2.2. Bacterial Isolates, DNA isolation and quantification

The thirty strains including five select agents representing twenty four bacterial genera, that were used to experimentally confirm the performance of PCR primers and probes are shown in Table 2. The bacterial cultures used in these experiments included both standard laboratory strains and de-identified patient isolates. Bacterial DNA was extracted using Phenol–Chloroform as described previously (Maloy, 1989). DNA was isolated from Bacillus anthracis, Yersinia pestis, Burkholderia mallei and Francisella tularensis in a biosafety level III laboratory certified to work with select agents (registration number 20011016-798; entity number C20031133-0125).

2.3. Primer and probe design

Universal PCR primers complementary to four conserved regions flanking two hypervariable sequences (V3 and V6) were designed using Primer Select software (Lasergene V6, DNASTAR). The primers were designed so that at least the 10-most 3’ nucleotides were complimentary to highly conserved 16 rRNA sequences. A molecular beacon probe complementary to the hypervariable region V6 was designed as described previously (Tyagi and Kramer, 1996). The sequences of all primers and the molecular beacon are shown in Table 3.

2.4. Real-time and end-point detection PCR

For the Staphylococcus epidermidis specific molecular beacon assay, real time PCR was carried out in a Smart Cycler II (Cepheid, Sunnyvale, California, USA) real time PCR thermal cycler. Each tube was loaded with 25μl final volume of a solution containing 1X Amplitaq Gold polymerase buffer, 0.03 U/μl of Amplitaq Gold polymerase enzyme (Applied Biosystems, California, USA), 4 mM MgCl₂, 250 μM dNTP mix, 0.5 pmol of each primer and 5ng/μl of the molecular beacon (Biosearch Technologies, California, USA) and the appropriate amount of chromosomal DNA (10⁶ to 1 genome equivalents) or an equal volume of water (for no DNA control reactions). PCR amplification were carried out as follows: Initial denaturation at 95°C for 10 min, followed by 45 cycles each of 95°C for 20 s, 52°C for 30s and 72°C for 30 s. Data were collected during the annealing step for analysis. End-point detection PCR was carried out in the thermal cycler Gene Amp PCR system 9700 (Applied Biosystems, California, USA). Reaction compositions were identical as described in case of the real time PCR assay, except that the reaction did not contain any molecular beacons, the annealing temperature was 55°C and 10 ng of DNA was used as templates in the assay. The PCR products for the end point assays were visualized under UV light by ethidium bromide staining after agarose gel electrophoresis. Prior to both real time and end-point detection PCR assays, the PCR cocktail (without any template DNA added) was digested with 0.1U/μl of AluI restriction enzyme (Invitrogen Life Technologies, Carlsbad, California, USA) to degrade any endogenous DNA present in the recombinant Taq Polymerase and then heat inactivated as described (20).

3.2. Region V1 (nucleotides 69-99)

The dendrogram of V1 (supplementary figs. 1) revealed that this region could be used to distinguish common pathogenic Streptococcus sp. and to differentiate between Staphylococcus aureus and coagulase negative Staphylococcus (CONS) species. An alignment of a shorter 28 nucleotide-long region within V1 spanning nucleotides 70 to 97 (numbering based on S. aureus 16S rRNA gene) for 122 sequences from 31 Staphylococcus species showed that the S. aureus 16S rRNA gene sequence contained at least 4 unique single nucleotide polymorphisms (SNPs) in this short region compared to other CONS species at positions (numbering based on S. aureus 16S rRNA gene) 73, 80, 89 and 90 (except for S. equorum and S. schleiferi which were identical at position 73 and S. lentus, S. pulvereri and S. sciuri, which were identical at positions 80, 89 and 90). S. aureus differed from the latter three CONS species at the positions 73 and 76 in V1; therefore, this short sequence is ideal for designing S. aureus specific probes.

