• Bacteria;
  • 16S rRNA;
  • PCR;
  • Subsurface;
  • Sediment


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

Culture-based techniques have traditionally been the primary tools utilized for studying the microbiology of terrestrial subsurface environments. Recently, nucleic acid-based methods have been employed to further characterize the microbial diversity in subsurface sediments and rocks, but the results have not been related to individual bacteria cultivated from the same environment. Restriction fragment length profiles of 16S rRNA genes derived from bulk community DNA or bacterial isolates were compared to determine the efficacy of PCR-based methods for studying microbial diversity and phylogeny in a deep (188 m) subsurface environment. The phylogenetic relatedness between 16S rRNA genes of enrichment cultures and individual clones was also determined through DNA sequence analysis of 16S rRNA genes. Restriction fragment length profiles from PCR clone libraries accounted for 64% of recovered isolates and 55% of the estimated culturable diversity based upon their 16S rDNA RFLP signatures. DNA sequence comparisons between the 16S rDNA of the most commonly occurring isolates and clones confirmed that similar DNA sequences were contained within the RFLP groups used to categorize the isolates and clones. For 7 of 8 RFLP groups for which DNA sequences were obtained, nearest neighbor assignments corresponded at the genus level but suggested that 16S rDNA sequences from multiple genera were contained within single RFLP profiles. Phylogenetic analysis of 16S rRNA sequences supported the nearest neighbor inferences and indicated that 16S rDNA clones derived from bulk sediment were specifically related to isolates recovered on enrichment plates. This study has shown that a majority of the cultivated aerobic heterotrophic bacteria in a subsurface sediment could be described by 16S rDNA clones obtained from directly extracted DNA, but that PCR-based methods cannot account for all organisms from a given sample. Consequently, a more comprehensive assessment of microbial diversity in subsurface (and probably other) environments can be obtained by using a combination of culture- and molecular-based techniques than by using either method alone.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

During the last 10 years, the U.S. Department of Energy's (DOE) Subsurface Science Program has sponsored numerous geomicrobiological investigations of subsurface environments. This effort has revealed surprising microbial and metabolic diversity at depths previously thought to be devoid of microbial life [1–6]. Most of these investigations have relied upon traditional culture-based methods to assess the presence, abundance, diversity, spatial distribution and phylogenetic traits of subsurface microbes. Frequently, these investigations are hampered by an inability to culture microbes, such that many microbiological properties of the sediment cannot be ascertained. The ‘culturability’ problem is a general and well recognized dilemma for microbial ecologists regardless of the environment in question, but is particularly relevant to subsurface studies where many organisms reside in dormant or low-activity states, or are viable but non-culturable [7].

The inability to culture organisms from specific habitats has been overcome to some extent by culture-independent nucleic acid techniques. 16S rRNA and DNA, in particular, have been attractive targets for ecological studies due to the phylogenetic inferences that arise from sequence analysis of these molecules [8, 9]. Investigations involving 16S rDNA characterization have served to emphasize the biases associated with culturing, and the power of nucleic acid approaches for describing natural communities and identifying novel organisms [10–13]. The uniqueness of 16S rDNA clones and microbial diversity estimates based on these results, however, are relative to baseline information of a particular environment. Biases associated with community DNA extraction, PCR amplification and cloning strategies result in clone libraries that are not quantitative indicators of microbial abundance [14–17], but may also result in a qualitatively misleading portrait of the microbial phylogenetic diversity.

As part of DOE's Subsurface Science Program, a collection of culturable aerobic and anaerobic subsurface bacteria have been recovered and preserved in the Subsurface Microbial Culture Collection (Florida State University, Tallahassee, FL and Oregon Graduate Institute, Beaverton, OR) [18]. It is still not known how representative these cultures are of their respective communities or of subsurface environments in general. We have recently generated a total community 16S rDNA clone library from a subsurface paleosol [17, 19] from which culturable aerobic heterotrophic bacteria were also isolated. The 16S rDNA clones were characterized by restriction enzyme profiles of full-length 16S rDNA inserts. Ninety-eight distinct RFLP types were identified, suggesting even greater microbial diversity than previously determined using culture-based methods. A comparison of 16S rDNA clones generated by PCR amplification of total community DNA and 16S rRNA genes obtained from individual isolates can provide a frame of reference for clone interpretation and, potentially, a broader assessment of microbial diversity in subsurface sediments. More importantly, such comparisons should provide insights into the effectiveness of direct 16S rDNA PCR methods for identifying known organisms in a particular environment. The purpose of this research, then, was to determine the efficacy of PCR-based methods for describing subsurface microorganisms that have been recovered using cultural methods, and to link the occurrence of specific 16S rDNA clones to those isolates.

2Materials and methods

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

2.1Sediment collection and handling

A subsurface paleosol was recovered from a borehole drilled on the Hanford Site in south-central Washington state at a depth of 187.7–188.4 m below the surface. Cable-tool drilling, which does not involve circulating mud or air, was used to obtain the subsurface core sample [5, 20]. The core was sealed inside a sterile anaerobic glovebag and immediately transported to the lab for processing in a laminar flow hood with sterile implements. The paleosol was physically and texturally homogeneous and contained approximately 1–8 g kg−1 total organic carbon; precise measurements for oxygen content were not obtained, but the Eh of sediment slurry was approximately 50–80 mV, the lowest among all samples obtained from the borehole. The upper half of the core sample was disaggregated into small crumbs (homogenized treatment), while the lower half was maintained as an intact core (intact treatment). One aliquot of homogenized sediment was taken immediately and frozen at −80°C (t=0 sample) for DNA extraction, and another aliquot processed immediately for culturable bacteria (see below). The remaining sediments were stored aerobically at the in situ temperature (17°C), and aseptically resampled several times thereafter (1, 3, 10 and 21 weeks) in an attempt to maximize the number and diversity of culturable organisms and evaluate a community-level response to treatment [17, 19].

