Insights into chemotaxonomic composition and carbon cycling
of phototrophic communities in an artesian sulfur-rich spring
(Zodletone, Oklahoma, USA), a possible analog for ancient
microbial mat systems
S. I. BU¨HRING,1S. M. SIEVERT,2H. M. JONKERS,3,* T. ERTEFAI,1,?M. S. ELSHAHED,4,?
L. R. KRUMHOLZ4AND K.-U. HINRICHS1,5
1Departmentof GeosciencesandMARUMCenterforMarine EnvironmentalSciences,Universita ¨t Bremen,Bremen,
2Biology Department, WoodsHole Oceanographic Institution,WoodsHole,MA,USA
3MaxPlanck InstituteforMarine Microbiology, MicrosensorGroup, Bremen, Germany
4Departmentof BotanyandMicrobiology, Universityof Oklahoma,Norman, OK,USA
5Departmentof GeologyandGeophysics,WoodsHole Oceanographic Institution, WoodsHole,MA,USA
Zodletone spring in Oklahoma is a unique environment with high concentrations of dissolved-sulfide (10 mM)
and short-chain gaseous alkanes, exhibiting characteristics that are reminiscent of conditions that are thought to
have existed in Earth’s history, in particular the late Archean and early-to-mid Proterozoic. Here, we present a
process-oriented investigation of the microbial community in two distinct mat formations at the spring source,
(1) the top ofthe sediment in the source pool and (2) the purple streamersattached to the side walls. We applied
a combination of pigment and lipid biomarker analyses, while functional activities were investigated in terms of
sis showed cyanobacterial pigments, in addition to pigments from purple sulfur bacteria (PSB), green sulfur bac-
teria (GSB) and Chloroflexus-like bacteria (CLB). Analysis of intact polar lipids (IPLs) in the source sediment
confirmed the presence of phototrophic organisms via diacylglycerol phospholipids and betaine lipids, whereas
glyceroldialkylglyceroltetraether additionally indicated the presence of archaea. No archaeal IPLs were found in
the purple streamers, which were strongly dominated by betaine lipids.13C-bicarbonate- and -acetate-labeling
experiments indicated cyanobacteria as predominant phototrophs in the source sediment, carbon was actively
fixed by PSB⁄CLB⁄GSB in purple streamers by using near infrared light. Despite the presence of cyanobacteria,
no oxygen could be detected in the presence of light, suggesting anoxygenic photosynthesis as the major meta-
bolic process at this site. Our investigations furthermore indicated photoheterotrophy as an important process in
both habitats. We obtained insights into a syntrophically operating phototrophic community in an ecosystem
that bears resemblance to early Earth conditions, where cyanobacteria constitute an important contributor to
carbonfixationdespitethepresence ofhigh sulfideconcentrations.
Received22 July2010;accepted 16 November2010
Corresponding authors: Solveig I. Bu ¨hring. Tel.:+49 421 21865744; fax: +49 421 21865715; e-mail: solveig.
StefanM.Sievert.Tel.: +1 508 2892305;fax: +1 508 4572076;e-mail: firstname.lastname@example.org
*Presentaddress:Civil Engineering & Geosciences, DelftUniversity of Technology,Stevinweg1, 2628 CNDelft, The Netherlands.
?Present address: WA – Organic & Isotope Geochemistry Centre, Curtin University of Technology, GPO Box U1987, Perth, WA
?Present address: Oklahoma State University, 1110 S. Innovation Way, Stillwater, OK 74074, USA.
? 2011 Blackwell Publishing Ltd
Geobiology (2011)DOI: 10.1111/j.1472-4669.2010.00268.x
The understanding of geochemical signals of extant microbial
communities in extreme environments is vital to the study of
ancient microbial ecosystems. Modern analogs can potentially
provide insights into the functioning of these ecosystems, yet
their presence on today’s Earth is largely restricted (e.g.
Crowe et al., 2008). In particular, surficial microbial mat-like
communities thriving under anoxic conditions in the presence
of light and at moderate temperatures, conditions thought to
have prevailed in the Archean and into most of the Protero-
zoic, are largely absent from today’s Earth, due to the high
atmospheric concentration of oxygen and the competition
with oxygenic phototrophs (e.g. Herman & Kump, 2005).
However, the geological record of the late Archean and early-
to-mid Proterozoic indicates that microbial mats in sunlit, lar-
gely anoxic waters were a dominant feature of the ancient
tivity (Grotzinger & Knoll, 1999; Tice & Lowe, 2004; Her-
man & Kump, 2005; Noffke et al., 2008; Heubeck, 2009).
