Matching phylogeny and metabolism in the
uncultured marine bacteria, one cell at a time
Ramunas Stepanauskas* and Michael E. Sieracki
Bigelow Laboratory for Ocean Sciences, P.O. Box 475, West Boothbay Harbor, ME 04575-0475
Edited by David M. Karl, University of Hawaii, Honolulu, HI, and approved April 11, 2007 (received for review January 18, 2007)
The identification of predominant microbial taxa with specific
metabolic capabilities remains one the biggest challenges in envi-
ronmental microbiology, because of the limits of current met-
agenomic and cell culturing methods. We report results from the
direct analysis of multiple genes in individual marine bacteria cells,
demonstrating the potential for high-throughput metabolic as-
signment of yet-uncultured taxa. The protocol uses high-speed
fluorescence-activated cell sorting, whole-genome multiple dis-
placement amplification (MDA), and subsequent PCR screening. A
pilot library of 11 single amplified genomes (SAGs) was con-
structed from Gulf of Maine bacterioplankton as proof of concept.
The library consisted of five flavobacteria, one sphingobacterium,
four alphaproteobacteria, and one gammaproteobacterium. Most
of the SAGs, apart from alphaproteobacteria, were phylogeneti-
rRNA gene sequence. Thus, single-cell MDA provided access to the
genomic material of numerically dominant but yet-uncultured
taxonomic groups. Two of five flavobacteria in the SAG library
contained proteorhodopsin genes, suggesting that flavobacteria
are among the major carriers of this photometabolic system. The
pufM and nasA genes were detected in some 100-cell MDA prod-
ucts but not in SAGs, demonstrating that organisms containing
bacteriochlorophyll and assimilative nitrate reductase constituted
<1% of the sampled bacterioplankton. Compared with met-
agenomics, the power of our approach lies in the ability to detect
the metabolic and phylogenetic markers are located far apart on
Flavobacteria ? flow cytometry ? multiple displacement amplification ?
proteorhodopsin ? single-cell genomics
unknown phylogenetic and metabolic diversity of prokaryotes
(1–9). Although yet-uncultured taxa are believed to comprise
?99% of all prokaryotes, their metabolic capabilities and eco-
logical functions remain enigmatic, largely because of method-
ological limitations. For example, PCR-based clone libraries are
intrinsically limited to the analysis of one gene at a time, with no
direct way of linking libraries of diverse genes. Large-scale
environmental shotgun sequencing, although extremely produc-
tive for finding novel genes, is prohibitively expensive, and so far
is limited to only partial genome assembly of the most numer-
ically dominant taxa in complex marine microbial communities
bacteria (2). However, large insert-based function assignment is
limited to situations where the metabolic gene of interest is
located near phylogenetic markers (e.g., ribosomal genes). Thus,
suited for identification of microorganisms with specific meta-
bolic characteristics. This significantly limits the progress in such
diverse fields as biogeochemistry, microbial ecology and evolu-
tion, and bioprospecting.
We propose sequencing multiple DNA loci in individual
bacterial cells rather than environmental DNA extracts, as a
he PCR- and direct cloning-based sequencing of environ-
mental DNA extracts has revealed the enormous previously
more productive alternative for metabolic mapping of uncul-
tured microorganisms. This strategy has been gaining momen-
tum, with recent implementations of single-cell multiplex PCR
in termite gut microbiota by Ottesen et al. (10) and partial
genome sequencing of single cells of Prochlorococcus by Zhang
et al. (11). Here we show important improvements in single-cell
separation and DNA analysis protocols and demonstrate a
proof-of-concept metabolic mapping of taxonomically diverse
One of the main challenges for single-cell studies is efficient
and contamination-free separation of individual cells from other
microorganisms and extracellular DNA in an environmental
sample. Prior studies of DNA in individual prokaryote cells used
serial sample dilution (11), dilution using microfluidics (10), or
FACS. Compared with alternative methods, FACS offers several
critical advantages, including high-throughput rates and the
ability to sort targeted plankton groups, based on cell size and
fluorescence signals of natural cell components and fluoro-
chromes (13). Furthermore, cell separation by FACS creates
microsamples containing the target cell and only 3–10 pl of
sample around it (13). This reduces the codeposition of extra-
cellular DNA, which in marine waters occurs at concentrations
similar to cell-bound DNA (14, 15).