2.2Recovery of viable bacteria

Populations of viable aerobic heterotrophic bacteria were recovered from the sediment by first blending samples in 0.1% Na4P2O7·10H2O (pH 7.0) and spread-plating 10-fold serial dilutions of sediment (in phosphate-buffered saline) in triplicate on 1% peptone-tryptone-yeast extract-glucose agar (1% PTYG) [21] or optimal plate count agar (OPCA) [22]. Plates were incubated for 2–3 weeks at 22°C, and the number of bacterial colonies on appropriate dilutions was determined. Morphologically distinct isolates from all samples were streak purified on 1% PTYG agar and deposited in the Subsurface Microbial Culture Collection (SMCC) maintained at Florida State University (Tallahassee, FL). Individual isolates were retrieved from the SMCC on 10% PTYG agar slants and shipped to our labs in Richland, WA at 4°C. Cultures were grown to exponential phase in 10% PTYG liquid media and stored frozen at −20°C for PCR amplification of 16S rRNA genes. Since colony morphotype was the primary criterion upon which isolates were selected, ‘morphotype’ and ‘isolate’ are considered to be synonymous within the context of this study.

2.3Acridine orange direct counts

Acridine orange direct counts (AODC) were performed as previously described [23] with modifications [24]. Twenty microscopic fields on each of five slides were counted at 400× total magnification, yielding a detection limit of approximately 1.5×104 cells g−1 wet sediment.

2.4Construction of 16S clone library from total community DNA

A 16S rDNA clone library was generated for each sediment treatment and time point, as previously described [17, 19]. For each of the 8 sediment treatments, total community DNA was extracted from 20 g sediment by a direct lysis technique, and 16S rDNA amplified with universal eubacterial primers corresponding to E. coli positions 7–27 and 1406–1392. The primers contained cloning tails specific for the CloneAmp™ system (Life Technologies, Gaithersburg, MD), which utilizes uracil DNA glycosylase to generate sticky ends for ligase-free, unidirectional cloning into the pAMP1 vector. All PCR reactions were performed in a model 480 thermal cycler (Perkin Elmer, Foster City, CA) under 50 μl oil. Final reaction conditions were 1 μl template DNA, 10 mM Tris pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 200 μM each dNTP, 0.2 μM each primer, 1.5 μg T4 gene 32 protein (Boehringer Mannheim), and 2.5 units ld-Taq polymerase (Perkin Elmer) in a final reaction volume of 100 μl. Template DNA and primers were initially heated under the oil overlay to 80°C (hot start) before the remaining reaction components were added. 16S rRNA genes were amplified for 5 cycles at 94°C for 1.5 min, 55°C for 30 s, 72°C for 2 min followed by 40 cycles at 94°C for 1 min, 65°C for 30 s, 72°C for 2 min with a 5 s extension per cycle. After amplification, tubes were incubated an additional 20 min at 72°C before cooling to 4°C. Five replicate PCR reactions were performed for each sediment extract and template dilution, and like reactions combined prior to gel purification.

Amplification products were gel purified and ligated into pAMP1 according to the manufacturer's instructions (Life Technologies, Gaithersburg, MD). Aliquots of each ligation reaction were transformed into competent MAX Efficiency DH5α™ cells, and recombinant clones detected by blue-white colony selection on LB plates containing 100 μg ml−1 ampicillin. Transformants were grown overnight in 2 ml LB medium containing 100 μg ml−1 ampicillin; one half of each overnight culture was frozen in 15% glycerol at −80°C, and the other half was frozen directly at −20°C.

2.5RFLP analysis of 16S genes from cultures and clones

Frozen bacterial culture was used directly as template for amplification of 16S rRNA genes. 1 μl of culture was resuspended in 9 μl water containing 0.2 μM each of primers 7f and 1392r [17]. Cells were lysed by heating in a Perkin Elmer 480 thermal cycler under 25 μl oil overlay at 99°C for 5 min. After initial denaturation, samples were held at 80°C while 15 μl of Taq master mix was added to each tube. Final reaction conditions were 10 mM Tris pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 200 μM each dNTP, 0.2 μM each primer, and 0.625 units ld-Taq polymerase (Perkin Elmer). 16S rRNA genes were amplified by 5 cycles at 94°C for 90 s, 55°C for 30 s, 72°C for 120 s followed by 30 cycles at 94°C for 60 s, 65°C for 30 s, 72°C for 120 s with a 5 s extension per cycle. PCR amplification was followed by a final incubation at 72°C for 20 min.

Plasmid DNA containing individual 16S rDNA inserts obtained from total community DNA was isolated by a modified alkaline lysis/polyethylene glycol precipitation protocol (Applied Biosystems). Plasmid DNA was diluted to approximately 1–2 ng μl−1, and 1 μl used as template for PCR. The template DNA and primers were initially heated to 80°C, and then the Taq master mix added as above. Cloned 16S rDNA inserts were reamplified by 30 cycles of 94°C for 60 s, 65°C for 30 s, 72°C for 120 s with a 5 s extension per cycle.

RFLP analysis of full-length amplification products from either bacterial isolates or cloned inserts was performed as described previously [17] with the enzyme CfoI, which routinely generated 4–7 bands that were easily resolved on 3% MetaPhor™ agarose gels.

