Until recently, the estimation of bacterial biodiversity has been hampered by limitations associated with cultivating bacteria from natural environments. An uncultured fraction has been recognized to be a major component of all microbial communities (7
). The application of molecular approaches to the characterization of bacterial communities has overcome the requirement for prior cultivation of community members. In particular, the analysis of 16S rRNA genes, aided by using PCR to amplify target sequences in environmental samples, has enabled molecular ecologists to provide better estimates of bacterial diversity (1
). PCR primers for amplification of 16S rRNA genes are widely available (6
). However, a problem with several commonly used amplimers is that they have been constructed theoretically, using the (incomplete) database of 16S rRNA sequences from cultured organisms, and have not been tested systematically. Hence, empirical testing is essential to confirm PCR primer specificity prior to their use in PCRs with environmental samples.
Primers for PCR described by Lane (6
), which have been used by many other workers (5
), consistently failed to work with some difficult samples generated in our work. DNA samples extracted from various sources, including deep sea sediment, oral bacteria, and bacteria isolated from epilithon (biofilms associated with stones in lotic habitats), were found to be poor templates for amplification of the 16S rRNA gene with amplimers such as 27f and either 1492r or 1392r (numbering is based on the Escherichia coli
16S rRNA gene [3
]). We had attempted to optimize the PCR conditions for 27f-1392r by altering the annealing temperature, Mg2+
concentration, and DNA template concentration and also by including the PCR additives bovine serum albumin, Triton X-100, T4 gene 32 protein, polyethylene glycol 8000 and glycerol. Finally, we decided to redesign the amplimers, and when we used 63f and 1387r with the difficult DNA samples, we were successful.
The wide adoption of amplimers 27f, 1392r, and 1492r is empirically based, and although their utility for investigating the molecular ecology of natural bacterial communities is often assumed, to our knowledge they have not been systematically tested. In this communication, we describe a new set of amplimers which were designed to be universal for the domain Bacteria
) and their testing on a range of pure cultures and difficult natural samples.
Unless otherwise noted, genomic DNA was extracted from pure cultures, reference strains, and environmental samples by a modification of the method of Ausubel et al. (2
). DNA from marine sediment samples (1 to 2 g), collected as described previously (9
), was extracted by a modification of the method of Rochelle et al. (10
). After the lysozyme step, 100 μl of proteinase K (18 mg/ml; Sigma) was added, and the solution was further incubated for 1 h at 37°C. Second, the phenol-chloroform step was replaced by adding one-half volume of 7.5 M ammonium acetate, and the mixture was centrifuged at 11,220 × gav
for 20 min at 4°C. DNA from the three Leptospira
strains was obtained as a thermolysate (2a
Two PCR primers were designed (Oligo, version 3.4; National Biosciences Inc.) to amplify approximately 1,300 bp of a consensus 16S rRNA gene (6
): forward primer 63f (5′-CAG GCC TAA CAC ATG CAA GTC-3′) and reverse primer 1387r (5′-GGG CGG WGT GTA CAA GGC-3′) (Pharmacia). Primers 27f and 1392r (6
) were also used.
The PCR mixtures (100 μl) contained 20 pmol of each appropriate primer, 200 μM each deoxynucleoside triphosphate, Taq
extender PCR buffer (Stratagene Ltd.), 0.5 U of Taq
extender (Stratagene), and 0.5 U of Taq
polymerase (Boehringer). Approximately 200 to 300 ng of DNA from a test strain culture and subnanogram quantities of sediment DNA were added to PCRs. In addition, 2 μg of T4 gene 32 protein (Pharmacia) was included in PCRs of sediment DNA. PCR was performed with one of the following thermal cyclers: Hybaid Omnigene, Omni-E, or TR1; Perkin-Elmer 460; or MJ Research PTC-100. All cyclers were programmed to perform 30 cycles consisting of 95°C for 1 min, 55°C for 1 min, and 72°C for 1.5 min followed by a final extension step of 5 min at 72°C. PCR products were visualized by agarose gel electrophoresis (11
The specificities of the new primers were tested in PCRs with template DNA from cultures of well-characterized species representing the major groups of the domain Bacteria
(Table ) (14
). Since DNA from the Thermotogales
, green non-sulfur bacteria, and Fibrobacteria
groups was not available for testing, the sequence similarities between amplimers 63f and 1387r and the 16S rRNA genes of selected species from those groups were assessed by aligning the amplimer sequences with the appropriate prealigned sequences obtained from the Antwerp rRNA database (12
) (Table ). In the comparison of alignments shown in Tables and , one main point emerged. Amplimers 63f and 1387r successfully amplified 16S rRNA genes from species showing higher levels of theoretical 5′ mismatches than amplimer pair 27f-1392r, in some cases with 3′ mismatches in 1387r.
TABLE 1 Species of bacteria and archaea tested in PCRs using amplimer 63f and 1387f to amplify 16S rRNAgenes
TABLE 2 Theoretical alignment of sequences of amplimers 63f and 1387r with database sequences of 16S rRNA genes from species not tested byPCR
TABLE 3 Alignment of sequences of amplimers 63f-1387r and 27f-1392r with the corresponding regions of 16S rDNA genes of representative species used inPCR
In a range of experimental studies carried out in the participating laboratories, primers 63f and 1387r were used successfully and consistently to amplify 16S rRNA genes from template DNA extracted from a variety of organisms: organisms identified as belonging to the coryneform and Micrococcus genera (gram-positive, high-G+C bacteria), cultured for the first time from concrete; Eubacterium species cultured from dental abscesses; novel δ-proteobacteria (sulfate- and iron-reducing bacteria); epilithic samples; and deep sea sediments. Conversely, amplimer pair 27f-1392r failed to amplify the 16S rRNA genes of many of these test samples.
Our results provide no clear theoretical explanation for why amplimer pair 63f-1387r was so much more successful than 27f-1392r. One suggestion is that the latter amplimers are not optimal for PCR since 27f may form an intramolecular duplex with a 5′ overhang and may thus be susceptible to the 5′→3′ exonuclease activity of Taq
polymerase. Any resultant removal of 5′ nucleotides from 27f (possibly six in total) would affect the annealing temperature of the primer pair (ΔTm
of 13.6°C instead of 1.6°C) and also result in unfavorable intermolecular complementarity between 27f and 1392r, leading to binding of the 3′ ends. An alternative explanation comes from a recent computer analysis of the potential of primers to hybridize with 16S rRNA genes (4
). In this study, a primer designed for the conserved area of the 16S rRNA gene, which was also used for 63f, was found to have a greater hybridization potential than the conserved area used for 27f.
In conclusion, although 63f and 1387r showed some theoretical bias, in practice they were more successful than amplimer pair 27f-1392r and amplified 16S rRNA genes from a wider range of bacteria than other primers which are commonly used for bacterial community analysis. So far as we are aware, the other primers have not been tested in the systematic way described for 63f-1387r in this paper. The results presented here suggest that the latter primer pair may be better suited for this type of molecular ecological analysis, the aim of which is to minimize PCR bias, and underline the point that the theoretical design of PCR amplimers is only the beginning and that systematic empirical testing of the amplimers is of paramount importance.