OUP user menu

Rapid identification of the species of the Bacteroides fragilis group by multiplex PCR assays using group- and species-specific primers

Chengxu Liu, Yuli Song, Maureen McTeague, Ann W. Vu, Hannah Wexler, Sydney M. Finegold
DOI: http://dx.doi.org/10.1016/S0378-1097(03)00296-9 9-16 First published online: 1 May 2003


We report a rapid and reliable two-step multiplex polymerase chain reaction (PCR) assay to identify the 10 Bacteroides fragilis group species –Bacteroides caccae, B. distasonis, B. eggerthii, B. fragilis, B. merdae, B. ovatus, B. stercoris, B. thetaiotaomicron, B. uniformis and B. vulgatus. These 10 species were first divided into three subgroups by multiplex PCR-G, followed by three multiplex PCR assays with three species-specific primer mixtures for identification to the species level. The primers were designed from nucleotide sequences of the 16S rRNA, the 16S–23S rRNA intergenic spacer region and part of the 23S rRNA gene. The established two-step multiplex PCR identification scheme was applied to the identification of 155 clinical isolates of the B. fragilis group that were previously identified to the species level by phenotypic tests. The new scheme was more accurate than phenotypic identification, which was accurate only 84.5% of the time. The multiplex PCR scheme established in this study is a simple, rapid and reliable method for the identification of the B. fragilis group species. This will permit more accurate assessment of the role of various B. fragilis group members in infections and of the degree of antimicrobial resistance in each of the group members.

  • Multiplex PCR
  • Identification
  • Bacteroides fragilis group

1 Introduction

The Bacteroides fragilis group, anaerobic, bile-resistant, non-spore-forming, Gram-negative rods are part of the endogenous human bowel flora [1]. Species most frequently isolated from that flora are Bacteroides vulgatus, B. thetaiotaomicron, B. distasonis and, less frequently, B. eggerthii and B. fragilis [2]. The B. fragilis group is commonly associated with a variety of human infections, such as intra-abdominal abscesses, wound infections and bacteremia [3]. The B. fragilis group bacteremia contributes significantly to morbidity and mortality [4]. The distribution of individual species in clinical infections is as follows: B. fragilis accounts for 63% of all the group isolates, B. thetaiotaomicron for 14%, B. vulgatus and B. ovatus for 7% each, B. distasonis for 6% and B. uniformis for 2% [5]. The choice of antibiotics for therapy is limited because the species of the B. fragilis group are among the most resistant of all anaerobes to antimicrobial agents, and this resistance has increased recently [6,7]. Accordingly, there is need for rapid, accurate identification of clinical isolates to permit an early, effective management of infected patients.

Phenotypic methods have been used in clinical laboratories to differentiate Bacteroides spp. These are time-consuming, and it may be difficult to differentiate Bacteroides species with them [8,9]. Automated methods currently used are also unreliable for identifying isolates of the B. fragilis group [10,11]. Incorrect identification of strains of the B. fragilis group may result in inappropriate antibiotic therapy.

Recently, genotypic-based techniques are emerging as alternatives or complements to phenotypic methods. Hybridization assays using a DNA or an RNA probe [12,13], polymerase chain reaction (PCR) amplification using species-specific primers [1416], rRNA restriction fragment length polymorphism [17], restriction endonuclease analysis [18], ribotyping [19], arbitrary primer PCR [20] and intergenic spacer region (ISR) PCR [21] have been used to study the B. fragilis species. However, for other species in the B. fragilis group, such as B. thetaiotaomicron, B. ovatus, B. vulgatus and B. distasonis, that are also clinically important, there has been little work carried out on molecular diagnosis.

With multiplex PCR, more than one locus is simultaneously amplified in the same reaction. It has the potential to save considerable time and effort for the laboratory without compromising test utility. Since its introduction, multiplex PCR has been successfully applied in many bacterial identifications [2225]. In the present study, based on the sequence analysis of the 16S rRNA gene, the 16S–23S rRNA ISR and a variable region of the 23S rRNA gene of the 10 ATCC type strains representing the 10 B. fragilis group species, a two-step multiplex PCR identification scheme was established to rapidly and accurately identify the B. fragilis group species.