2.6DNA sequencing and DNA sequence comparisons

Plasmid DNA, containing cloned 16S rRNA genes from total community DNA, was isolated as described above and sequenced with DyePrimer chemistry and a cycle sequencing protocol, all according to the manufacturer's instructions (Perkin Elmer). For isolates, 16S rRNA genes were PCR-amplified from genomic DNA and PCR products sequenced as described elsewhere [25]. Sequences were resolved on 6% acrylamide gels run on an Applied Biosystems 373 or 377 sequencer (Perkin Elmer) and edited within the ABI Analyst software package. DNA sequences were compared against the unaligned small subunit rRNA database of the Ribosomal Database Project [26] for the nearest phylogenetic neighbors, and phylogenetic trees constructed with distance matrix and maximum likelihood algorithms contained within GDE (Genetic Data Environment, available by anonymous ftp from


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

3.1Direct microscopic and viable plate counts

Plate counts of viable aerobic heterotrophs increased by a factor of 103–104 for homogenized sediments and 102–103 for intact sediments following sampling, whereas total microscopic direct counts only increased by 101.7 in the homogenized sediment and 101 in the intact sediment (Table 1). By recording the dilution from which the isolates were recovered, an estimate of their in situ abundance could be calculated. For example, the estimated culturable population from the homogenized sediment at 10 weeks post-sampling was 3.4×105 cfu g−1 on 1% PTYG agar. The individual morphotypes isolated at this time point were recovered from the 10−4 dilution plate, and were therefore present in the sediment at ∼104 cfu g−1. Thus, the isolates from the homogenized sediment at 10 weeks post-sampling represented ≥10% of the culturable isolates from this sediment and time point. Based upon these estimates for the 7 remaining sediment samples, distinct colony morphotypes recovered from the dilution plates represented approximately 1–10% of the cultured aerobic heterotrophic population at any given time. Relative to the total biomass, however, these individual morphotypes ranged from as little as 0.0001% of the total population at the initial sampling, to ∼1% for homogenized sediments and 10% for intact sediments at 21 weeks post-sampling. In total, 100 morphologically distinct isolates were recovered from dilution plates, representing the most dominant morphotypes at the various time points. Most of the isolates were from the 21 week post-sampling time point, due in part to a more intense effort to select all available colonies regardless of their morphotype, and a general increase in the culturable population typically observed in stored subsurface sediments [27].

Table 1.  Recovery of culturable aerobic heterotrophs from a subsurface sediment
  1. H=homogenized sediment condition; I=intact sediment condition. Standard deviations are shown for acridine orange direct counts (AODC).

Sampling timeNo. of recovered morphotypesPlate counts (log)AODC (log)
1 week462.±4.9
3 weeks324.±5.34.8±4.5
10 weeks735.±5.45.3±5.1
21 weeks42315.±5.95.5±5.1

3.2RFLP analysis of 16S genes from culturable isolates

RFLP profiles for individual isolates were assigned to 20 unique RFLP groups (A–T, Fig. 1) based on identical RFLP patterns. Forty-eight of the isolates (48%) from the various time points grouped into three RFLP types (A, B, C). The next three most dominant RFLP groups accounted for an additional 19 isolates (19%). The remaining 14 RFLP groups each accounted for 1–4 additional isolates. Isolates unique to homogenized sediments fell into RFLP groups B, E, F, G, L, and M. Isolates unique to intact sediments were described by RFLP groups P, Q, R, S, and T. All other RFLP groups contained isolates recovered in both the homogenized and intact sediments. The RFLP profile for one isolate (Q) contained more bands than could have arisen from a single 1400 bp molecule, and may therefore represent a mixed culture or more than one distinct 16S rRNA sequence within a single microorganism.


Figure 1. Appearance of specific RFLP groups with time. Parentheses indicate the percentage of total culturable isolates or clones identified by a single RFLP group. Total number of cultures=100. Total number of clones=305.

Download figure to PowerPoint

3.3Comparison of RFLP profiles of isolates to those from a 16S rDNA clone library

The community of morphologically distinct, heterotrophic aerobes numbered 100 isolates and was represented by 20 CfoI RFLP groups. In contrast, the total community 16S rDNA clone library from all intact and homogenized sediments and all time points contained 98 different RFLP types after analysis of 744 full-length 16S rDNA PCR products [19]. The RFLP profiles developed for the original clone library included some vector sequence, which resulted in 2 bands of altered mobility relative to the RFLP profiles generated from individual isolates. However, a preliminary comparison of the two data sets revealed 38 RFLP patterns from the clone library that were closely related to the profiles generated from the culturable isolates. To more rigorously compare RFLP patterns of clones and isolates, plasmid DNA was isolated for each of these 38 clones, and RFLP profiles generated using amplification and restriction enzyme conditions identical to those for the isolates. In total, the RFLP profiles from 64% of the isolates, representing 11 (55%) of the 20 RFLP groups, were among those found in the clone libraries from PCR-amplified community DNA. Of the 744 16S rDNA clones from the original clone libraries, 305 (41%) matched RFLP patterns identified among the isolates (Fig. 1). Of these 305 16S rDNA clones, 43% clustered into RFLP group A, and 26% clustered into group E. Groups F and R accounted for an additional 21% of clones. Clones unique to homogenized sediments were assigned to RFLP groups F, M, and P. Clones unique to intact sediments were assigned to RFLP groups C, H, and I. Clones found in both homogenized and intact sediments were described by RFLP groups A, D, E, O, and R.

Whereas the clone library constructed from directly extracted DNA accounted for 55% of the culturable diversity, RFLP types from culturable aerobic heterotrophic bacteria represented only 11% of the genetic diversity from the entire clone library based upon RFLP patterns. There was no apparent relationship between the number of isolates within an RFLP group and clone abundance in the same RFLP group, although 4 of the 6 most dominant groups of isolates (6–21% of cultures) were detected in the clone library. Three of these 6 RFLP groups were also dominant (9–43% of clones) in the clone library. With respect to sediment treatment (homogenized vs. intact), 75% of the homogenized clone RFLP groups matched RFLP profiles from cultures obtained from the homogenized sediment. Likewise, 88% of the intact clone RFLP groups matched RFLP profiles from intact sediment isolates. The direct PCR method generally did not identify specific RFLP types at time points earlier than detected by culturing (Fig. 1). For example, the PCR technique recovered 16S rDNA sequences from groups D, E, O, and R before they were identified as culturable isolates on plates, but groups A, B, C, and I were detected among the culturable isolates before they appeared among the 16S rDNA clones.