Current sequencing technologies require nanogram-to-
microgram DNA templates and are not capable of direct se-
quencing of individual DNA molecules. Thus, DNA preampli-
fication is necessary to sequence genes or genomes from
individual cells. For the analysis of up to two loci per cell, single
cell multiplex PCR has been used in medical research since the
1980s (16) and was recently used in an environmental microbi-
ology study (10). As a more versatile alternative, allowing for
analysis of an unlimited number of loci, several methods have
been suggested for whole-genome amplification, including de-
generated oligonucleotide-primed PCR, primer extension pre-
amplification, ligation-mediated PCR, and multiple displace-
ment amplification (MDA) using phi29 or Bst DNA polymerases
(17, 18). Among them, phi29-based MDA appears the most
suitable for efficient whole-genome amplification with low error
and bias (17, 18) and is capable of generating micrograms of
genomic DNA from nanogram-sized samples (19–21). Recently,
Phi29-based MDA was used on single human (18, 22, 23),
Escherichia coli (24), and Prochlorococcus (11) cells. Direct
Author contributions: R.S. and M.E.S. designed research; R.S. and M.E.S. performed re-
search; R.S. analyzed data; and R.S. and M.E.S. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Abbreviations: MDA, multiple displacement amplification; SAG, single amplified genome;
SSU rRNA, small-subunit rRNA; HNA, high nucleic acid; T-RFLP, terminal restriction frag-
ment length polymorphism.
database (accession nos. EF202334–EF202347 and EF508145–EF508148).
*To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2007 by The National Academy of Sciences of the USA
May 22, 2007 ?
vol. 104 ?
cloning and whole-genome sequencing of single-cell MDA prod-
ucts remains technically challenging because of template frag-
mentation, amplification biases, self-primed DNA synthesis, and
the multibranched nature of MDA products (11, 25). However,
downstream PCR appears insensitive to these MDA artifacts
(24), is simple and inexpensive, and can be effectively used in the
analysis of multiple genes.
The aim of this study was a proof-of-concept metabolic
mapping of marine bacterioplankton. We developed procedures
for clean single-cell separation by FACS, cell lysis, and subse-
quent single-cell whole-genome MDA for downstream PCR-
based analyses of multiple loci. We successfully constructed a
pilot library of single amplified genomes (SAGs) from a tem-
perate coastal marine site. We analyzed them for the presence
and DNA sequences of genes representing phylogenetic markers
[small-subunit rRNA (SSU rRNA)] and several significant bio-
geochemical functions in marine ecosystems (proteorhodopsin,
bacteriochlorophyll, nitrogenase, and assimilatory nitrate reduc-
tase). Bacterial proteorhodopsins (2) and bacteriochlorophylls
(26) are photometabolic systems, recently recognized for their
ubiquity and likely significance in the global carbon and energy
fluxes. Nitrogenase is a key enzyme in the fixation of N2,
effectively controlling primary production in vast areas of the
ocean, and appears to be possessed by some heterotrophic
bacterioplankton (27). Assimilatory nitrate reductase enables
some heterotrophic bacteria to use nitrate and in this way
compete with phytoplankton for the upwelled nitrogen (28). So
far, little is known about the taxonomic composition of micro-
organisms carrying these genes in marine environments.
Results and Discussion
Taxonomic Composition of the SAG Library. The SSU rRNA gene
was successfully PCR-amplified and sequenced from 12 of 48
single-cell MDA reactions (Table 1). The SAG MS024–2C was
identified as a contaminant and excluded from further analyses
[see supporting information (SI) Text]. The remaining SAG
library consisted of five flavobacteria, one sphingobacterium,
four alphaproteobacteria of the Roseobacter lineage, and one
gammaproteobacterium, all most closely related to marine iso-
lates and clones. Diverse representatives of the Roseobacter
lineage are readily isolated and are relatively well studied (29,
30). Accordingly, SSU rRNA genes of the four alphaproteobac-
terial SAGs were 99% identical to existing isolates. In contrast,
all flavobacterial, sphingobacterial, and gammaproteobacterial
SAGs were phylogenetically distant from established cultures,
with 88–97% identities in the SSU rRNA gene. Flavobacteria as
a group are proficient degraders of complex biopolymers, in-
cluding cellulose, chitin, and pectin (31). Thus, certain Flavobac-
teria taxa may play important and specialized roles in microbial
food webs and may be attractive for bioprospecting. Single-cell
MDA provided access to the unique genomic material of these
yet-uncultured taxa at the individual organism level. Different
from single-cell multiplex PCR (10, 16), which enables analysis
of high molecular-weight whole-genome amplification products.
