|Home | About | Journals | Submit | Contact Us | Français|
Specific SYBR green-based quantitative-PCR assays targeting conserved regions in the 16S-23S rRNA internal transcribed spacer regions were developed for five subgroups of the environmentally abundant and biogeochemically active Roseobacter clade of marine bacteria. The assays were applied to field samples demonstrating their utility in investigations of abundant Roseobacter group phylotypes in the environment.
The Roseobacter clade (referred to as roseobacter hereafter) represents a phylogenetically diverse and biogeochemically relevant group of marine bacteria that has been the subject of several recent reviews (3, 6, 27). This large lineage is well defined by 16S rRNA gene phylogeny, and high abundances in diverse marine habitats have been revealed in gene surveys, as well as targeted approaches (e.g., fluorescence in situ hybridization). Roseobacter abundance tends to be highest in coastal environments where they can account for upward of 30% of bacterioplankton communities, but its members also contribute significantly (~10%) to open ocean microbial assemblages (6, 27). Recently, investigators have been able to delineate several subgroups within this clade, representing the most prevalent sequences in environmental inventories (see, for example, references 6, 10, and 22). Since many of these clusters lack cultured isolates, investigations of specific subgroup abundances and distributions would facilitate our understanding of their ecological roles. Furthermore, studies of cultivated representatives have revealed genomic and metabolic versatility that is expected to be reflected in the majority of lineage members and suggest discrete subgroups likely occupy distinct niches in the environment (16). In fact, evidence is building that distributions of some subgroups may be correlated with environmental conditions (see, for example, references 15, 21, and 22). Quantitative tools to monitor specific subpopulations are therefore needed for detailed studies of roseobacter dynamics over temporal and spatial scales.
We describe here the development and testing of SYBR green-based real-time quantitative-PCR (qPCR) assays for five different subgroups of the clade. These assays target the 16S-23S rRNA internal transcribed spacer (ITS), a region that has been used to discriminate between closely related strains of bacteria and therefore is an ideal target for quantitative tools (see, for example, references 1, 12, and 20). Collectively, we show that these qPCR tools are appropriate for marine plankton samples and will be useful in future investigations of the distributions and dynamics of subgroups of this important clade. As proof of concept, these assays, along with a previously published 16S rRNA gene assay (25), were applied to a collection of water samples from the Chesapeake Bay, a region known to harbor diverse and dynamic roseobacter lineages (10, 11).
To aid in the design of quantitative PCR primer sets for roseobacter subgroups, partial rrn operons (hereafter referred to as ribosomal and intergenic spacer [RIS] regions) were amplified from DNA collected from Station RM6 (37°06′N, 75°43′W) in the Chesapeake Bay in September 2005. PCR was carried out using the Roseo536MSF (5′-CGGAGGGGGTTAGCGTTGT-3′) and 23Sr (5′-GGGTTBCCCCATTCRG-3′) primers. The Roseo536MSF primer was adapted from primer ROSEO536R (4) to be used as a forward primer. The 23Sr is a general bacterial primer that targets the 23S rRNA gene of most known bacteria (8). Amplification was conducted using Platinum Taq DNA polymerase high fidelity (Invitrogen, Carlsbad, CA). After amplification the products were cloned with the TOPO TA cloning kit for sequencing (Invitrogen) and subject to ITS-length heterogeneity PCR (ITS-LH-PCR ) to identify unique roseobacter clones. Sixteen clones were sequenced in their entirety using M13F, M13R, 1390R (5′-CCC ATC ATN ARI ATN GT-3′), 1406F (5′-ACG GGC GGT GTG TRC AA-3′) (13), and the novel primers tRNArosF (5′-CTG GGA GAG CGC CTG-3′) and tRNArosR (5′-CAG GCG CTC TCC CAG-3′) targeting the tRNA-alanine sequences in the ITSs of roseobacters. Although small, this clone library augments the previously existing databases and was sufficient for the design of qPCR primers to groups of interest.