2 Materials and methods

2.1 Bacterial strains and culture conditions

Strains used in this study included 10 ATCC strains representing the 10 B. fragilis group species and 155 Bacteroides isolates previously recovered from clinical specimens and identified phenotypically (Tables 1 and 2). The clinical isolates were chosen to represent 10 commonly isolated Bacteroides species. In addition, four ATCC strains of Bacteroides species other than the B. fragilis group and 47 ATCC or NCTC strains of Gram-negative anaerobic bacilli other than Bacteroides species that are either phylogenetically related species or may grow on Bacteroides bile esculin agar were used to verify the specificity of the established multiplex PCR assay (Tables 1 and 2). All strains were cultured anaerobically overnight on Brucella blood agar (Anaerobe Systems, CA, USA) at 37°C and were characterized by a combination of conventional tests as described in the Wadsworth Anaerobe Manual [26] and the BD BBL Crystal™ Identification System (Becton Dickinson Microbiology Systems, MD, USA).

View this table:
Table 1

List of strains used in this study and the multiplex PCR results

StrainStrain No.Accession No.Bac-FG23S-1G23S-11392A1392A1392A1392A1392ABegg-FBuni-FBste-FBdis-FBmer-F
B. thetaiotaomicronATCC29148TAY155588++
B. vulgatusATCC8482TAY155596++
B. fragilisATCC25285TAY155459++
B. caccaeATCC43185TAY155590++
B. ovatusATCC8843TAY155589++
B. eggerthiiATCC27754TAY155591+++
B. uniformisATCC8492TAY155592+++
B. stercorisATCC43183TAY155593+++
B. distasonisATCC8503TAY155594+++
B. merdaeATCC43184TAY155595+++
B. splanchnicusATCC29572T+
B. tectusATCC43331T+
Pr. zoogleoformansATCC33285T+
Po. cansulciNCTC12858T+
Po. endodontalisATCC35406T+
Po. gingivalisATCC33277T+
  • Genera are abbreviated as: B., Bacteroides; Pr., Prevotella; Po., Porphyromonas.

  • GenBank accession numbers for the partial 23S rRNA gene.

  • Sequences of the 16S–23S rRNA ISR were determined in this study. The GenBank accession numbers are AY153428, AY153429, AY153427 for B. caccae, B. stercoris and B. merdae, respectively.

View this table:
Table 2

Other strains used in this study

StrainStrain No.No. of strains
Bacteroides putredinisATCC29800T1
Bacteroides ureolyticusATCC33387T1
Bilophila wadsworthiaATCC515811
Bilophila wadsworthiaATCC49260T1
Campylobacter rectusATCC33238T1
Eubacterium sulciATCC35585T1
Fusobacterium gonidiaformansATCC25563T1
Fusobacterium mortiferumATCC98171
Fusobacterium naviformeATCC25832T1
Fusobacterium necrogenesATCC25556T1
Fusobacterium necrophorum subsp. necrophorumATCC25286T1
Fusobacterium nucleatum subsp. nucleatumATCC25586T1
Fusobacterium nucleatum subsp. polymorphumATCC10953T1
Fusobacterium perfoetensATCC29250T1
Fusobacterium periodonticumATCC33693T1
Fusobacterium pseudonecrophorumATCC51644T1
Fusobacterium russiiATCC25533T1
Fusobacterium simiaeATCC33568T1
Fusobacterium ulceransATCC49185T1
Fusobacterium variumATCC8501T1
Fusobacterium variumATCC277251
Mitsuokella multiacidaATCC27723T1
Mobiluncus curtisii subsp. curtisiiATCC35241T1
Mobiluncus mulierisATCC35243T1
Porphyromonas asaccharolyticaATCC25260T1
Porphyromonas cangingivalisNCTC12856T1
Porphyromonas canorisNCTC12835T1
Porphyromonas gingivalisATCC494171
Porphyromonas leviiATCC29147T1
Porphyromonas macacaeATCC33141T1
Porphyromonas salivosusNCTC116321
Prevotella biviaATCC29303T1
Prevotella buccaeATCC33574T1
Prevotella buccalisATCC35310T1
Prevotella corporisATCC33547T1
Prevotella denticolaATCC331851
Prevotella disiensATCC29426T1
Prevotella intermediaATCC25611T1
Prevotella loescheiiATCC15930T1
Prevotella melaninogenicaATCC25845T1
Prevotella nigrescensATCC33563T1
Prevotella oralisATCC33269T1
Prevotella oulorumATCC43324T1
Prevotella veroralisATCC33779T1
Sutterella wadsworthiiATCC51579T1
Bacteroides caccaeClinical isolates12
Bacteroides distasonis/merdaeClinical isolates25
Bacteroides eggerthiiClinical isolates4
Bacteroides fragilisClinical isolates51
Bacteroides ovatusClinical isolates13
Bacteroides stercorisClinical isolates3
Bacteroides thetaiotaomicronClinical isolates20
Bacteroides uniformisClinical isolates12
Bacteroides vulgatusClinical isolates15
  • Same as Sutterella wadsworthensis.