This material can be used in a virtually unlimited number of
downstream PCRs (see below) and hybridization analyses and
may be suitable for genomic sequencing (11).
Table 1. Phylogeny of bacterial SSU rRNA genes obtained from single amplified genomes
protocol Genus* Closest isolate†
Kordia, 26 Flavobacterium sp. 3034
Flavobacterium sp. 3034
Cellulophaga sp. CC12
Sponge bacterium Zo9
Ulvibacter litoralis AY243096,
Clone NorSea37 AM279169, 9690 283
Kordia, 36 Clone NorSea43 AM279191, 9994 No cut
Cellulophaga, 80Clone 1D10 AY274838, 99 9632
Tenacibaculum, 98Clone WLB13–197 DQ015841,
Clone PB1.23 DQ071072, 99
Ulvibacter, 99 94284
Heliscomenobacter, 55Saprospiraceae bacterium
MS-Wolf2-H AJ786323, 88
Roseobacter sp.AY167254, 99
symbiont U63548, 99
O. aculeata symbiont
KOPRI 13313 DQ167247, 97
Clone F3C24 AY794157, 100
Rhodobacteraceae bact. 183
HTCC2120 AY386340, 90
Clone Ant4D3 DQ295237, 99575413
*Determined by RDP Classifier. Both type and nontype ?good? ?1,200-bp sequences were used. Numbers indicate confidence level (%) for genus identification.
†Determined by RDP Seqmatch. Provided are sequence ID, GenBank accession no., and sequence identity to the SAG (%).
‡Determined by National Center for Biotechnology Information BLAST-N. Provided are sequence ID, GenBank accession no., and sequence identity to the
Stepanauskas and SierackiPNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?
Only high nucleic acid (HNA) bacterioplankton was analyzed
in this study, which comprised 57% of all heterotrophic pro-
karyotes in the sample. This may have biased the taxonomic
composition of the SAG library, likely explaining the lack of
predominance of Bacteroidetes in the SAG library was unex-
pected. Alphaproteobacteria typically dominate SSU rRNA
PCR clone libraries of marine surface bacterioplankton, whereas
Bacteroidetes constitute ?3% of all marine clones (33). In
contrast, studies employing fluorescent in situ hybridization (34,
35), quantitative PCR (36), and metagenomics (5, 8, 9) suggest
a higher proportion of Bacteroidetes (particularly Flavobacte-
ria), in some cases ?70% of the total bacterioplankton (31). This
contradiction may be caused by PCR and/or cloning biases
against Flavobacteria (31). Interestingly, the ratio of Alphapro-
teobacteria vs. Bacteroidetes in our SAG library was 0.7 (Table
1), whereas the corresponding ratio of community terminal
restriction fragment length polymorphism (T-RFLP) analyses
peak areas at 55 bp (assumed Roseobacter lineage) and at 90–96
bp (assumed Bacteroidetes) was 1.0 (100 cell MDA-PCR) and
3.5 (1,000-cell seminested PCR) (SI Fig. 5). This discrepancy is
supportive of a PCR bias against Bacteroidetes, which would
affect T-RFLP profiles, especially those based on two rounds of
seminested PCR. On the other hand, the construction of the
SAG library was insensitive to PCR biases and did not involve
cloning. Thus, scaled-up SAG libraries may become an ultimate
tool for quantitative bacterioplankton analyses at high phyloge-
netic resolution. The advantage of SAG screening over fluores-
cent in situ hybridization, another taxon-specific quantification
method, is demonstrated by the extraction of high-resolution
phylogenetic information through sequencing of the entire SSU
rRNA gene, as well as protein-encoding loci (see below).