3.4DNA sequence comparisons

As part of a separate study, an internal region of the 16S rDNA from the bacterial isolates was sequenced and compared to the Ribosomal Database Project for the nearest phylogenetic neighbors. For those RFLP groups that were common to both the clones and the isolates and for which DNA sequences were available, the nearest neighbor check showed clones and isolates to be within the same genera in all but two RFLP types (C and I; Table 2). Some of the discrepancy between isolate and clone nearest neighbor assignments and/or sequence similarity values may be related to the region of the 16S molecule that was sequenced or the length of the sequence used for comparison against the RDP. Therefore, the 16S rDNA from some of the more abundant isolates was resequenced from the 5′ end (as indicated in Table 2) and used for direct sequence comparison with individual 16S rDNA clones. Where both the isolate and clone sequences overlapped (e.g. the 5′ end of the 16S rRNA gene), sequences were manually aligned with reference strains taken from the nearest neighbor comparisons (Table 2) and subjected to phylogenetic analysis with distance matrix and maximum likelihood methods. In all cases, the phylogenetic trees supported the nearest neighbor assignments based upon the newly sequenced region of the 16S rRNA genes (Fig. 2). Clone and isolate sequences from RFLP groups E and O formed a coherent clade most closely related to Variovorax paradoxus, RFLP group D formed a cluster of sequences related to Acinetobacter, and sequences representing RFLP group H were closely related to each other and environmental strain UW103/A31. Clones from RFLP group C were clearly unrelated to the isolates categorized by this RFLP profile.

Table 2.  Comparison of Ribosomal Database Project nearest neighbor assignments for isolates and total community 16S rDNA clones
RFLP GroupIsolate16S RegionRDP Nearest NeighborRDP SimilarityCloneRDP Nearest NeighborRDP Similarity
  1. Letter prefixes before numbers indicate sediment condition where H=homogenized sediments and I=intact sediments. t0=initial sediment sample. Numbers immediately after sediment condition indicate timepoints post-sampling; e.g. H1=homogenized sediment, 1 week post-sampling. Int=internal sequence from 16S rDNA; 5′=5′ sequence from 16S rDNA. ND=not determined. All clones were sequenced from the 5′ end and share the sediment sample and timepoint prefixes.

At0-31-003IntPseudomonas flavescens0.822H10.1Pseudomonas putida0.862
 H3-32-2055′Xanthomonas maltophilis0.677H10.2Pseudomonas putida0.856
 H10-32-3025′Pseudomonas testosteroni0.688H10.4Pseudomonas putida0.885
 H10-32-310Int‘Flavobacterium’ lutescens0.805H10.11Pseudomonas putida0.846
 H21-32-403Int‘Flavobacterium’ lutescens0.735H10.24Pseudomonas putida0.913
 H21-32-4075′Erwinia herbicola0.859H21.10Pseudomonas putida0.887
 H21-32-411IntPseudomonas putida0.847H21.11Pseudomonas putida0.896
 H21-32-416Int‘Flavobacterium’ lutescens0.729H21.13Pseudomonas putida0.892
 H-2132-419IntPseudomonas putida0.863H21.17Pseudomonas putida0.881
 H-2132-422IntPseudomonas putida0.755I3.6Pseudomonas putida0.889
 H-2132-436IntPseudomonas putida0.846I10.11Pseudomonas putida0.880
 H-2132-438IntPseudomonas putida0.821I21.4Pseudomonas putida0.884
 H-2132-440Int‘Flavobacterium’ lutescens0.617   
 H-2132-4415′Pseudomonas putida0.837   
 I21-33-403Int‘Flavobacterium’ lutescens0.735   
 I21-33-414IntPseudomonas putida0.777   
 I21-33-4225′‘Flavobacterium’ lutescens0.828   
 H10-32-301IntXanthomonas campestris0.757   
 H21-32-405IntXanthomonas campestris0.760   
 H21-32-420IntGordona amarae0.744 No match 
 H-21-32-430IntXanthomonas campestris0.822   
 H21-32-449IntXanthomonas campestris0.806   
 H21-32-450IntXanthomonas campestris0.829   
CH3-32-201IntMycoplana bullata0.922I3.19Bacillus gordonae0.719
 H21-32-4245′Mycoplana bullata0.861I3.29Bacillus gordonae0.691
 H21-32-425IntMycoplana bullata0.903I10.54Bacillus gordonae0.722
 H21-32-426IntMycoplana bullata0.876I21.77Pseudomonas testosteroni0.716
 I10-33-3055′Mycoplana bullata0.886   
 I21-33-413IntMycoplana bullata0.869   
 I21-33-417IntMycoplana bullata0.919   
 I21-33-4215′Mycoplana bullata0.773   
 I21-33-4345′Mycoplana bullata0.820   
DH3-32-2035′Acinetobacter calcaceticus0.764t0.29.fAcinetobacter anitratus0.818
 I3-33-2035′Acinetobacter anitratus0.696I3.45Acinetobacter anitratus0.835
EH10-32-3055′Pseudomonas testosteroni0.694H1.5.fRhodocyclus tenuis str. 37600.725
 H10-32-3085′Pseudomonas testosteroni0.621H1.7.fRhodocyclus tenuis str. 37600.710
 H10-32-3155′Variovorax paradoxus0.849H3.1Variovorax paradoxus0.741
 H21-32-4395′Rhodocyclus tenuis0.808H3.3Pseudomonas testosteroni0.726
 H21-32-4435′Variovorax paradoxus0.857H3.15Pseudomonas testosteroni0.751
 H21-32-4475′Variovorax paradoxus0.809H21.7Pseudomonas testosteroni0.759
     H21.8Pseudomonas testosteroni0.740
     H21.12Pseudomonas testosteroni0.742
     H21.20Pseudomonas testosteroni0.736
     I3.31Variovorax paradoxus0.723
FH21-32-402ND  H3.4Erythromicrobium ramosum0.818
 H21-32-437ND  H3.7Erythromicrobium ramosum0.874
     H3.13Erythromicrobium ramosum0.857
     H3.17Erythromicrobium ramosum0.868
     H3.18Erythromicrobium ramosum0.861
     H3.19Erythromicrobium ramosum0.867
     H21.22Erythromicrobium ramosum0.876
GH21-32-404IntDermatophilus congolensis0.613   
 H21-32-446IntDermatophilus congolensis0.798 No match 
 H21-32-452IntDermatophilus congolensis0.773   
HH21-32-4065′str UW 1030.777I21.81str. UW 103/A10.797
II21-33-4245′‘Flavobacterium’ lutescens0.840I10.23Alcaligenes xylosoxidans0.829
JH21-32-415IntAzospirillum sp.0.885   
 H21-32-417IntAzospirillum sp.0.904   
 I21-33-409IntAzospirillum sp.0.906   
 I21-33-425Int   No match 
 I21-33-428IntAgrobacterium tumefaciens0.837   
KH21-32-421IntArthrobacter globiformis0.893   
 I3-33-204Int   No match 
 I10-33-307IntArthrobacter globiformis0.907   
LH21-32-429IntBacillus thermoruber0.079 No match 
MH21-32-431ND  H21.84Clavibacter xyli 
NH21-32-434IntCaulobacter sp. FWC140.832   
 I21-33-427IntCaulobacter sp. FWC140.806   
 I21-33-433IntCaulobacter sp. FWC140.788 No match 
OH21-32-4515′Variovorax paradoxus0.809I3.46Rhodoferax fermentans0.786
 I21-33-4295′Agrobacterium tumefaciens0.815I10.65Pseudomonas testosteroni0.726
     I21.10Variovorax paradoxus0.726
PI3-33-113ND  H1.6Methylobacterium sp. str. F730.943
Q*  Probable Contaminant    
RI21-33-406IntMycoplana bullata0.926H1.39.fArthrobacter simplex0.705
 I21-33-423IntNocardioides plantarum0.793H21.82Arthrobacter simplex0.743
 I21-33-431IntNocardioides plantarum0.814I3.18Nocardioides luteus0.751
     I10.1Nocardioides luteus0.733
     I10.17Nocardioides luteus0.735
     I21.2Nocardioides luteus0.740
     I21.8Nocardioides luteus0.739
     I21.16Nocardioides luteus0.723
SI21-33-410IntGordona amarae0.764   
 I21-33-418IntClavibacter michiganensis0.805 No match 
TI21-33-412ND   No match 

