|Home | About | Journals | Submit | Contact Us | Français|
Isolation of multiple carbon monoxide (CO)-oxidizing Burkholderia strains and detection by culture-independent approaches suggest that Burkholderia may be an important component of CO-oxidizing communities in Hawaiian volcanic deposits. The absolute and relative abundance of the bacteria in these communities remains unknown, however. In this study, a quantitative PCR (Q-PCR) approach has been developed to enumerate Burkholderia coxL genes (large subunit of carbon monoxide dehydrogenase). This represents the first attempt to enumerate coxL genes from CO oxidizers in environmental samples. coxL copy numbers have been determined for samples from three sites representing a vegetation gradient on a 1959 volcanic deposit that included unvegetated cinders (bare), edges of vegetated sites (edge), and sites within tree stands (canopy). Q-PCR has also been used to estimate copy numbers of Betaproteobacteria 16S rRNA gene copy numbers and total Bacteria 16S rRNA. coxL genes could not be detected in the bare site (detection limit, ≥4.7 × 103 copies per reaction) but average 1.0 × 108 ± 2.4 × 107 and 8.6 × 108 ± 7.6 ×107 copies g−1 (dry weight) in edge and canopy sites, respectively, which differ statistically (P = 0.0007). Average Burkholderia coxL gene copy numbers, expressed as a percentage of total Bacteria 16S rRNA gene copy numbers, are 6.2 and 0.7% for the edge and canopy sites, respectively. Although the percentage of Burkholderia coxL is lower in the canopy site, significantly greater gene copy numbers demonstrate that absolute abundance of coxL increases in vegetated sites and contributes to the expansion of CO oxidizer communities during biological succession on volcanic deposits.
Aerobic carbon monoxide (CO)-oxidizing bacteria are widespread in soil environments and consume about 15% of annual CO emissions, indirectly impacting the tropospheric chemistry of greenhouse gases (5, 13). Recent enrichment and isolation efforts as well as genomic sequencing have revealed that CO oxidizers comprise a phylogenetically broad and metabolically diverse group of taxa including Firmicutes, Actinobacteria, Proteobacteria, and Chloroflexi (16, 31). Laboratory studies of CO-oxidizing members of these taxa have demonstrated that some are able to grow on CO as the sole carbon and energy source while most function preferentially as heterotrophs using CO when suitable substrates are lacking or are present at low concentrations (16).
Elucidation of the enzyme structure, function (5), and operon sequence (26) of aerobic carbon monoxide dehydrogenase (CODH) made it possible to develop PCR primers targeting a 1,260-bp fragment of the catalytic (large) subunit in the coxL gene (15). Cloning and sequencing of this gene fragment in young volcanic deposits have revealed that CO oxidizers are among the primary colonists on young unvegetated basalts and include bacteria spanning the currently known diversity of CO-oxidizing taxa as well as uncultured taxa (6, 18). As vegetation cover and organic carbon increase during biological succession on volcanic deposits, CO-oxidizing Proteobacteria become increasingly dominant and diverse, most likely due to increased organic matter availability for heterotrophic growth (6, 30).
A recent molecular ecological survey of CO oxidizers across a vegetation gradient on Kilauea volcano demonstrated that Betaproteobacteria coxL genes, in particular, were highly correlated with increasing vegetation (30). Total Proteobacteria comprised 2.6% of a coxL (large subunit of carbon monoxide dehydrogenase) clone library generated for unvegetated cinders (bare site) but 70 and 75% of libraries generated for transition (edge site) and vegetated sites (canopy), respectively. Although Alphaproteobacteria comprised the majority of the Proteobacteria at all sites, Betaproteobacteria coxL comprised 0, 1.7, and 32.9% of bare, edge, and canopy site coxL clone libraries, respectively. Several of the canopy coxL sequences were phylogenetically similar to sequences from Burkholderia xenovorans LB400 or Burkholderia strain PP52-1, an isolate previously obtained from the canopy site (29).