3.3. Region V2 (nucleotides 137-242)

The dendrogram of V2 (supplementary fig. 2) revealed that this region could distinguish all of the 110 bacterial species in this study to the genus level except for (i) most Escherichia sp. and Shigella sp. and (ii) K. pneumoniae and E. aerogenes. This region was also able to distinguish among the common Staphylococcal and Streptococcal pathogens and among Clostridium, Haemophilus and Neisseria species. V2 also appeared to be the best target for distinguishing among Mycobacterial species. V2 has an average length of 100 bp, which is a relatively large target region for DNA probes. A separate analysis of three 15 – 35 bp regions (nucleotides 137-168, 182-194 and 205-242) within V2 revealed that the 13 nucleotides spanning 182-194 contained the maximum SNP variations between mycobacterial and haemophilus species and the region spanning 205-242 contained most of the SNP variation between the Staphylococcal, Streptococcal, Haemophilus, Fusobacterium, Clostridium and Neisseria species as well as other bacterial genus analyzed.

3.4. Region V3 (nucleotides 433-497)

The dendrogram of V3 (Fig. 1a), demonstrated that V3 was similar to V2 in its ability to distinguish among the 110 bacteria to the genus level. Examination of shorter regions within V3 revealed that nucleotides 456 through 479 contained the maximum number of SNPs between most bacterial species, producing almost the same dendrogram (not shown) as the full V3 sequence except for identities among Mycobacterium fortuitum, Nocardia sp., Propionibacterium acnes and Rhodococcus equi. V3 appeared to be better than V2 in distinguishing between the closely related enterobacteriaceae K. pneumoniae and E. aerogenes, and the SNP variation among different Haemophilus species in this region was also more than that in V2. V3 is shorter than V2 (65 nucleotides versus 106 nucleotides) and PCR amplification of V3 with universal primers specific to flanking sequences would result in a smaller amplicon compared to V2. The smaller amplicon size may be preferred in some real time PCR assays (Chakravorty et al., 2006).

3.5. Region V6 (nucleotides 986-1043)

Although the V6 region was only 58 bp in length, it showed considerable sequence variability and was able to distinguish among most bacterial species except for the closely related enterobacteriaceae Escherichia sp., Shigella sp. and Salmonella sp. (Fig. 1b). V6 appeared to be the best target region for assays designed to distinguish between B. anthracis and B. cereus due to a SNP at the position 988 in the B. anthracis 16S rRNA gene. This SNP was the only 16S rRNA polymorphism detected between these two closely related bacteria that was present in all the copies of their 16S rRNA genes (B. anthracis 11 copies and B. cereus 12-13 copies) (Fogel et al., 1999; Klappenbach et al., 2001). An examination of 11 CDC classified “select” agents B. anthracis, B. mallei, B. pseudomallei, Coxiella burnetti, Clostridium botulinum, C. perfringens, E. coli O157H7, F. tularensis, Rickettsia prowazekii, R. rickettsii and Y. pestis demonstrated that V6 contained unique DNA sequences that were specific for each select agent species except for E. coli O157H7. E. coli O157H7 could not be distinguished from Salmonella sp. in this region. An inspection of shorter overlapping 20-30 bp sequences in the V6 region (998-1032 and 1018-1043) revealed that nucleotides 998-1032 provided a better target for species and genus specific probes than 1018-1043 because B. anthracis and B. cereus were identical and the different Neisseria species were indistinguishable in the second region.

3.6. Regions V4, V5, V7 and V8 (nucleotides 576-682, 822-879, 1117-1173 and 1243-1294)

These four hypervariable regions (supplementary figs. 3, 4, 5 and 6) particularly V5, were less suitable for species identification due to a higher degree of sequence conservation compared to the other hypervariable regions. However, V7 may be useful for designing probes to detect most select agents. We found that B. anthracis and B. cereus differed by two SNPs in this region. However, the SNP differences were present only in three and five respectively out of the 11 copies of the B. anthracis 16S rRNA genes (Ueda et al., 1999). This might result in a less sensitive assay, as the specific probes designed to this region might not bind to all of the amplified targets.