2.2 Sequencing of the 16S–23S rRNA ISR and partial 23S rRNA gene

The primer pair G23S-1 and G23S-2 (Table 3) [27], corresponding to the positions 1508–1525 and 2114–2132 of the 23S rDNA gene of Escherichia coli, was used to amplify partial 23S rRNA genes of 10 ATCC type strains of the B. fragilis group. PCR amplification was performed as follows: one or two colonies of bacterial strains were suspended in 50 µl of Tris–HCl–EDTA saline (pH 8.0), incubated for 10 min at 95°C and centrifuged at 18 600×g for 2 min to obtain the DNA as the PCR template. PCR amplification was performed in 50 µl of reaction mixture containing 1.25 U of Taq polymerase (Promega, WI, USA), 50 mM KCL, 10 mM Tris–HCl (pH 9.0), 0.1% Triton, 2.5 mM MgCl2, 0.5 mM (each) primer, 0.2 mM dNTPs, and 3 µl of the bacterial lysate as the DNA template. After initial denaturation for 2 min at 95°C, the reactions were subjected to 35 cycles comprising 30 s at 95°C, 30 s at 52°C, and 1 min at 72°C, with a final extension at 72°C for 5 min.

View this table:
Table 3

Oligonucleotide primers used in this study

PrimerSequence (5′–3′)Reference
1392AGTACACACCGCCCGTpresent study

The 16S–23S rDNA ISR regions of B. merdae, B. ovatus, and B. stercoris were amplified as described previously [21], as there were no sequence data of these species in the GenBank. The PCR was carried out as described above except the annealing temperature was 58°C.

The 16S rRNA gene fragments were amplified as previously described [28]. Briefly, almost the full length of the 16S rRNA gene was amplified by using two pairs of primers (8UA and 907B; 774A and 1485B). The PCR was carried out as described above except the annealing temperature was 45°C.

The major PCR products were excised from a 1% agarose gel after electrophoresis, were purified using a QIAquick Gel Extraction kit (Qiagen Inc., Chatsworth, CA, USA) and were sequenced directly with a Biotech Diagnostic (Biotech Diagnostic, CA, USA) Big Dye Sequencing kit on an ABI 377 sequencer (Applied Biosystems, Foster City, CA, USA). Sequences obtained have been registered at the GenBank; the accession numbers are listed in Table 1. All sequences were analyzed by multialignment using CLUSTAL-W (http://genome.kribb.re.kr).

2.3 Development of group-specific and species-specific primers

Based on the multialignment analysis data, a potential B. fragilis group-specific primer pair, Bfr-F and Bfr-R, was selected from the 16S rRNA gene, and two potential subgroup-specific downstream primers, BFR-G2 (for subgroup-II which includes B. eggerthii, B. stercoris and B. uniformis) and BFR-G3 (for subgroup-III which includes B. distasonis and B. merdae), were selected from the 23S rRNA gene. In addition, 10 species-specific primer pairs were designed from the 16S–23S rRNA ISR regions (Table 3). The primer sequences were analyzed for secondary structure formation, G+C content, and primer–dimer formation with the NetPrimer analysis software (http://www.premierbiosoft.com/netprimer/netprlaunch/netprlaunch.html). The specificities of these primers were predicted by comparison to the aligned SSU_rRNA database of the Ribosomal Database Project using the CHECK_PROBE utility [29]. These primers were designed with minimal differences in their annealing temperature within each primer set, and to yield amplification products that ranged between 200 and 700 bp and differed by at least 50 bp. The relative locations of the primers in the E. coli rRNA gene sequence are indicated in Fig. 1.