Proteorhodopsin genes were detected by PCR and confirmed
by sequence analysis in 2 of 11 SAGs (Fig. 1). In addition,
MDA reactions. Accordingly, Sabehi et al. (37) estimate that
13% bacterioplankton in the photic zone of the Mediterranean
Sea and the Red Sea carried proteorhodopsin genes. Our study
provides further evidence that proteorhodopsin-containing
microorganisms comprise a significant fraction of marine
Interestingly, both proteorhodopsin-positive SAGs were Fla-
vobacteria, providing evidence that proteorhodopsins are com-
mon in the numerically abundant representatives of this taxo-
nomic group (Fig. 1). The presence and photometabolic
functionality of proteorhodopsins in Flavobacteria were recently
confirmed by genome sequencing of four isolates (38). The first
indication of proteorhodopsins in Flavobacteria was obtained
from shotgun sequencing of Sargasso Sea microbes, where a
proteorhodopsin gene was found on a scaffold also containing a
DNA-directed RNA polymerase sigma subunit (rpoD) typical of
Bacteroidetes (5). Bacterial proteorhodopsins were first discov-
ered by screening environmental BAC libraries (2). Using this
technique, several Gammaproteobacteria, Alphaproteobacteria,
and Euryarchaea were identified as proteorhodopsin hosts (39,
of proteorhodopsins in Flavobacteria, possibly because the pro-
teorhodopsin and SSU rRNA genes are too far apart. Studies
based on community proteomics (41), community PCR (42, 43),
community shotgun sequencing (5, 9), and PCR screening of
metagenomic BAC libraries (39) demonstrated high diversity
of proteorhodopsins in the ocean, although the vast majority of
their hosts remain unknown. So far, only five marine isolates
have been reported to contain proteorhodopsin genes, including
alphaproteobacterium Pelagibacter ubique (44) and four Fla-
vobacteria (38). Here we demonstrate how single-cell MDA-
PCR can provide a powerful and relatively inexpensive tool for
the phylogenetic mapping of this biogeochemically important
gene, independent of the gene’s position on the chromosome or
The two SAG proteorhodopsins were most closely related (up
to 71% identity) to proteorhodopsins from four Flavobacteria
isolates and to a group of environmental clones from the North
Atlantic (Fig. 1). Consistent with the flavobacterial isolates and
near-surface environmental sequences, both SAGs had methi-
onine at amino acid position 105 (eBAC31A08 numbering),
indicative of absorption maxima near 530 nm (green light) (38).
In general, phylogenetic relationships among proteorhodopsins
and SSU rRNA genes mirrored each other, providing no evi-
dence for recent cross-taxa horizontal transfer events like those
observed in Archaea (40). On the other hand, the presence of
proteorhodopsin genes was inconsistent among some closely
related Flavobacteria, e.g., Polaribacter filamentus 215 and P.
irgensii 23-P (Fig. 1), suggesting recent proteorhodopsin gene
losses. Interestingly, proteorhodopsin genes closely related to
Flavobacterial SAGs and isolates were present among environ-
mental clones from the North Atlantic, Mediterranean Sea, and
Red Sea, indicating that Flavobacteria may be major carriers of
proteorhodopsin genes in diverse marine environments.
Other Genes. The pufM and nasA were not detected in any of the
single-cell MDA products. However, they were present in six
clone AAT38636, North Atlantic, 0 m
clone AAT39130, North Atlantic, 0 m
clone AAT389128, North Atlantic, 40 m
clone AAT39129, North Atlantic, 40 m
clone AAT38634, North Atlantic, 40 m
clone AAT38635, North Atlantic, 40 m
clone AAT38630, North Atlantic, 40 m
clone AAT38633, North Atlantic, 0 m
clone AAT38637, North Atlantic, 0 m
clone AAT38631, North Atlantic, 40 m
clone AAT38632, North Atlantic, 0 m
Tenacibaculum MED152, Mediterranean Sea, 1 m
Dokdonia MED134, Mediterranean Sea, 1 m
P. torquis ATCC700755, Antarctica Prydz Bay, sea ice
P. irgensii 23-P, Western Antarctic Peninsula, 10-25 m
clone ABA90994, Sargasso Sea, 40 m
clone ABA90894, Mediterranean Sea, 55 m
clone ABC72937, Sargasso Sea, 80 m
clone ABD85002, Mediterranean Sea, 20 m
clone AAY68048, Mediterranean Sea, 12 m
16S rRNA genes
Candidatus Sulcia muelleri str. Hc
Croceibacter atlanticus HTCC2559
Robiginitalea biformata HTCC2501
Flavobacterium johnsoniae UW101
Psychroflexus torquis ATCC700755
Leeuwenhoekiella blandensis MED217
Dokdonia sp. MED134
Kordia algicida OT-1
Polaribacter irgensii 23-P
Tenacibaculum sp. MED152
Polaribacter filamentus 215
most closely related sequences in GenBank, based on a BLASTP search. Colors indicate SAGs (yellow background), isolates (blue text), and environmental clones
(black text). GenBank accession nos. are provided for each clone. Nodes marked with circles have ?70% neighbor-joining bootstrap support.