3.7. Combining hypervariable regions

Our dendrogram analysis implied that combining V2, V3 and V6 together would provide sufficient sequence diversity to identify all the 110 bacterial species in this study to the level of the bacterial genus. Species level identification would also be possible for most but not all of the bacterial species. We therefore individually concatenated the V2, V3 and V6 sequences of each of the 110 species in this study and then generated a new dendrogram of the concatenated sequence (Fig 1c). All of the 110 bacterial species clustered within their genus on independent branches except for E. coli BL21 and S. flexneri using this approach, indicating that species-level discrimination would be possible for 97 of the 110 bacterial species. The bacteria that could be detected only to genus level using these three regions were: (i) Salmonella (spp. typhimurium / enterica typhi) (ii) B. mallei and B. pseudomallei (iii) B. parapertussis and B. pertussis (iv) Enterococcus faecalis and E. faecium (v) Brucella (spp. melitensis / suis) (vi) M. bovis and M. tuberculosis.

3.8. Universal primers and PCR sensitivity and specificity

Each hypervariable region of the 16S rRNA gene is flanked by a conserved sequence (Baker et al., 2003); this has made it possible to design “universal” PCR primers that can amplify 16S rRNA hypervariable regions from a large number of different bacterial species (Baker et al., 2003). We designed universal primers to conserved regions flanking hypervariable regions V3 and V6 based on our alignment data (Table 3). The universal primers flanking the region V3 are similar but not identical to those described previously (Baker et al., 2003). V6 is the shortest hypervariable region with the most nucleotide diversity in the sequences among bacteria analyzed. Paired with V3, most of the bacterial species analyzed here can be differentiated, as evident from their individual dendrograms (Figs. 1a and 1b). We amplified the hypervariable regions V3 and V6 from 30 bacteria, which included 20 of the most common pathogens described here, and 5 select agents (Table 2) using V3 and V6 primers. Amplicons were produced from all the bacteria tested, with the expected amplicon sizes (204 bp and 125 bp respectively) (Fig 2a and b). The small differences in the band sizes observed in some bacteria after amplification of V3 (Fig 2a) is consistent with the sequence data showing partial V3 deletions in many bacteria. In order to demonstrate that the V6 primers could be used for assays that were both sensitive and specific, we repeated the V6 PCR assay, this time using a S. epidermidis species-specific molecular beacon as a reporter molecule (SEP V6 probe; Table 3). The results show that S. epidermidis DNA could be detected with a sensitivity of 10 molecules added to the initial reaction, while control DNA from the closely related S. aureus species (which differs by three SNPs in the molecular beacon target region) was not detected even when 1,000,000 molecules of S. aureus DNA were added to the initial reaction (Fig. 3). It should be noted that universal primer-based assays would be most relevant in studies of sterile body fluids. Tests of the environment or of contained body sites would require that more specific primer sets be developed.

Open in a separate window

Fig. 2

PCR amplification of 16S rRNA hypervariable regions of from 30 bacterial species using universal primers. Lane M: 100 kb molecular weight DNA marker; Lanes 1-30: Aerococcus viridans, Arcanobacterium pyogenes; Clostricium difficile; Bacillus anthracis, Bacteroides uniformis, Campylobacter jejuni, Burkholderia mallei, Corynebacterium pseudodiptheriticum, Enterococcus galinarium, Escherichia coli O157:H7, Francisella tularensis, Klebsiella pneumoniae, Haemophilus influenzae, Haemophilus parahemolyticus, Listeria grayi, Moraxella cattarhalis, Mycobacterium tuberculosis, Oligella urethralis, Proteus vulgaris, Pseudomonas aeruginosa, Neisseria gonorrhoeae, Neisseria lactamica, Neisseria mucosa, Serratia marcescens, Streptococcus agalactiae, Streptococcus pyogenes, Streptococcus pneumoniae, Staphylococcus aureus, Staphylococcus epidermidis and Yersinia pestis. A: universal primers amplifying V3; B: universal primers amplifying V6.

Open in a separate window

Fig.3

Real time PCR plot showing amplification of S. aureus and S. epidermidis genomic DNA using universal primers to V6 and a Staphylococcus epidermidis specific molecular beacon. The number of bacterial genomes added to each PCR reaction are indicated.

10⁶ genomes 10⁵ genomes 10⁴ genomes 10³ genomes 100 genomes 10 genomes 1 genome ― DNA negative control S. aureus 10⁶ genomes

This work was supported by Public Health Service grant AI-056689 from the National Institutes of Health and Department of Defense grant DAMD 17-01-1-0787 from the United States Army Medical Research Material Command.