, Shoop DS, Bradbury WC, Reid R, Santosham M. Enterotoxigenic Bacteroides fragilis epidemiologic studies of its role as a human diarrheal pathogen. J Diarrheal Dis Res 1992;10:4–9. [25] Caceres M, Zhang G, Weitraub A, Nord CE. Prevalence and antimicrobial susceptibility of enterotoxigenic Bacteroides fragilis in children with diarrhea in Nicaragua. Anaerobe 2000;6: 143–8. [26] Kato N, Liu C, Kato H, Watanabe K, Nakamura H, Iwai N, Ueno K. Prevalence of enterotoxigenic Bacteroides fragilis in children with diarrhea in Japan. J Clin Microbiol 1999;37:801–3.

[27] Menozzi MG, Malpeli M, Covan Sl, Rossi S, Genaglia G, Montanarini G, Vitali P, Pantosti A, Gherardini G, Chezii C. Bacteroides fragilis diarrhea in children admitted to Parma University Hospital: a case-control study. Second world congress on anaerobic bacterial and infections (abstract), 1998 [28] Zhang G, Svenungsson B, Karnell A, Weintraub A. Prevalence of enterotoxigenic Bacteroides fragilis in adult patients with diarrhea and healthy controls. Clin Infect Dis 1999;29:590–4. [29] Szo¨ke I, Dosa E, Nagy E. Enterotoxigenic Bacteroides fragilis in Hungary. Anaerobe 1997;3:87–9. [30] Claros MC, Claros ZC, Tang YJ, Cohen SH, Silva Jr J, Goldstein EJC, Rodloff AC. Occurrence of Bacteroides fragilis enterotoxin gene-carrying strains in Germany and the United States. J Clin Microbiol 1999;38:1996–7.