Figure 2. Relationship between 16S rDNA clones recovered from total community DNA and 16S rRNA genes from aerobic, heterotrophic isolates inferred from maximum likelihood analysis. Sequence alignment included 198 homologous nucleotides from positions 127–183 and 218–358 (E. coli numbering). Affilitations based upon RFLP typing are shown following clone or isolate numbers. Isolates and clones can be found in Table 2. Scale bar represents 5 mutations per 100 nucleotide positions.

Download figure to PowerPoint

Interestingly, sequences among isolates from RFLP group A were related to 4 different genera based upon the RDP nearest neighbor check, whereas all corresponding 16S rDNA clones (130) recovered from bulk sediment DNA corresponded to Pseudomonas putida. The multiple genera of RFLP group A isolates are closely related phylogenetically and physiologically, so that PCR or sequencing errors might result in differing nearest neighbor assignments. However, distinct genera were isolated at different times post-sampling, which suggests that the multiple genera identified within RFLP group A were not due to PCR or sequencing errors. For example, isolate H1-32-113 (Pseudomonas flavescens) was isolated from the original sediment but was not isolated at any time thereafter. At 3 weeks post-sampling, isolates related to Pseudomonas testosteroni and Xanthomonas maltophila were recovered. By 21 weeks post-sampling, relatives of Pseudomonas putida were the most dominant biotype within RFLP group A isolates. The exclusive occurrence of Pseudomonas putida clones within RFLP group A, and their detection in 5 different clone libraries (3, 10 and 21 weeks post-sampling in both intact and homogenized sediment), may have therefore resulted from PCR amplification biases for these sequences. For broader-scale phylogenetic affiliations, however, DNA sequence analysis supported the grouping of clones and isolates into RFLP groups described by the restriction enzyme CfoI.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

The purpose of this research was to determine the efficacy of total community DNA extraction and PCR amplification of 16S rDNA for describing the phylogenetic diversity of subsurface microorganisms, some of which were concurrently isolated by enrichment culture. One hundred aerobic, heterotrophic bacteria were selected on the basis of distinct colony morphology and were categorized into 20 distinct groups based on RFLP analysis of 16S rRNA genes. Relative to the total microbial population, however, the most abundant isolates (recovered at 21 weeks) represented ≤1% of the total count as determined by AODC. The 18 distinct RFLP groups retrieved from the 21 week time point, then, still constituted a minority of the total microbial population. While the PCR-amplified clones represented at least 98 distinct microbial types, we cannot asses their actual abundance in situ due to the potential for PCR amplification bias [14–16]. In those cases where the RFLP profiles of the isolates and clones matched, however, a qualitative relationship between isolate abundance and clone abundance may be established. Invariably, the proportion of clones ascribed to any one RFLP group did not correspond with the in situ abundance of the isolates having the same RFLP profile.

Subsurface bacteria show limited physiological and biochemical diversity within a colony morphotype [28], but they can be highly pleiomorphic on solid media (unpublished observations). Therefore, colony morphology alone cannot always differentiate between genetically distinct subsurface isolates. The actual genetic diversity of aerobic heterotrophs recovered during this study was clearly less than the 100 morphologically distinct isolates (e.g. 9 isolates of Mycoplana bullata comprising RFLP group C). Alternatively, RFLP analysis of a single gene using a single restriction enzyme characterizes only a very small portion of the genome and individual genes. Therefore, more closely related sequences may remain grouped together within a single RFLP profile. This, too, was observed and supported by DNA sequence data, such that the 20 RFLP groups underestimated the total genetic diversity contained within the 16S rRNA gene sequences in the isolates (e.g. RFLP group A isolates) and clones.