Additional efforts to enrich novel CO-oxidizing bacteria from these sites have yielded 12 isolates, the closest relatives (>97% 16S rRNA gene sequence similarity) of which include Burkholderia sacchari, Burkholderia unamae, Burkholderia nodosa, Burkholderia mimosarum, Burkholderia soli, Burkholderia bryophila and Burkholderia caledonica (29). These observations suggest that the capacity for CO oxidation may be widespread within the Burkholderia genus, particularly among plant-associated members, and that Burkholderia species may be important contributors to the expansion of CO-oxidizing communities during biological succession on volcanic deposits. However, the abundance of CO oxidizers in general and CO-oxidizing Burkholderia species in particular remains unknown.
We describe here a quantitative PCR (Q-PCR) approach to quantify Burkholderia coxL gene copy numbers and to compare them to Q-PCR-based estimates of Betaproteobacteria 16S rRNA and total Bacteria 16S rRNA gene abundance. This represents the first molecular ecological approach to enumerate coxL gene copy numbers. To date, estimates of CO oxidizer abundance have been based on most probable number (MPN) assays (2) or have been inferred from maximum potential CO uptake rates (17). Results indicate that copy numbers of 10 Burkholderia coxL genes increased significantly with increasing vegetation. Ratios of Burkholderia coxL to total 16S rRNA gene copy numbers were similar to or greater than analogous ratios reported for other functional genes (e.g., nifH and narG) in previous Q-PCR studies (10, 12).
The three sampling sites used in this study, bare, edge, and canopy, were located along a vegetation gradient on a 1959 volcanic deposit (Pu'u Puai) on Kilauea volcano (Hawaii). Chemical and physical characteristics of these sites as well as the diversity and structure of CO-oxidizing and total bacterial communities have been described previously (6, 9, 14, 17, 24, 30). The Pu'u Puai deposit consisted of centimeter-size basalt cinders that support patches or islands of woody vegetation (edge and canopy sites), including Meterosideros polymorpha and Morella faya, which are surrounded by unvegetated patches (bare sites). Canopy sites are located inside well-developed tree islands characterized by a distinct litter layer and an accumulation of organic-rich soil within the cinder matrix. Edge sites are located at the perimeter of tree islands and are also characterized by a litter layer and organic-rich soil within the cinder matrix. Bare sites are located about 5 m from the perimeter of tree islands; they lack a litter layer, rooted vegetation, and soil.
In August 2008, triplicate surface samples (0 to 2 cm) were collected aseptically from bare, edge, and canopy sites and placed into Whirlpak bags. Samples were transported to Louisiana State University on dry ice and stored at −80°C until DNA was extracted. Major roots and debris were removed from all samples prior to extraction; cinders in edge and bare samples were crushed using a sterile mortar and pestle. After crushed materials were subjected to three cycles of freezing and thawing at −80°C and 65°C, DNA was extracted from triplicate canopy samples (0.25 to 0.5 g of fresh weight [gfw]) using a MoBio Power Soil DNA Extraction Kit (MoBio Laboratories, Carlsbad, CA) and from bare and edge samples (5 to 7 g and 7 to 12 gfw) using a MoBio Power Max Soil DNA Extraction Kit (MoBio Laboratories, Carlsbad, CA). Parallel subsamples were used for determining water content. DNA concentrations in the extracts were determined using a Nanodrop spectrophotometer (Nanodrop, Wilmington, DE).
coxL sequences from Burkholderia xenovorans LB400 and three CO-oxidizing Burkholderia strains isolated from canopy soil (strains CP11, PP52-1 and PP51-3), as well as putative Burkholderia coxL sequences obtained from edge and canopy site clone libraries (30), were aligned using ClustalX, version 1.0.1 (28). Primaclade (http://www.umsl.edu/services/kellogg/primaclade.html) was used to identify primer sites that would amplify a 259-bp fragment of coxL (Table (Table1),1), which includes the CODH active-site motif. Primers were chosen to exclude form I coxL sequences from all taxa other than Burkholderia and to exclude form II putative coxL genes.