Maximum-likelihood phylogenetic trees of bacterial SSU rRNA genes and proteorhodopsins. The SSU rRNA tree includes SAGs from this study and all
www.pnas.org?cgi?doi?10.1073?pnas.0700496104Stepanauskas and Sieracki
(pufM) and three (nasA) 100-cell MDA reactions (of a total of
12), indicating that ?1% of bacterioplankton in the sample
carried either of these genes. Accordingly, bacteriochlorophyll
was previously found to be expressed (infrared fluorescence) in
?1% of bacteria in coastal Maine waters at this time of year (45).
All six pufM were 100% identical to each other and were most
closely related to bacteriochlorophylls from the Roseobacter
lineage (Fig. 2A). Thus, it appears that a single Roseobacter taxon
dominated bacteriochlorophyll-containing bacterioplankton in
the studied sample. Two nasA were most closely related to
assimilatory nitrate reductases in Roseobacter lineage, whereas
one nasA was most closely related to marine Gammaproteobac-
teria (Fig. 2B). The pilot SAG library failed to unambiguously
identify these relatively rare but biogeochemically important
microorganisms. Screening of a larger SAG library would be an
ideal tool for this task. Alternative community genomics-based
analyses have proven less effective to match SSU rRNA, and
functional genes in such rare taxa.
Genes encoding archaeal SSU rRNA and nitrogenases were
not detected in any of the sorted wells, suggesting that Archaea
and nitrogen-fixing organisms were extremely rare or absent in
the analyzed heterotrophic HNA bacterioplankton. Eukaryote
SSU rRNA genes were also not detected, confirming effective
separation of prokaryotes from protists by FACS.
We demonstrate how a combination of single-cell FACS, MDA,
and PCR can be used in metabolic mapping of taxonomically
diverse uncultured marine bacterioplankton. Large quantities of
high molecular-weight whole-genome amplification products
were obtained from individual cells, allowing for a virtually
unlimited number of downstream analyses. In this proof-of-
concept study, we detected proteorhodopsin genes in two of five
flavobacteria, providing evidence that Flavobacteria are major
carriers of this photometabolic gene. We also determined that
Flavobacteria were a major component of HNA bacterioplank-
ton in the analyzed coastal sample. Fewer than 1% of the
analyzed cells carried nasA, pufM, and nifH.
We used standard configuration flow cytometry instrumen-
tation that is available on most major research campuses and
increasingly used aboard oceanographic research vessels. Work-
ing at the single-cell level requires especially stringent instru-
ment cleaning, sample handling, and quality-control methods to
prevent DNA contamination. We show that our methods were
The cost of MDA and subsequent PCR sequencing is on the
order of tens of U.S. dollars per cell and thus is significantly less
expensive than metagenomic sequencing. In addition to high-
throughput screening by PCR or hybridization, SAG libraries
may provide material for genomic sequencing of selected un-
cultured microorganisms. Two of our SAGs are currently in the
process of whole-genome sequencing.
Materials and Methods
Sample Collection and Single-Cell Sorting.Thecoastalwatersample
was collected from Boothbay Harbor, Maine, from 1-m depth at
the Bigelow Laboratory dock (43°50?40??, 69°38?27??W) on
March 28 at 9:45 a.m. during high tide (water temperature
7.0°C). The unmanipulated sample was 10-fold diluted with
filtered (0.2 ?m pore size) sample water and stained with 5 ?M
(final concentration) SYTO-13 nucleic acid stain (Invitrogen,
Carlsbad, CA) for prokaryote detection as in delGiorgio et al.
(46). Individual bacterioplankton were sorted into 96-well plates
containing 5 ?l per well of PBS. Only HNA cells were sorted to
reduce the probability of depositing dead cells with partially
degraded genomes. Single cells were sorted into four of the eight
rows on each plate. Of the remaining rows, two were dedicated
to background controls, consisting of single drops generated
of the side scatter/green fluorescence plot. One row of 12 wells
was dedicated to blanks with no drop deposition, and one row
received 100 HNA bacterioplankton cells per well. Sorting was
done with a MoFlo (Dako Cytomation, Carpenteria, CA) flow
cytometer equipped with the CyClone robotic arm for sorting
into plates, using a 488-nm argon laser and a 70-?m nozzle
orifice. The cytometer was triggered on side scatter, the sort gate
was based on side scatter and SYTO-13 fluorescence, and the
‘‘purify 0.5 drop’’ sort mode was used for maximal sort purity.
Extreme care was taken to prevent sample contamination by any
nontarget DNA. New sheath fluid lines were installed before
each sort day. Sheath fluid and sample lines were cleaned by a
succession of warm water, 5% bleach solution, and an overnight
by dissolving combusted (2 h at 450°C) NaCl in DNA-free
deionized water for a final concentration of 1%. Sorted plates
were stored at ?80°C until MDA.