Figure 1

The alignment of the approximate location of the PCR primers and amplicon sizes of each PCR amplification system; multiplex PCR-G for subgrouping the B. fragilis group species, multiplex PCR-I, multiplex PCR-II, and multiplex PCR-III for identification to the species level. The primer set used for multiplex PCR-G comprised primers Bfr-F, Bfr-R, G23S-1, Bfr-G2 and Bfr-G3. The primer set used for multiplex PCR-I comprised primers 1392A, Bth-R, Bvul-R, Bfra-R, Bcac-R and Bova-R. The primer set used for multiplex PCR-II comprised primers Begg-F, Buni-F, Bste-F and 23R4. The primer set used for multiplex PCR-III comprised primers Bdis-F, Bdis-R, Bmer-F and Bmer-R. Arrows indicate the direction of primers.

2.4 Identification of the B. fragilis group species by two-step multiplex PCR assays

Ten species of the B. fragilis group were first grouped by multiplex PCR (designated multiplex PCR-G) and then further identified to the species level by three multiplex PCR assays (named multiplex PCR-I, multiplex PCR-II and multiplex PCR-III).

PCR amplification was carried out in a total volume of 50 µl containing 1.25 U of Taq polymerase (Promega, WI, USA), 50 mM KCL, 10 mM Tris–HCl (pH 9.0), 0.1% Triton, 2.5 mM MgCl2, 0.5 mM (each) primer, 0.2 mM dNTPs, and 3 µl of bacterial lysate as the DNA template. PCR was carried out for 35 cycles. Each cycle consisted of 95°C for 20 s for denaturation, annealing for 1 min at 52°C for multiplex PCR-G, 62°C for multiplex PCR-I, 60°C for multiplex PCR-II, and 55°C for multiplex PCR-III; extension was performed at 72°C for 30 s. A cycle of 72°C for 5 min was added to the final extension. PCR products were analyzed by electrophoresis on a 6% polyacrylamide gel followed by ethidium bromide staining.

2.5 Sensitivity of multiplex PCR assays

The sensitivities of the multiplex PCR assays were evaluated by titrating cultures of 10 ATCC strains of the B. fragilis group species (CFU 106). We made serial 10-fold dilutions of cultures with Tris–HCl (pH 7.5) and plated equal volumes (100 ml) of dilutions onto Brucella blood agars. The cultures and the dilutions were taken for DNA preparation and subsequent multiplex PCR assays. Colonies were counted after 3 days of incubation. The detection limits of multiplex PCR assays were determined with known numbers of bacteria diluted in Tris–HCl (pH 7.5).

3 Results and discussion

Members of the B. fragilis group are the anaerobes most commonly recovered from clinical specimens and the most resistant to antimicrobials. The members of this group exhibit species-to-species variability in both virulence and drug resistance. PCR assays have been developed to identify various microorganisms, including Bacteroides species [1416]. However, they target only one species; therefore, a large number of individual PCR assays may be required. In this study, using the sequences of rRNA genes as targets for specific primer selection, we developed a two-step multiplex PCR that allows rapid detection of the 10 B. fragilis group species; first, the 10 species of the B. fragilis group are differentiated into three subgroups by one multiplex PCR (multiplex PCR-G), followed by one multiplex PCR assay for each subgroup for species identification.

Although Bacteroides species have been classified phylogenetically on the basis of the 16S rRNA sequence similarity, it is not useful for differentiation of Bacteroides species because sequence diversity, except for B. distasonis, is less than 10%. The 16S–23S rRNA ISRs are more variable within Bacteroides species. The present study showed that the ISR sequence similarity ranged from 56.4 to 78.2%, lower than the 16S rRNA sequence similarity (84.2–95.4%) among Bacteroides species. The ISR sequence expresses genetic diversity among Bacteroides species better than the 16S rRNA sequence. Therefore, the ISR sequences of Bacteroides species that are available from the GenBank database and those determined in this study were used for species-specific primer selection. A part of the 23S rRNA gene was sequenced in this study, and the sequences were used for designing subgroup-specific primers that were unique to each subgroup as well as common within the species of each subgroup. In addition, a Bacteroides genus-specific primer pair was selected from the16S rRNA gene.