Primer specificity was examined using PCRs with DNA extracts from various CO-oxidizing Actinobacteria and Alphaproteobacteria and Betaproteobacteria, including Mycobacterium strain B8HB, Mycobacterium neoaurum, Mycobacterium parafortuitum, Mycobacterium marinum, Stappia aggregata, Stappia stellulata, Stappia strain BrT7, Stappia strain MIO, Stappia strain HI, and Stappia strain M4, Shinella strain FG1M5, Sulfitobacter strain P10, Mesorhizobium strain KP12W, Burkholderia sp. PO-04-17-38, and a series of Burkholderia strains isolated from the canopy (strains DNBP18, DNBP16, DNBP20, DNBP22, DNBP6-1, I2, I7, GA, WA, YA, B2of, PP51-2, PP52-1, and CP11) and other sites (strains EB-2 and Rim) (29). To further confirm the specificity of Burkholderia coxL Q-PCR primers, edge and canopy site DNA was used in PCR amplifications as described below.
PCR products from edge and canopy DNA extracts were cloned using a TOPO TA cloning kit (Invitrogen, Carlsbad, CA) (6, 30). A PureYield Plasmid Miniprep System (Promega, Madison, WI) was used to isolate plasmids from individual clones grown in 0.6 ml of Luria broth with kanamycin (50 μg ml−1). Inserts were sequenced bidirectionally from the plasmids using vector primers M13R and T7. Sequences were assembled, analyzed by BLAST, and then aligned and used in phylogenetic analyses to determine their affiliations (15). Q-PCR products were also cloned in a similar manner and screened to confirm specificity of the Q-PCR.
16S rRNA and coxL gene copy numbers were quantified using standard curves generated from a series of 10-fold dilutions of B. xenovorans LB400 DNA extracted as described previously (15). Possible DNA contamination from the kit reagents or processing was assessed with a culture-free extraction blank.
DNA concentrations in culture and culture-free extraction blanks were determined using a Nanodrop spectrophotometer (Nanodrop, Wilmington, DE). The number of 16S rRNA and coxL gene copies present in the DNA extract was determined as previously described (11) based on the average molecular mass for double-stranded DNA (660 Da), the known size of the B. xenovorans LB400 genome (9,731,138 bp) (3), and the 16S rRNA and coxL gene copy numbers in the genome (six and two, respectively) (3, 20, 22).
Standard curves were generated using five 10-fold dilutions of B. xenovorans LB400 DNA which ranged from 24 to 2.4 × 105 genome copies. Triplicate sets of standards and no template controls (NTC) were run in parallel with all samples. The lower detection limit was determined from the highest DNA dilution that consistently amplified at a threshold cycle (CT) lower than the NTC. Standard curves were plotted as CT versus log of the calculated gene copy number.
PCR conditions were optimized by amplifying genes from B. xenovorans LB400 DNA and from soil extracts using a variety of primer concentrations and annealing temperatures. All Q-PCRs were carried out in 96-well ABI Prism MicroAmp optical plates using 25-μl reaction mixtures and the primers listed in Table Table1.1. Burkholderia coxL was amplified in reaction mixtures containing the following components: 12.5 μl of Absolute QPCR SYBR green mix (Thermo Scientific, Epsom, Surrey, United Kingdom), 0.8 μl of each primer (0.32 μM final concentration of each primer), 1 μl of 6-carboxy-X-rhodamine (ROX) dye (80-fold dilution in sterile H2O; Thermo Scientific), and 0.5 to 4 ng of template DNA. Amplification was carried out in an ABI Prism 7000 Sequence Detection System (Applied Biosystems, Carlsbad, CA) using the following program: Taq polymerase activation at 95°C for 15 min, followed by 35 cycles of denaturation at 95°C (1 min), annealing at 60°C (30 s), and extension at 72°C (1 min). Betaproteobacteria 16S rRNA genes and total Bacteria 16S rRNA genes were amplified in the same manner using annealing temperatures of 60°C and 53°C (7) and final primer concentrations of 0.5 μM and 0.3 μM, respectively. Reactions were conducted in triplicate. Dissociation curve analysis and PCR product visualization using gel electrophoresis were carried out after each amplification to confirm reaction specificity. Gene copy numbers in soil DNA were determined based on the CT values and by calculating the log of the gene copy number using the standard curves previously described.