Lysis and MDA. We compared three protocols for cell lysis, DNA
denaturing, and MDA in this study.
Congregibacter litoralis KT71 (outgroup)
Loktanella vestfoldensis SKA53
Roseovarius sp. 217
Rhodopseudomonas palustris HaA2
Jannaschia sp. CCS1
Roseobacter denitrificans OCh-114
Roseobacter denitrificans OCh 114
Rhodopseudomonas palustris BisB18
Erythrobacter sp. NAP1
Rhodopseudomonas palustris BisB5
Rhodopseudomonas palustris CGA009
Rhodopseudomonas palustris BisA53
Dinoroseobacter shibae DFL-12
clone DelRiverFos06H03, AAX48200
Sulfitobacter sp. NAS-14-1
Roseovarius sp. 217
Roseobacter denitrificans OCh-114
Sulfitobacter sp. EE-36
Rhodobacterales bacterium HTCC2654
Hahella chejuensis KCTC2396
Jannaschia sp. CCS1
Roseovarius sp. HTCC2601
Roseobacter sp. MED193
Thiomicrospira crunogena XCL2
Burkholderia mallei GB8 (outgroup)
Erwinia carotovora SCRI1043
Chromohalobacter salexigens DSM3043
text), isolates of Betaproteobacteria (green text), isolates of Gammaproteobacteria (blue text), and environmental clones (black text). GenBank accession nos.
are provided for each clone. Nodes with yellow circles have ?70% neighbor-joining bootstrap support.
Maximum-likelihood phylogenetic trees of the PufM and NasA. Included are protein sequences obtained from 100-cell MDA reactions and most closely
Stepanauskas and SierackiPNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?
used for cell lysis and DNA denaturing, after which 18-h MDA
was performed by using REPLI-g Mini (Qiagen, Chatsworth,
CA) Phi29 polymerase and reaction buffer. For each well
containing 5 ?l of PBS, we used 0.5 ?l of polymerase, 14.5 ?l of
buffer, and 5 ?l of DNA-free deionized water.
Protocol B. Alkaline lysis on ice and 18-h MDA were performed
by using REPLI-g Mini (Qiagen) kit reagents and following the
manufacturer’s protocol for blood samples.
Protocol C. As protocol B, except that REPLI-g Midi kit (Qiagen)
was used, and PicoGreen DNA stain (Invitrogen) was added to
the reaction at 0.5? (final concentration). DNA synthesis was
monitored with IQ5 real-time PCR system (Bio-Rad, Hercules,
CA). Duplicate standards containing 0.05, 5, 500, and 50,000 fg
of human genomic DNA (Promega, Madison, WI) were ampli-
fied simultaneously with the sort samples.
Initially, each of the three protocols were applied on 24 wells:
12 with single cells, 3 no-drop controls, 6 background controls,
and 3 with 100 cells. Protocols A, B, and C were used after 7, 8,
and 94 days of sorted cell storage at ?80°C, respectively. An
additional 24 wells were analyzed by using protocol B after 350
days of storage. The DNA concentration in MDA reactions was
determined by using a ND-1000 spectrophotometer (Nanodrop)
after a cleanup with MinElute PCR Purification Kit (Qiagen).
PCR-Based Analyses of MDA Products. The MDA products were
diluted 10-fold (protocols A and B; REPLI-g Mini kit products)
or 200-fold (protocol C; REPLI-g Midi kit products). Two
microliters of the dilute products served as templates in 25 ?l of
PCR. Previously described primers and PCR conditions were
used to amplify genes encoding bacterial, archaeal, and eukaryal
SSU rRNA, proteorhodopsin, bacteriochlorophyll, nitrogenase,
and assimilative nitrate reductase (SI Table 2). The PCR prod-
ucts were cleaned with QIAquick (Qiagen). For the T-RFLP of
bacterial SSU rRNA genes, PCR amplicons obtained with
27F-FAM and 907R primers were digested with either HhaI or
MD). Sequencing and fragment analyses were performed with a
3730xl analyzer (Applied Biosystems, Foster City, CA) at the
W. M. Keck Center for Comparative and Functional Genomics,
GeneScan 1000 ROX Size Standard (Applied Biosystems) was
gene analysis with GenBank BLASTN (47) and Ribosomal
Database Project (RDP) Classifier and Seqmatch search tools
(48). The SSU rRNA sequences were checked for chimeras by
were translated with National Center for Biotechnology Infor-
mation ORF Finder and their identities verified by GenBank
BLASTP searches for closest relatives. Evolutionary trees were
constructed by using PHYLIP (49) after an automatic sequence
alignment with ClustalX (50).