Nearly 1,000 bp of the 3′ end of the 16S rRNA gene sequence was obtained for each of the ITS clones, facilitating phylogenetic identification. Homology searches using this region reveal all clones belong to seven roseobacter subgroups that are poorly represented by cultured isolates (Fig. (Fig.1)1) (6, 10). These sequences have been submitted to GenBank with the accession numbers GQ342302 to GQ342317. qPCR assays were developed for five of these subgroups. The DC-5-80 subgroup (also known as the RCA for roseobacter clade-associated cluster) is one of the 20 most abundant phylotypes in marine surface waters and is likely the most cosmopolitan of the roseobacter clusters identified to date (6, 15, 21, 26). A systematic global survey of this subgroup revealed that they are found worldwide in temperate and polar waters and can comprise upwards of 10% of coastal bacterial assemblages (21). Recent evidence suggests group members are most abundant under eutrophic conditions, and may comprise one of the most successful marine copiotrophic bacterial lineages (15). The CHAB-I-5 subgroup is also abundant in coastal waters but appears in open ocean metagenomic libraries as well (6). In surveys of coastal California bacterioplankton and Sargasso Sea surface waters, CHAB-I-5 group members represented ~20% of the roseobacter 16S rRNA gene containing clones (17, 22, 26). The ChesI-A, -B, and -C groups are common in Chesapeake Bay waters, where they show a seasonal distribution (10). BlastN searches indicate the GOS survey recovered sequences from all ChesI subgroups from tropical and temperate coastal and open ocean sites (26). Finally, a novel subgroup designated M0-Ar2-P4F09 has not commonly been retrieved in 16S rRNA and metagenomic surveys. This subgroup likely represents a lineage with relatively limited distribution and abundance. qPCR assays for this subgroup will help to elucidate the physiochemical conditions leading to this apparent geographic constraint.
For phylogenetic placement, partial 16S rRNA sequences were first imported into an ARB (14) database based on ssujun02.arb containing ca. 29,000 sequences, aligned and added by ARB_PARSIMONY to an updated version of a previously described full-length roseobacter tree (10). Based on this placement, 16S rRNA-ITS-23S rRNA sequences were exported and combined with published ITS-containing sequences from previous studies (5, 22, 23, 26) and one unpublished sequence associated with a fosmid library from Monterey Bay (22). These sequences were aligned by using CLUSTAL W and imported into a new ARB database. Inspection of the alignment revealed obvious sequence differences that were used to design specific primers for the five subgroups, with the aid of the sequence matching functions of ARB_EDIT4. Criteria for primer design included (i) a targeted length of the amplicon that was ca. 150 bp when possible and always less than 500 bp, (ii) a target primer Tm of 59°C at 500 nM and estimated with the Applied Biosystems online calculator (http://www.appliedbiosystems.com/support/techtools/calc/), and (iii) mismatches to nontargets were designed toward the 3′ end of the primer.
Primer sets specific for the five selected roseobacter subgroups target unique sequences in the ITS region (Table (Table1).1). In all cases, these primers are located exterior to both tRNA coding regions and generate an amplicon of 209 to 311 bp. As a first approximation of primer specificity, each primer set was tested against all RIS clones, as well as representative roseobacter strains by using endpoint PCR. No cross-reactivity was detected in these assays. Annealing temperatures were initially determined empirically by gradient endpoint PCR analysis (50 to 70°C) of each primer set with standards containing the targeted sequence (purified plasmids from the clones shown in Table Table1).1). Optimal annealing temperatures were further refined following primer concentration optimization in qPCR assay conditions. To identify the most appropriate primer concentrations, a matrix of various concentrations (100, 500, 1,000, and 1,500 nM) of forward and reverse primers was tested with 105 copies of the corresponding standard. The primer concentration combinations yielding low threshold cycle number (CT) values were selected as most appropriate for use in assays. Optimal primer concentrations varied significantly between primers used in the same assay (up to 2-fold), as well across those used in all assays (10-fold) (Table (Table1).1). The high primer concentration required for the previously published 16S rRNA gene assay (25) is likely due to differences in the melting temperature of these primers, which are constrained in terms of length to be universal. The roseobacter ITS primers are more specific in nature, and concentration variation may be reflective of unpredictable interactions between primers under the defined assays conditions. Similar discrepancies in primer concentrations have been reported previously (25) and highlight the importance of empirically determining the primer concentrations for real-time assays that yield the highest amplification efficiencies.