In cases where amplification was detected in the NTC, the baseline was manually adjusted above the NTC to ensure that fluorescence signal did not contribute to standard or sample CT values. For each gene quantified, all samples were run in two separate Q-PCR assays, at most, to minimize error due to run-to-run variability; in these cases, careful attention was paid to the behavior of the standard curves in each assay to ensure comparability of gene copy numbers. Samples were quantified only if they fell within the linear range of the standard curves. Gene copy numbers in cinder and soil extracts were expressed per ng of DNA and per gram of dry weight (gdw) of soil, which were estimated from the total amount of DNA extracted from a known soil dry mass.
The presence of PCR inhibitors in cinder and soil DNA extracts was examined by amending extracts exhibiting gene copy numbers ranging from 1 × 102 to 1 × 104 copies μl−1 with plasmid DNA containing either 16S rRNA or coxL genes at final concentrations of 1 × 106 copies μl−1. Gene copy numbers were also examined in 10-fold serial dilutions of DNA to determine if PCR inhibitors were present.
The Burkholderia coxL clone sequences from the edge and canopy libraries were deposited in the GenBank under accession numbers FJ713673 to FJ713692.
A 259-bp coxL fragment was successfully amplified from the Burkholderia strains used for primer design (B. xenovorans LB400 , Burkholderia strain CP11 , Burkholderia strain PP52-1 , and Burkholderia strain PP51-3 ), from 10 other CO-oxidizing Burkholderia strains isolated from canopy soil (Burkholderia strain DNBP18, Burkholderia strain DNBP16, Burkholderia strain DNBP20, Burkholderia strain DNBP22, Burkholderia strain DNBP6-1, Burkholderia strain I7, Burkholderia strain GA, Burkholderia strain WA, Burkholderia strain YA, and Burkholderia strain B2of) (29), a strain isolated from a passalid beetle hindgut (EB-2), and from a strain isolated from soil from Pico de Orizaba, Mexico (Burkholderia strain PO-04-17-38). The coxL fragment from one CO-oxidizing isolate, Burkholderia strain Rim, could not be amplified using primers F3 and R5 even though the larger coxL fragment (ca. 1,260 bp) could be amplified using previously designed primers (15). No PCR products were amplified from CO-oxidizing Actinobacteria or Proteobacteria other than Burkholderia. Dissociation curves from Q-PCRs using soil DNA extracts contained only one peak corresponding to the amplified fragment in the standards, which confirmed the reaction specificity. Similarly, visualization of PCR products by gel electrophoresis typically revealed the presence of a single band. Occasionally, a faint larger band (about 1,000 bp) could be seen in some replicates, but this did not impact CT values.
Twenty coxL clones from libraries generated from edge and canopy site DNA extracts were chosen arbitrarily and sequenced. All of the clone sequences were identified as form I coxL by analysis of translated nucleic acid sequences; these sequences were most closely related to other Burkholderia coxL genes by BLAST analysis and clustered phylogenetically with Burkholderia coxL genes (Fig. (Fig.1).1). Clones generated from the Q-PCR products had the same size inserts as the form I coxL-containing clones, demonstrating the specificity of the Q-PCR.
Standard curves generated from serial dilutions of B. xenovorans LB400 DNA for coxL, Betaproteobacteria 16S rRNA, and total Bacteria 16S rRNA gene fragments were linear from 4.7 × 103 to 4.7 × 106, 1.42 × 103 to 1.42 × 106, and 1.42 × 103 to 1.42 × 106 gene copies, respectively. The lower detection limits for coxL, Betaproteobacteria 16S rRNA, and total Bacteria 16S rRNA gene fragment assays were 4.7 × 103, 1.42 × 103, and 1.42 × 103 copies per reaction, respectively. Culture-free extraction blanks and NTC behaved similarly during amplification, indicating that there was no contamination in the culture-free extraction blanks. Amplification of serial dilutions of environmental DNA yielded the expected decrease in copy numbers when 0.5 to 10 ng was loaded in a PCR, indicating that there was no inhibition. Amplification efficiencies for all genes were similar, ranging from 1.6 to 1.7, allowing for comparisons of gene copy numbers.