Based on our two-step multiplex PCR strategy, the 10 B. fragilis group species were first differentiated into three subgroups by multiplex PCR-G using a primer set comprising Bfr-F, Bfr-R, Bfr-G2, Bfr-G3, and G23S-1. Multiplex PCR-G produced a DNA fragment of about 230 bp from all 10 ATCC strains, as expected (Fig. 2, lanes 1–10). B. caccae, B. fragilis, B. ovatus, B. thetaiotaomicron and B. vulgatus were identified as subgroup-I by producing only this common DNA fragment of 230 bp (Fig. 2, lanes 1–5). B. eggerthii, B. stercoris, and B. uniformis were identified as subgroup-II by yielding a specific DNA fragment 450 bp in size in addition to the common band of 230 bp (Fig. 2, lanes 6–8), and similarly, B. distasonis and B. merdae were identified as subgroup-III by the occurrence of an additional 400-bp PCR product (Fig. 2, lanes 9 and 10). The slight size differences of the PCR products are due to length polymorphism found in amplified gene portions. The specificity of the multiplex PCR-G was evaluated by testing four ATCC strains of the non-B. fragilis group Bacteroides species and 47 ATCC strains representing closely or more distantly related Gram-negative anaerobic bacilli. Of the 51 ATCC or NCTC strains tested, six strains: B. splanchnicus ATCC 29572T, B. tectus ATCC 43331T, Porphyromonas endodontalis ATCC 35406T, Prevotella gingivalis ATCC 33277T, Prevotella cansulci NCTC 12858T and Prevotella zoogleoformans ATCC 33285T, also produced amplicons corresponding in size to the common band of 230 bp, but no others did (data not shown).

Figure 2

Polyacrylamide gel electrophoresis of PCR products from multiplex PCR assays. Lane M: 50-bp DNA ladder; lanes 1–5: subgroup-I species with an amplicon of ca. 230 bp (B. caccae ATCC 43185T, B. fragilis ATCC 25285T, B. ovatus ATCC 8483T, B. thetaiotaomicron ATCC 29148T and B. vulgatus ATCC 8482T, respectively); lane 6–8: subgroup-II species with an amplicon of ca. 450 bp (B. eggerthii ATCC 27754T, B. stercoris ATCC 43183T, and B. uniformis ATCC 8492T, respectively); lanes 9 and 10: subgroup-III species with an amplicon of ca. 400 bp (B. distasonis ATCC 8503T and B. merdae ATCC 43184T); lanes 11–15: PCR products from multiplex PCR-I, B. thetaiotaomicron ATCC 29184T (ca. 180 bp), B. vulgatus ATCC 8482T (ca. 250 bp), B. fragilis ATCC 25285T (ca. 420 bp), B. caccae ATCC 43185T (ca. 500 bp) and B. ovatus ATCC 8483T (ca. 610 bp), respectively; lanes 16–18: PCR products from multiplex PCR-II, B. eggerthii ATCC 27754T (ca. 250 bp), B. uniformis ATCC 8492T (ca. 350 bp) and B. stercoris ATCC 43183T (ca. 400 bp), respectively; lanes 19 and 20: PCR products from multiplex PCR-III, B. distasonis ATCC 8503T (ca. 220 bp) and B. merdae ATCC 43184T (ca. 310 bp), respectively.