qPCR was carried out by using a DNA Engine Opticon 2 real-time PCR detector with the Opticon Monitor 3.1.32 software package (Bio-Rad Laboratories, Inc., Hercules, CA). qPCR reactions were performed in a 25-μl volume with 1× SYBR Premix Ex Taq cocktail (Perfect Real Time; Takara Bio, Inc., Shiga, Japan), forward and reverse primers at the concentrations listed in Table Table1,1, and 2.5 μl of template DNA. The amplification programs for each primer set are shown in Table Table1.1. Fluorescence measurements were conducted at the end of each cycle at 82°C through channel 1 (523 to 543 nm). Standards were developed from plasmids containing cloned RIS sequences (Table (Table1)1) and 10-fold serial dilutions of these samples in molecular-grade water were used in assays. Standard curves were determined as the correlation between the log of gene copy numbers and the CT. In all cases, correlation coefficients for standard curves assays were greater than 0.98.
The lower limits of detection for each assay were set by determining the limit at which DNA standards were reliably amplified and ranged from 25 to 250 copies per reaction volume (25 μl) (Table (Table1).1). Melting curves were generated after each assay to verify the specificity of the amplification by heating from 50 to 95°C at a rate of 0.2°C s−1 and taking fluorescence measurements every 1.0°C. These melting curves suggested that minor nonspecific amplification may occur when the ChesI-B qPCR assays are applied to planktonic samples. Furthermore, at lower copy numbers, primer dimers were evident in some melting curves. In order to avoid over- or underestimations of gene copy numbers, fluorescence readings were taken at 82°C, a temperature several degrees above the melting point of nonspecific double-stranded DNAs and yet several degrees below the melting temperature of the desired amplicons (Table (Table1).1). This approach ensures that nonspecific products and primer dimers are denatured and thus not bound by the SYBR green reporter dye (18). Theoretical amplification efficiencies were calculated from the slope of the standard curve by using the equation “efficiency = 10−1/slope − 1” (19) and were found to be high (Table (Table1).1). This indicates amplification of target products was not significantly compromised by these primer dimers. Finally, in order to conclusively demonstrate that the assays are amplifying the intended products, 10 randomly selected amplicons from each of the five primer sets applied to DNA extracted from Chesapeake Bay were sequenced. In all cases, homology searches verified clones from all amplifications were the correct roseobacter ITS regions.
The Roseobacter subgroup qPCR assays were applied to a series of water samples collected from 16 to 18 July 2007 that represent four stations and multiple depths along a transect from the middle to the mouth of the Chesapeake Bay (Table (Table2).2). Water samples (250 to 500 ml) were prefiltered through a 47-mm Whatman GF/A (1.6-μm nominal retention; Schleicher & Schuell, Keene, NH), and the microbial biomass was ultimately collected on 25-mm 0.2-μm-pore-size Supor filters (Pall Corp., East Hills, NY). The filters were immediately placed in 250 μl of lysis buffer (2 mM sodium EDTA, 20 mM Tris-HCl, 1.2% [vol/vol] Triton X-100 [pH 8.0]) and frozen. DNA extractions from the filters were carried out according to a slightly modified protocol for the DNeasy blood and tissue kit (Qiagen, Inc., Valencia, CA) (24). Briefly, the filter-containing tubes were thawed, and 20 mg of lysozyme was added (from a 72-mg/ml stock). Tubes were briefly vortexed to dislodge bacterial cells and then incubated for 1 h at 37°C. Subsequent steps followed the manufacturer's protocol. Each DNA sample was analyzed in duplicate reactions of at least two dilutions ranging from 0.5× to 0.05× in molecular-grade water. To each 25 μl of qPCR mixture, 2.5 μl of diluted DNA (final concentrations of 0.75 to 5 ng) was added, and qPCR was performed as described above. For each run, a standard curve was determined by analyzing a dilution series (101 to 105 gene copies per 25-μl reaction mixture) of the appropriate standard in duplicate. For each measurement, a standard deviation of a minimum of two duplicates was determined.