Gene copy numbers for coxL, Betaproteobacteria 16S rRNA, and total Bacteria 16S rRNA increased across the vegetation gradient from the bare to canopy sites (Table (Table2).2). Burkholderia coxL gene numbers were below the detection limit in the bare site (4.7 × 103 copies per reaction), while coxL copy numbers for the edge and canopy sites ranged from 7.2 × 103 to 1.48 × 104 copies ng DNA−1 and did not differ significantly (t test, P = 0.913). When expressed per gdw of soil, however, edge and canopy coxL gene copy numbers differed significantly (t test, P = 0.001) and ranged from 7.58 × 107 to 1.47 × 108 and from 7.1 ×108 to 9.52 × 108, respectively (Table (Table22).
Copy numbers of Betaproteobacteria 16S rRNA genes in bare and canopy sites ranged from 1.47 × 102 to 7.04 × 102 and from 3.93 × 103 to 9.91 × 103 copies ng DNA−1, respectively, and differed statistically (analysis of variance [ANOVA], P = 0.001). The number of Betaproteobacteria 16S rRNA gene copies ng DNA−1 in the edge site ranged from 3.07 × 103 to 4.61 × 103 and was significantly greater than values for the bare site (ANOVA, P = 0.003) but did not differ significantly from canopy numbers (ANOVA, P = 0.150). Gene copy numbers per gdw of soil increased significantly from bare to canopy sites, ranging from 1.42 × 105 to 8.4 × 108 (ANOVA, P ≤ 0.001) (Table (Table22).
Copy numbers of 16S rRNA genes in the bare, edge, and canopy sites ranged from 3.4 × 104 to 1.97 × 106 copies ng DNA−1 and did not differ significantly (ANOVA, P > 0.108). In contrast, when expressed per gdw of soil, copy numbers ranged from 3.4 × 104 to 1.78 × 1011, with canopy copy numbers significantly greater than those of the bare (ANOVA, P = 0.001) and edge sites (ANOVA, P = 0.008).
coxL copy numbers corrected per gdw of soil for the edge and canopy sites did not differ significantly from the corresponding Betaproteobacteria 16S rRNA gene copy number (ANOVA, P > 0.19) but were significantly less than total Bacteria 16S rRNA gene copy numbers (ANOVA, P ≤ 0.001) (Table (Table22).
Average numbers of coxL genes (corrected per gdw of soil) expressed as a percentage of the total Bacteria 16S rRNA genes for the edge and canopy sites were 6.2 and 0.7%, respectively. Average numbers of Betaproteobacteria 16S rRNA genes (corrected per gdw of soil) expressed as a percentage of total Bacteria 16S rRNA genes for the bare, edge, and canopy sites were 0.1, 2.0, and 0.5%, respectively (Table (Table22).
Results from dissociation curve analyses, coxL assays from known organisms, and sequencing and phylogenetic analysis of coxL clones have shown that primers F3 and R5 specifically amplify form I coxL genes from Burkholderia isolates. Of 16 isolates tested, only Burkholderia strain Rim does not yield an amplification product. Subsequent assays have revealed that a larger coxL gene fragment (ca. 1,260 bp) clusters with sequences from Alphaproteobacteria, e.g., Bradyrhizobium (Fig. (Fig.1),1), which were excluded as targets for amplification. Therefore, the primers used in this study amplify coxL genes that can be defined phylogenetically as Burkholderia and provide a conservative index of CO-oxidizing Burkholderia coxL gene copy numbers when used in Q-PCR applications.
Results from Q-PCR assays in this study reveal increasing Burkholderia coxL gene copy numbers with increasing vegetation cover. Burkholderia coxL gene copy numbers in the canopy site significantly exceed numbers in the edge and bare sites when expressed per gdw of soil (Table (Table2)2) (t test, P < 0.001). This is consistent with results from clone libraries, which show that Betaproteobacteria coxL sequences account for a much higher percentage of coxL sequences in canopy than in other sites (30). Increases in abundance are also consistent with the well-known association of Burkholderia with rhizosphere soil (1).
Betaproteobacteria 16S rRNA gene copy numbers per gdw also increase (1,400-fold) across the vegetation gradient (Table (Table2).2). In addition, Betaproteobacteria 16S rRNA gene copy numbers corrected per ng of DNA are statistically greater for edge and canopy sites than for the bare site (Table (Table2).2). Collectively, these results suggest that Betaproteobacteria in general and Burkholderia specifically are important contributors to the expansion of CO oxidizer diversity as vascular plants develop on volcanic material.