Recently, a computer analysis of terminal restriction fragment lengths of small subunit rRNA sequences indicated that the majority (73%) of restriction fragment length polymorphisms in these genes are the result of insertions and deletions, rather than single base-pair substitutions [29]. Consequently, we would expect that DNA sequences within an RFLP type would be closely related, while those sequences from separate RFLP groups would be more widely divergent. To substantiate the Ribosomal Database Project nearest neighbor assignments and more precisely determine the relationship between 16S rDNA sequences within an RFLP group, phylogenetic trees were constructed with all genera represented in Table 2 and the 5′ 16S rRNA sequences derived from the isolates and clones from this study. In all cases, phylogenetic trees substantiated the nearest neighbor assignment, and DNA sequence analysis generally supported broader-scale (genera, family) phylogenetic affiliations based upon CfoI, 16S rDNA RFLP signatures. One exception is RFLP group C, where the isolates were all related to each other and Mycoplana bullata, while the clones were related to Bacillus and Variovorax. A similar situation occurred for the two sequences from RFLP group I.

Some of the sequence variation among individual isolates within an RFLP group (e.g. RFLP group A) resulted from sequence ambiguities and base-calling errors. That is, the rRNA genes from individual isolates were amplified directly from genomic DNA extracts and not individual clones. Consequently, DNA sequence heterogeneity between different rRNA operons [30] probably resulted in a mixed pool of 16S rRNA amplification products that were simultaneously sequenced. The frequency of base-calling ambiguities was approximately 2% for isolates described in Table 2, which may alter inferred phylogenetic affiliations in a group of microorganisms as closely related as Xanthomonas and Pseudomonas (Fig. 2). The unrelatedness of clones and isolates from RFLP groups C and I, however, cannot be ascribed to sequencing ambiguities alone. In these cases, improper RFLP classification may have resulted from the limited resolving power of agarose gels, an inability to visualize smaller (e.g. ≤30 bp) restriction fragments in an RFLP profile, or errors resulting from manual comparison of RFLP profiles across multiple gels. As a first step in analyzing large collections of clones or isolates, however, RFLP profiling provided a reasonably accurate measure of phylogenetic relatedness and a guide for more intense sequencing efforts.

An abundance of Pseudomonas putida clone types in the PCR clone libraries and lack of any other species within type A clones suggests that the PCR was biased for these sequences at the exclusion of others, such as Xanthomonas and related templates. PCR bias could result from a number of factors, including differential denaturation of template, differential primer annealing, different target gene length, degraded or impure template, or a small number of total genomes within a PCR reaction [31–33]. In addition, differential DNA recovery or altered PCR conditions might affect the recovery of specific 16S rDNA sequences. Numerous precautions were taken to minimize bias during the construction of the total community 16S rDNA clone library [17], yet 45% of the RFLP groups (36% of isolates) corresponding to culturable aerobic heterotrophs were not detected in the clone library. Some of the discrepancy may be related to the primer sequences used during the generation of the clone libraries. According to Brunk et al. [29], primer 1392r (without the cloning tail) has significant hybridization potential to 97% of all eubacterial small subunit 16S rRNA gene sequences contained within the Ribosomal Database Project [26]. However, primer 7f (also without the cloning tail) would be expected to hybridize with only 70% of the same eubacterial sequences. That we accounted for 64% of the isolates within the clone library is an intriguing coincidence, and, coupled with the observation that most isolates and 16S rDNA clones recovered from this sediment are fairly unique (i.e. ≤0.8 Sab to Ribosomal Database Project sequences [34]), may indicate that a more critical evaluation of 16S rDNA primers will be required for further 16S rDNA analyses of microbial diversity in terrestrial subsurface environments. The observation that PCR of directly extracted DNA identified certain RFLP patterns only after they had also been detected as cultures on plates for several weeks also suggests that amplification of specific 16S rDNA sequences may be influenced by the genetic background in which it exists. Such a possibility warrants further study.

PCR methods identified far more diversity (98 RFLP types [19]) than did selective plating (20 RFLP types), reflecting the bias imposed by specific culture conditions (aerobic, heterotrophic). However, fully 36% of the isolates and 45% of the estimated diversity among the isolates (RFLP groups) were not detected by the PCR method, indicating that the PCR and cloning techniques are also unable to amplify or capture all 16S rRNA genes contained within a given sample. In addition, PCR did not detect several of the isolates that were relatively abundant (∼1%) members of the total microbial community. Most notable are those isolates within RFLP group B (identified as Xanthomonas), which were isolated at early and late time points, and in high abundance (≥10% of culturable bacteria, ≥1% of total bacteria). Related sequences were not found among the clones with matching RFLP profiles, or from any of the other clones analyzed from this sediment [34]. However, there was a fairly strong relationship between the RFLP groups identified among the homogenized and intact isolates and clones, with 75% and 88% of the clone RFLP groups obtained from homogenized sediment and intact core, respectively, corresponding to RFLP groups among isolates obtained from the same sediment treatment. This result suggests that the 16S rDNA clones obtained from each sediment treatment (homogenized or intact) are fairly reliable indicators of the genetic diversity of culturable aerobic microorganisms recovered from the same (homogenized or intact) sediment environment.

It is possible that we could have identified more of the isolates (and RFLP groups) within the clone libraries had we examined a greater number of the clones. In total, 1994 clones were obtained from this sediment, but only 841 (42%) were analyzed for full-length 16S rDNA inserts. Of the 841, 744 (88%) contained full-length 16S rDNA fragments [19]. It is unlikely that continued analysis of clones from the homogenized sediment would have significantly reduced the discrepancy between the culturable and clone RFLP types, because this set of clones was dominated by only a few RFLP groups [19, 35]. On the other hand, the clone libraries generated from the intact core consisted of many RFLP groups that contained only a few clones within each RFLP type. Further analysis of the intact clone libraries, then, may indicate that more of the aerobic, culturable isolates are represented among the clones than reported here.