In addition to the Roseobacter-specific primers, general bacterial 16S rRNA primers (25) were also applied to the field samples for comparative purposes. In order to provide the data in a relevant frame of reference, subgroup abundances were expressed as relative contributions of the total bacterial 16S rRNA gene pool. A caveat to this measurement is that general bacterial 16S rRNA gene PCR primers fail to amplify all recognized bacterial targets with equal efficiencies (see reference 2 and references within). However, the presumption is that that any amplification bias would be relatively consistent across the samples analyzed, making between sample comparisons valid. Another well-documented issue is that of multiple rrn operons and their confounding effects on abundance estimates (7). The rrn copy number is not known for the Roseobacter subgroups for which assays were developed but ranges between one and five in organisms for which complete genome sequences are available (16). However, it is unlikely that a large variation in rrn copy exists among members of a given subgroup. Thus, the value of these assays in determining biogeographical patterns of specific subgroups is not compromised.
Application of the qPCR assays to field samples revealed distinct spatial variation of the five subgroups along the transect. With the exception of CHAB-I-5 members, Roseobacter subgroup abundances decreased with increasing distance from the mouth of the bay (Table (Table2).2). This is consistent with the earliest report identifying the roseobacter clade (9) and multiple subsequent investigations that indicate roseobacter abundances are well correlated with salinity (6). The relative contributions of each subgroup to the bacterial 16S rRNA gene pool varied along this transect. The DC-5-80 and ChesI-C groups were detected in every sample analyzed and were the most abundant groups assayed. DC-5-80 comprised 8% of the bacterial 16S rRNA genes in samples from the mouth of the bay, but their relative contributions fell to <0.1% by mid-bay. A similar trend was seen with the ChesI-C group, although their relative contributions were often an order of magnitude lower than those for DC-5-80 group members (Fig. (Fig.22 and Table Table2).2). DC-5-80 abundance was positively correlated with salinity (Pearson correlation: r = 0.95, P < 0.01) and generally increased with depth (Fig. (Fig.2).2). An exception was found at the lower bay station, where there is evidence of three distinct water strata and DC-5-80 relative abundances do not show an obvious trend with depth (Fig. (Fig.2B).2B). A similar correlation with salinity, although positive, was not significant for the ChesI-C group. Correlations between abundance of these two subgroups with one another and with temperature were not significant.
The CHAB-I-5 subgroup displayed relatively low and sporadic abundances at all stations. Members were detectable in surface waters at all four stations and at mid-water depths at stations 724 and CB61. The ChesI-B subgroup was only detected in the lower bay stations. Overall, the distributions of ChesI-B and -C were in agreement with a previous report showing that the ChesI clade was retrieved in 16S rRNA clone libraries from Chesapeake Bay in the fall, where average salinities were higher than in late winter (10). The M0-Ar2-P4F09 group was only found in one sample, from surface waters just outside the mouth of the bay, and was present in relatively low numbers. Application of the qPCR assay for this group to another surface water sample collected at this station in August 2005 yielded similarly low abundances (data not shown) and indicates that these organisms may be numerically minor representatives of roseobacters in this system during late summer and fall.
The Roseobacter lineage is a large and heterogeneous collection of marine bacteria that are globally distributed. The prevalence of specific phylogenetic subgroups within the clade is evident from multiple, independent inventories of diverse marine bacterioplankton communities. This raises questions regarding the biogeographical and temporal patterns of these clusters. The development of qPCR assays for five roseobacter subgroups provides a series of rapid response tools that will be useful for describing roseobacter dynamics at greater resolution than is typically possible with sequence-based inventories.
We sincerely thank the crews of the R/V Hugh R. Sharp and R/V Cape Henlopen for aid in sample collection. We thank Robert Mason and Cynthia Gilmour for inviting us to participate in the 2005 Cape Helopen cruise. We also thank Christina Preston and Edward DeLong for the sequence of clone EB08011F12 and Gary LeCleir for providing constructive comments on the manuscript.
This study was funded by the National Science Foundation (OCE-0550485 to A.B. and OCE-0550547 to M.T.S.).
Published ahead of print on 2 October 2009.