T-RFLP Profiling of Bacterioplankton Communities. Triplicate 1,000
cell aliquots of HNA bacterioplankton were sorted as above into
microcentrifuge tubes preloaded with 5 ?l of Lyse-N-Go (Pierce,
Rockford, IL) and then stored at ?80°C. Cell lysis was per-
formed according to Lyse-N-Go instructions. Entire lysate vol-
umes were used as templates in 50-?l 30-cycle PCRs by using
primers 27F and 1492R (SI Table 2). Two-microliter aliquots of
these PCR products served as templates in a second semi nested
25-?l and 30-cycle PCR with primers 27F-FAM and 907R. PCR
products were cleaned and digested, and fragment analyses were
performed as above.
We thank Nicole Poulton and Wendy Bellows (Bigelow Laboratory) for
technical help and Janet Hart (Duke University, Durham, NC) for early
work on single-cell PCR optimization. Arumugham Raghunathan (Qia-
gen) is acknowledged for expert advice on whole-genome amplification
by using REPLI-g kits. This study was funded by the State of Maine
Technology Institute, Marine Research Initiative, and National Science
Foundation SGER award no. EF-0633142.
1. Pace NR, Stahl DA, Lane DJ, Olsen GJ (1986) Adv Microb Ecol 9:1–55.
2. Beja O, Aravind L, Koonin EV, Suzuki MT, Hadd A, Nguyen LP, Jovanovich
S, Gates CM, Feldman RA, Spudich JL, et al. (2000) Science 289:1902–1906.
Loiacono KA, Lynch BA, MacNeil IA, Minor C, et al. (2000) Appl Environ
4. Schloss PD, Handelsman J (2003) Curr Opin Biotechnol 14:303–310.
5. Venter JC, Remington K, Heidelberg JF, Halpern AL, Rusch D, Eisen JA, Wu
DY, Paulsen I, Nelson KE, Nelson W, et al. (2004) Science 304:66–74.
6. Tringe SG, von Mering C, Kobayashi A, Salamov AA, Chen K, Chang HW,
Podar M, Short JM, Mathur EJ, Detter JC, et al. (2005) Science 308:554–557.
7. DeLong EF, Karl DM (2005) Nature 437:336–342.
8. DeLong EF, Preston CM, Mincer T, Rich V, Hallam SJ, Frigaard N-U,
Martinez A, Sullivan MB, Edwards R, Brito BR, et al. (2006) Science
9. Rusch DB, Halpern AL, Sutton G, Heidelberg KB, Williamson S, Yooseph S,
Wu D, Eisen JA, Hoffman JM, Remington K, et al. (2007) PLoS Biol 5:e77.
10. Ottesen EA, Hong JW, Quake SR, Leadbetter JR (2006) Science 314:1464–1467.
11. Zhang K, Martiny AC, Reppas NB, Barry KW, Malek J, Chisholm SW, Church
GM (2006) Nat Biotechnol 24:680–686.
12. Frohlich J, Konig H (1999) Syst Appl Microbiol 22:249–257.
13. Sieracki M, Poulton N, Crosbie N (2005) in Algal Culturing Techniques, ed
Andersen R (Elsevier Academic, New York), pp 101–116.
14. Karl DM, Bailiff MD (1989) Limnol Oceanogr 34:543–558.
15. Brum JR (2005) Aq Microb Ecol 41:103–113.
17. Pinard R, de Winter A, Sarkis GJ, Gerstein MB, Tartaro KR, Plant RN,
Egholm M, Rothberg JM, Leamon JH (2006) BMC Genomics 7:216.
18. Spits C, Le Caignec C, De Rycke M, Van Haute L, Van Steirteghem A,
Liebaers I, Sermon K (2006) Hum Mutat 27:496–503.
19. Dean FB, Hosono S, Fang LH, Wu XH, Faruqi AF, Bray-Ward P, Sun ZY,
Zong QL, Du YF, Du J, et al. (2002) Proc Natl Acad Sci USA 99:5261–5266.
20. Hutchison CA, Smith HO, Pfannkoch C, Venter JC (2005) Proc Natl Acad Sci
21. Abulencia CB, Wyborski DL, Garcia JA, Podar M, Chen W, Chang SH, Chang
HW, Watson D, Brodie EL, Hazen TC, et al. (2006) Appl Environ Microbiol
22. Hellani A, Coskun S, Benkhalifa M, Tbakhi A, Sakati N, Al-Odaib A, Ozand
P (2004) Mol Hum Reprod 10:847–852.