Ward et al. [36] have reviewed several instances where DNA sequence information from 16S rDNA clones have been compared to the 16S rDNA sequences of isolates. In extreme environments, it appears that 16S rDNA clones are reasonably representative of the total microbial diversity, due in part to relatively less complex microbial community structure in such environments. In environments where the microbial community is more complex, however, the community profiles obtained by culturing or clones can vary dramatically. For example, when an acidic Australian soil was analyzed for Streptomyces sp., more than 50 isolates were recovered but none of the 113 16S rDNA clones derived from bulk DNA matched any of the 16S rDNA sequences from the isolates. In fact, only 2 of the 113 clones were related to Streptomyces sequences even though a Streptomyces-specific primer was used for PCR from soil DNA [37]. Our research has shown that the majority of RFLP types among culturable subsurface aerobic bacteria are represented in the 16S rDNA clone libraries constructed from directly extracted DNA, more than previously demonstrated in soil or sediment. That the approaches used in this study described up to 55% of the (broad-scale) genetic diversity among known aerobic, heterotrophic bacteria in a subsurface sediment may indicate that the microbial community in this subsurface environment is less complex (i.e. less microbial diversity) than those in near-surface environments. Alternatively, the relatively high representation of isolates within the clone libraries may have resulted from the specific sediment treatments and resampling over time, or reflect technical improvements in DNA extraction and cloning that have not been previously used.

Continued DNA sequence analysis of 16S rDNA clones and 16S rRNA genes from subsurface isolates will reveal the phylogenetic identity of these organisms and provide the necessary information for improved primer design, such that future molecular studies might describe even more of the total diversity in subsurface microbial communities, including culturable and non-culturable organisms. Even though PCR or selective enrichment techniques alone did not describe the full extent of microbial diversity in this subsurface sediment, it is clear that nucleic acid-based methods will be essential for continued, rigorous assessment of the phylogeny and genetic diversity of microbial communities in subsurface environments. A more complete analysis of microbial phylogeny, coupled with geochemical and geological characterizations of their environment, can then be used to establish stronger linkages between ‘phylotype’ and ‘ecotype’[38], and extend our understanding of biogeochemical processes in subsurface environments.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References

This research was supported by the Subsurface Science Program, Office of Health and Environmental Research, U.S. Department of Energy (DOE). The continued support of Dr. F.J. Wobber is greatly appreciated. Pacific Northwest National Laboratory is operated for DOE by Battelle Memorial Institute under Contract DE-AC06-76 RLO 1830.