After successfully grouping 10 ATCC strains of the B. fragilis group into three subgroups by multiplex PCR-G, these 10 strains were identified to the species level by using one of the three two-step multiplex PCR assays. Five strains of subgroup-I were identified to the species level by producing major specific DNA fragments (B. thetaiotaomicron 180 bp; B. vulgatus 250 bp; B. fragilis, 420 bp; B. caccae, 500 bp; B. ovatus, 610 bp), as expected by multiplex PCR-I (Fig. 2, lanes 11–15). B. fragilis produced an additional amplicon (ca. 650 bp) (Fig. 2, lane 13) other than the expected amplicon. Three strains of subgroup-II were identified by yielding signature DNA fragments (B. eggerthii, 250 bp; B. uniformis 350 bp; B. stercoris, 400 bp) by multiplex PCR-II (Fig. 2, lanes 16–18). B. eggerthii also produced an amplicon corresponding in size to that of B. stercoris, but was distinguished from B. stercoris by its unique amplicon (250 bp) (Fig. 2, lane 16). Two strains of subgroup-III were identified as their corresponding species by yielding signature DNA fragments (B. distasonis, 220 bp, and B. merdae, 310 bp, by multiplex PCR-III) (Fig. 2, lanes 19 and 20). B. distasonis produced an additional amplicon (ca. 250 bp) other than the expected amplicon (Fig. 2, lane 19). The specificities of these two-step multiplex PCR assays were verified by PCR amplification with DNA samples of the 51 ATCC or NCTC type strains or reference strains mentioned above. All these multiplex PCR assays generated no amplicons with DNA other than the target organisms. Although six ATCC strains of the non-B. fragilis group species were differentiated into subgoup-I by multiplex PCR-G, they generated no amplicon by multiplex PCR-I for species identification.

All DNA templates used in PCR reactions were obtained by heating cells at 95°C. The multiplex PCR assays detected between 50 and 500 CFU of each species. By decreasing the annealing temperature and increasing the magnesium concentration, the sensitivities of our procedures could be increased to 10 cells but resulted in weak cross-reactivity and non-specific amplification bands (data not shown).

The established scheme for identification of human clinical isolates was evaluated with 155 clinical isolates of the B. fragilis group that were identified to the species level by phenotypic tests previously. One hundred and thirty-one (84.5%) of the strains had concordant results between the original and PCR-based identifications, but the other 24 strains showed discordance (Table 4). B. fragilis (n=51), B. distasonis (n=19), B. eggerthii (n=4), B. stercoris (n=2) were identified correctly by both methods. However, only 45.5% of B. caccae, 53.3% of B. ovatus, 66.7% of B. merdae and 78.6% of B. thetaiotaomicron were correctly identified phenotypically. Isolates with discrepant results were further analyzed by 16S rDNA sequencing. The identification obtained from 16S rDNA sequencing showed 100% agreement with the multiplex PCR-based identification. The comparison of the two methods showed that the initial phenotypic identification was accurate only 84.5% of the time, whereas the multiplex PCR-based scheme always gave clear-cut results. It even distinguished bacteria with similar phenotypic characteristics, such as B. distasonis and B. merdae.

View this table:
Table 4

Comparison of multiplex PCR-based identification with phenotypic identification

Multiplex PCR identificationNo. of strainsPhenotypic identificationNo. of strainsNo. of strains that match
B. caccae22B. caccae1010 (45.5%)
B. distasonis2
B. ovatus5
B. stercoris1
B. thetaiotaomicron4
B. distasonis19B. distasonis/B. merdae1919 (100%)
B. eggerthii4B. eggerthii44 (100%)
B. fragilis51B. fragilis5151 (100%)
B. merdae3B. distasonis/B. merdae22 (66.7%)
B. caccae1
B. ovatus15B. ovatus88 (53.3%)
B. distasonis2
B. caccae1
B. thetaiotaomicron4
B. stercoris2B. stercoris22 (100%)
B. thetaiotaomicron14B. thetaiotaomicron1111 (78.6%)
B. uniformis3
B. uniformis10B. uniformis99 (90%)
B. thetaiotaomicron1
B. vulgatus15B. vulgatus1515 (100%)
Total131/155 (84.5%)
  • Identification results obtained from PCR assays that were different from phenotypic identification were reconfirmed by 16S rDNA sequencing.

In conclusion, a simple, rapid and reliable multiplex PCR-based method for identification of the B. fragilis group species has been established. This multiplex PCR-based identification scheme is easy, reliable, and inexpensive. It is a powerful tool for routine identification of clinical isolates of the B. fragilis group. Further studies are required to determine conditions needed to detect the B. fragilis group species directly from clinical specimens.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
View Abstract