23. Jiang ZW, Zhang XQ, Deka R, Jin L (2005) Nucleic Acids Res 33:e91.
(2005) Appl Environ Microbiol 71:3342–3347.
25. Hutchison CA, Venter JC (2006) Nat Biotechnol 24:657–658.
26. Kolber ZS, Plumley FG, Lang AS, Beatty JT, Blankenship RE, VanDover CL,
Vetriani C, Koblizek M, Rathgeber C, Falkowski PG (2001) Science 292:2492–
27. Zani S, Mellon MT, Collier JL, Zehr JP (2000) Appl Environ Microbiol
28. Allen AE, Booth MG, Frischer ME, Verity PG, Zehr JP, Zani S (2001) Appl
Environ Microbiol 67:5343–5348.
29. Buchan A, Gonzalez JM, Moran MA (2005) Appl Environ Microbiol 71:5665–
30. Moran MA, Buchan A, Gonzalez JM, Heidelberg JF, Whitman WB, Kiene RP,
Henriksen JR, King GM, Belas R, Fuqua C, et al. (2004) Nature 432:910–913.
31. Kirchman DL (2002) FEMS Microbiol Ecol 39:91–100.
32. Mary I, Heywood JL, Fuchs BM, Amann R, Tarran GA, Burkill PH, Zubkov
MV (2006) Aq Microb Ecol 45:107–113.
33. Giovannoni SJ, Rappe MS (2000) in Microbial Ecology in the Oceans, ed
Kirchman DL (Wiley–Liss, New York), pp 47–84.
Aq Microb Ecol 42:119–126.
35. Jaspers E, Nauhaus K, Cypionka H, Overmann J (2001) FEMS Microbiol Ecol
36. Abell GCJ, Bowman JP (2005) FEMS Microbiol Ecol 51:265–277.
37. Sabehi G, Loy A, Jung KH, Partha R, Spudich JL, Isaacson T, Hirschberg J,
Wagner M, Beja O (2005) PLoS Biology 3:1409–1417.
38. Gomez-Consarnau L, Gonzalez JM, Coll-Llado M, Gourdon P, Pacher T,
Neutze R, Pedros-Alio C, Pinhassi J (2007) Nature 445:210–213.
www.pnas.org?cgi?doi?10.1073?pnas.0700496104Stepanauskas and Sieracki
39. Sabehi G, Beja O, Suzuki MT, Preston CM, DeLong EF (2004) Environ Download full-text
40. Frigaard NU, Martinez A, Mincer TJ, DeLong EF (2006) Nature 439:847–850.
41. Beja O, Spudich EN, Spudich JL, Leclerc M, DeLong EF (2001) Nature
42. Man DL, Wang WW, Sabehi G, Aravind L, Post AF, Massana R, Spudich EN,
Spudich JL, Beja O (2003) EMBO J 22:1725–1731.
43. Sabehi G, Massana R, Bielawski JP, Rosenberg M, Delong EF, Beja O (2003)
Environ Microbiol 5:842–849.
44. Giovannoni SJ, Tripp HJ, Givan S, Podar M, Vergin KL, Baptista D, Bibbs L,
Eads J, Richardson TH, Noordewier M, et al. (2005) Science 309:1242–1245.
45. Sieracki ME, Gilg IC, Thier EC, Poulton NJ, Goericke R (2006) Limnol
46. delGiorgio P, Bird DF, Prairie YT, Planas D (1996) Limnol Oceanogr 41:783–789.
47. Altschul SF, Madden TL, Schaffer AA, Zhang JH, Zhang Z, Miller W, Lipman
DJ (1997) Nucleic Acids Res 25:3389–3402.
48. Cole JR, Chai B, Marsh TL, Farris RJ, Wang Q, Kulam SA, Chandra S,
McGarrell DM, Schmidt TM, Garrity GM, Tiedje JM (2003) Nucleic Acids Res
49. Felsenstein J (1989) Cladistics 5:164–166.
50. Chenna R, Sugawara H, Koike T, Lopez R, Gibson TJ, Higgins DG, Thompson
JD (2003) Nucleic Acids Res 31:3497–3500.
Stepanauskas and SierackiPNAS ?
May 22, 2007 ?
vol. 104 ?
no. 21 ?