Perhaps more important, all canopy site Betaproteobacteria 16S rRNA gene sequences from a prior clone library study have been classified as Burkholderia (30). This provides a clear indication of the relative abundance of the genus. Numbers of Burkholderia coxL and Betaproteobacteria 16S rRNA genes were also statistically equivalent at the edge and canopy sites, an additional indication that Burkholderia may dominate the Betaproteobacteria at these sites.
When expressed as a percentage of total Bacteria 16S rRNA gene copy numbers, Betaproteobacteria represent 2.0% and 1.5% of the edge and canopy sites, respectively. This is consistent with estimates from analyses of the incidence of Betaproteobacteria in 16S rRNA gene libraries from these sites. Consistency between clone libraries and Q-PCR results indicate that trends in these data sets are reliable.
Comparisons of gene copy numbers with cell numbers are constrained by the fact that copy numbers for coxL and 16S rRNA genes in individual bacterial genomes vary from 1 to 2 (16) and from 1 to 15 (8, 19), respectively. Comparisons of absolute gene numbers among studies are also limited by differences in DNA extraction methods and efficiencies (27). Expressing functional gene numbers as a percentage of total 16S rRNA gene copy numbers, however, facilitates estimates of changes in relative abundance and comparisons among studies. Average numbers of Burkholderia coxL gene copies in the edge and canopy sites expressed as a percentage of the total 16S rRNA gene copies were 6.2 and 0.7%, respectively, demonstrating that, even though absolute numbers of coxL may increase with increasing vegetation, CO-oxidizing Burkholderia species comprise a relatively smaller fraction of the total bacterial community in the canopy than the edge site.
The percentages estimated for both sites are similar to or higher than those reported for several functional genes in other soil ecosystems, indicating that the relative abundance of coxL may be similar to or higher than that of other functional genes. For instance, nirK and nosZ have been reported at levels from 0.1% to 6% of 16S rRNA gene copy numbers in agricultural, marsh, and Himalayan soils (10). A study of denitrification gene abundance across a successional gradient on a glacial foreland has reported similar percentages for narG, nirK, nirS, and nosZ (12). Because Burkholderia genes are only a fraction of the coxL gene pool, total coxL abundance must account for a somewhat larger percentage of 16S rRNA genes for edge and canopy sites than has been reported for other functional genes.
Assuming that CO-oxidizing Burkholderia species have 1 to 2 copies of the coxL gene, the average number of Burkholderia coxL genes in the canopy (8.6 × 108 copies gdw−1) corresponds to 4.3 × 108 to 8.6 × 108 organisms. In a previously generated coxL clone library for the canopy site, Betaproteobacteria comprised 32.9% of the sequence (30). If Burkholderia species are assumed to dominate this fraction, total CO oxidizer abundance extrapolates as 3.04 times the Burkholderia estimate, or 1.3 × 109 to 2.6 × 109 gdw−1. Since Burkholderia dominates but does not account for all of the Betaproteobacteria CO oxidizers and since the F3 and R5 primers may not amplify all Burkholderia coxL genes, this is a conservative estimate of canopy CO oxidizer community size.
Since bacteria contain 1 to 15 copies of the 16S rRNA gene per genome (8, 19), the total canopy 16S rRNA gene copy number per gdw of soil represents 8.0 × 109 to 1.2 × 1011 genomes or organisms (average, 6.4 × 1010). The average number of total CO oxidizers estimated above (1.95 × 109 organisms) represents about 3% of the average estimated number of total organisms, based on 16S rRNA gene copy numbers for the canopy site. This suggests that CO oxidizers occupy a significant fraction of the canopy bacterial community as a whole.
In summary, the absolute abundance of Burkholderia coxL genes increases with increasing vegetation, indicating that Burkholderia species contribute to CO oxidizer community expansion during biological succession. Extrapolations from numbers of Burkholderia coxL genes provide the first quantitative estimates of CO oxidizer abundances using molecular techniques. The results indicate that CO oxidizers are more abundant than previously estimated from MPN approaches. Further development of taxon-specific Q-PCR assays and group-specific estimates of activity will help link a newly emerging picture of extensive diversity among CO-oxidizing bacteria with their functional contributions in situ.
Published ahead of print on 5 February 2010.