  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results
  6. 4Discussion
  7. Acknowledgements
  8. References
  • 1
    Balkwill, D.L., Fredrickson, J.K. and Thomas, J.M. (1989) Vertical and horizontal variations in the physiological diversity of the aerobic chemoheterotrophic bacterial microflora in deep southeast coastal plain subsurface sediments. Appl. Environ. Microbiol. 55, 10581065.
  • 2
    Fredrickson, J.K., Garland, T.R., Hicks, R.J., Thomas, J.M., Li, S.W. and McFadden, K.M. (1989) Lithotrophic and heterotrophic bacteria in deep subsurface sediments and their relation to sediment properties. Geomicrobiol. J. 7, 5366.
  • 3
    Fredrickson, J.K. et al. (1993) Microbiological characteristics of pristine and contaminated deep vadose sediments from an arid region. Geomicrobiol. J. 11, 95107.
  • 4
    Brockman, F.J., Kieft, T.L., Fredrickson, J.K., Bjornstad, B.N., Li, S.W., Spangenburg, W. and Long, P.E. (1992) Microbiology of vadose zone paleosols in south-central Washington state. Microb. Ecol. 23, 279301.
  • 5
    Fredrickson, J.K. et al. (1995) Microbial community structure and biogeochemistry of Miocene subsurface sediments: implications for long-term microbial survival. Mol. Ecol. 4, 619626.
  • 6
    Haldeman, D.L., Pitonzo, B.J., Story, S.P. and Amy, P.S. (1994) Comparison of the microbiota recovered from surface and deep subsurface rock, water, and soil along an elevational gradient. Geomicrobiol. J. 12, 99111.
  • 7
    Brockman, F.J., Ringelberg, D.B., White, D.C., Fredrickson, J.K., Balkwill, D.L., Kieft, T.L., Phelps, T.J. and Ghiorse, W.C. (1993) Estimates of intact but nonviable and viable but nonculturable microorganisms in subsurface sediments from six boreholes located in wet and dry climatic regions of the United States. 1993 International Symposium on Subsurface Microbiology, B-32.
  • 8
    Ward, D.M., Bateson, M.M., Weller, R. and Ruff-Roberts, A.L. (1992) Ribosomal RNA analysis of microorganisms as they occur in nature. Adv. Microb. Ecol. 12, 219286.
  • 9
    Pace, N.R., Stahl, D.A., Lane, D.J. and Olsen, G.J. (1986) The analysis of natural microbial populations by ribosomal RNA sequences. Adv. Microb. Ecol. 9, 155.
  • 10
    Fuhrman, J.A., McCallum, K. and Davis, A.A. (1992) Novel major archaebacterial group from marine plankton. Nature 356, 148149.
  • 11
    Barns, S.M., Fundyga, R.E., Jeffries, M.W. and Pace, N.R. (1994) Remarkable archaeal diversity detected in a Yellowstone National Park hot spring environment. Proc. Natl. Acad. Sci. USA 91, 16091613.
  • 12
    Liesack, W. and Stackebrandt, E. (1992) Occurrence of novel groups of the domain Bacteria as revealed by analysis of genetic material isolated from an Australian terrestrial environment. J. Bacteriol. 174, 50725078.
  • 13
    Liesack, W. and Stackebrandt, E. (1992) Unculturable microbes detected by molecular sequences and probes. Biodiversity Conservation 1, 250262.
  • 14
    Suzuki, M.T. and Giovannoni, S.J. (1996) Bias caused by template annealing in the amplification of mixtures of 16S rRNA genes by PCR. Appl. Environ. Microbiol. 62, 625630.
  • 15
    Reysenbach, A.-L., Giver, L.J., Wickham, G.S. and Pace, N.R. (1992) Differential amplification of rRNA genes by polymerase chain reaction. Appl. Environ. Microbiol. 58, 34173418.
  • 16
    Farrelly, V., Rainey, F.A. and Stackebrandt, E. (1995) Effect of genome size and rrn gene copy number on PCR amplification of 16S rRNA genes from a mixture of bacterial species. Appl. Environ. Microbiol. 61, 27982801.
  • 17
    Chandler, D.P., Brockman, F.J. and Fredrickson, J.K. (1997) Effect of PCR template concentration on the composition and distribution of total community 16S rDNA clone libraries. Mol. Ecol. 6, 475482.
  • 18
    Balkwill, D.L. (1993) DOE makes subsurface cultures available. ASM News 59, 504506.
  • 19
    Chandler, D.P., Brockman, F.J. and Fredrickson, J.K. (1997) Use of 16S rDNA clone libraries to study changes in a microbial community resulting from ex situ perturbation of a subsurface sediment. FEMS Microbiol. Rev. (in press).
  • 20
    Kieft, T.L., Fredrickson, J.K., McKinley, J.P., Bjornstad, B.N., Rawson, S.A., Phelps, T.J., Brockman, F.J. and Pfiffner, S.M. (1995) Microbiological comparisons within and across contiguous lacustrine, paleosol, and fluvial subsurface sediments. Appl. Environ. Microbiol. 61, 749757.
  • 21
    Balkwill, D.L. (1989) Numbers, diversity, and morphological characteristics of aerobic, chemoheterotrophic bacteria in deep subsurface sediments from a site in South Carolina. Geomicrobiol. J. 7, 3352.
  • 22
    Stevens, T.O. (1995) Optimization of media for enumeration and isolation of aerobic heterotrophic bacteria from the deep terrestrial subsurface. J. Microbiol. Methods 21, 293303.
  • 23
    Ghiorse, W.C. and Balkwill, D.L. (1983) Enumeration and morphological characterization of bacteria indigenous to subsurface environments. Dev. Ind. Microbiol. 24, 213223.
  • 24
    Sinclair, J.L. and Ghiorse, W.C. (1989) Distribution of aerobic bacteria, protozoa, algae, and fungi in deep subsurface sediments. Geomicrobiol. J. 7, 1532.
  • 25
    Balkwill, D.L. et al. (1997) Taxonomic study of aromatic-degrading bacteria from deep terrestrial subsurface sediments and description of Sphingomonas aromaticivorans sp. nov., Sphingomonas subterrae sp. nov., and Sphingomonas stygialis sp. nov. Int. J. Syst. Bacteriol. 47, 191201.
  • 26
    Maidak, B.L., Larsen, N., McCaughey, M.J., Overbeek, R., Olsen, G.J., Fogel, K., Blandy, J. and Woese, C.R. (1994) The Ribosomal Database Project. Nucleic Acids Res. 22, 34853487.
  • 27
    Brockman, F.J., Li, S.W., Fredrickson, J.K., Ringelberg, D.B., Kieft, T.L., Spadoni, C.M., McKinley, J.P. and White, D.C. (1996) Post-sampling changes in microbial community structure and activity in a subsurface paleosol. Microb. Ecol. (submitted).
  • 28
    Haldeman, D.L. and Amy, P.S. (1993) Diversity within a colony morphotype: implications for ecological research. Appl. Environ. Microbiol. 59, 933935.
  • 29
    Brunk, C.F., Avaniss-Aghajani, E. and Brunk, C.A. (1996) A computer analysis of primer and probe hybridization potential with bacterial small-subunit rRNA sequences. Appl. Environ. Microbiol. 62, 872879.
  • 30
    Clayton, R.A., Sutton, G., Hinkle, P.S.J., Bult, C. and Fields, C. (1995) Intraspecific variation in small-subunit rRNA sequences in GenBank: why single sequences may not adequately represent prokaryotic taxa. Int. J. Syst. Bacteriol. 45, 595599.
  • 31
    Mutter, G.L. and Boynton, K.A. (1995) PCR bias in amplification of androgen receptor alleles, a trinucleotide repeat marker used in clonality studies. Nucleic Acids Res. 23, 14111418.
  • 32
    Walsh, P.S., Erlich, H.A. and Higuchi, R. (1992) Preferential PCR amplification of alleles: mechanisms and solutions. PCR Methods Appl. 1, 241250.
  • 33
    Sarkar, G., Cassady, J., Bottema, C.D.K. and Sommer, S.S. (1990) Characterization of polymerase chain reaction amplification of specific alleles. Anal. Biochem. 186, 6468.
  • 34
    Chandler, D.P., Brockman, F.J. and Fredrickson, J.K. (1997) Molecular phylogeny of a deep subsurface microbial community. Microb. Ecol. (submitted).
  • 35
    Moyer, C.L., Dobbs, F.C. and Karl, D.M. (1994) Estimation of diversity and community structure through restriction fragment length polymorphism distribution analysis of bacterial 16S rRNA genes from a microbial mat at an active, hydrothermal vent system, Loihi Seamount, Hawaii. Appl. Environ. Microbiol. 60, 871879.
  • 36
    Ward, N., Rainey, F.A., Goebel, B. and Stackebrandt, E. (1995) In: Microbial Diversity and Ecosystem Function (Allsopp, D., Colwell, R.R. and Hawksworth, D.L., Eds.), pp. 89–110. CAB International.
  • 37
    Stackebrandt, E., Liesack, W. and Goebel, B.M. (1993) Bacterial diversity in a soil sample from a subtropical Australian environment as determined by 16S rDNA analysis. FASEB J. 7, 232236.
  • 38
    Stahl, D.A. (1995) Application of phylogenetically based hybridization probes to microbial ecology. Mol. Ecol. 4, 535542.