There is considerable debate about the exact timing of the
origin of oxygenic photosynthesis (Summons et al., 1999;
Kaufman et al., 2007; Fischer, 2008; Kump, 2008; Rasmus-
sen et al., 2008; Sessions et al., 2009; Waldbauer et al.,
2009). However, it is clear that the first mats occurring at
about3.4 Gawereprimarilycomposedofso-called anoxygen-
ic phototrophs, like purple sulfur- and green sulfur bacteria
(PSB and GSB), or even procyanobacteria, possibly using
hydrogen, ferrous iron and⁄or sulfide as electron donors for
photosynthesis (Beukes, 2004; Tice & Lowe, 2004, 2006;
Herman & Kump, 2005; Mulkidjanian et al., 2006; Noffke
et al., 2008; Heubeck, 2009). It is in this setting that micro-
organisms carrying out reductive sulfur transformations are
likely to have evolved, leading to the establishment of an
anaerobic, light driven sulfur cycle in an otherwise sulfate-
depleted environment (Shen et al., 2001; Philippot et al.,
in a microbial mat-like environment, leading to the stepwise
oxidation of Earth’s atmosphere and the ocean starting at
about 2.4 Ga (Bekker et al., 2004; Kaufman et al., 2007;
Scott et al., 2008; Sessions et al., 2009). Oxygenic photosyn-
thesis evolved only once in Earth’s history and was the result
of combining photosystem I (PS I) and II (PS II) known to
operate in either GSB or PSB, respectively, within one organ-
ism (Mulkidjanian et al., 2006). Paradoxically, the initial
increase in atmospheric oxygen concentration lead to a stimu-
lation of sulfate reduction and thus sulfide production due to
an increased availability of sulfate as a consequence of elevated
weathering of sulfide minerals on land (Canfield, 1998; Rein-
hard et al., 2009). Large scale photic euxinia followed, pro-
moting again the growth of anoxygenic phototrophic bacteria
in the water column as well as in widespread microbial mats,
largely driven by light-mediated sulfide oxidation (e.g. Brocks
et al., 2005; Johnstonet al., 2009).
Zodletone spring in southwestern Oklahoma is a unique
environment that supports ecosystems resembling those that
might have existed in Earth’s history, in particular the late
Archean and early Proterozoic (Elshahed et al., 2003). The
highly sulfidic (10 mM), yet sulfate-poor (149 lM) spring-
water, that also contains abundant gaseous alkanes, creates
anoxic conditions in the light exposed source pool and its
effluent stream (Senko et al., 2004). This leads to the estab-
lishment of highly diverse microbial communities that are
largely driven by light-induced sulfur cycling (Elshahed et al.,
2003). While previous work has focused on the cycling of
sulfur and investigated the diversity of the microbial commu-
nities in this system (Elshahed et al., 2003, 2004), less is
known about the cycling of carbon in these communities, a
potentially important aspect for better understanding the
productivityof similar systems inEarth’s past.
In this study, we focused on the structure and function of
the phototrophic microbial community of two closely coex-
isting habitats of Zodletone spring, the top layers of the
sediment in the source pool as well as the purple streamers,
microbial mat-like structures attached to the side wall of the
source pool. The incorporation of both structure- and func-
tion-based approaches is especially useful as a number of
phototrophic bacteria, including cyanobacteria, are known
for their metabolic versatility, being able to switch between
different metabolisms (e.g. oxygenic- and anoxygenic photo-
autotrophy, photoheterotrophy, chemoautotrophy, anoxy-
genic heterotrophic sulfur respiration and fermentation).
For structural community analysis, we applied microscopy,
high-performance liquid chromatography (HPLC) to iden-
tify dominant photopigments, intact polar lipid (IPL) analy-
ses, and 16S ribosomal RNA (16S rRNA) gene-based
approaches. The analysis of lipids in the intact form (IPLs) is
a cultivation- and gene-independent approach, allowing an
unbiased assessment of microbial community composition
(e.g. Sturt et al., 2004). The lipid patterns represent finger-
prints of the microbial community with all members con-
tributing mixtures of different lipids of varying specificity
and in varying amounts. Together with the analysis of polar
lipid-derived fatty acids, IPLs can have a high taxonomic
specificity (e.g. Fang et al., 2000a; Ru ¨tters et al., 2002b).
Due to their finite lifetime after cell death (Lipp et al.,
2008), IPLs are useful markers of living cells, as opposed to
gene-based analysis, where DNA from dead cells and live
cells is indistinguishable. Another benefit of lipid analysis
occurs when lipid remnants become fossilized, allowing their
preservation in the geological record. Identification of these
and comparison with similar modern lipids can help establish
a source for orphan lipids in the geological past. For func-
tional analyses of phototrophic carbon metabolism, we per-
formed stable carbon isotope (13C)-labeling experiments in
combination with microsensor measurements of oxygen and
sulfide. This strategy enabled us to decipher carbon flow
within the community as well as to assess the potential for
2S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
photoautotrophic and photoheterotrophic processes, and
pinpoint organisms actively involved in these processes. The
geobiological relevance of the microbial communities inhab-
iting Zodletone spring for understanding Earth’s ancient
microbial ecosystems prior to the presence of atmospheric
oxygen is discussed.
MATERIALS AND METHODS
Site descriptionand sampling
Zodletone spring is an artesian sulfide-rich spring located in
Kiowa County in southwestern Oklahoma, USA. The water
that originates deep within the Anadarko Basin emerges at a
flow rate of 8 L min)1, degassing methane, ethane and pro-
pane. The dissolved sulfide concentration is high (10 mM),
whereas sulfate is low (149 lM). The source is artificially
dug out and cemented, encompassing an area of about 1 m2
overlaid by approximately 60 cm of water. The bottom is
filled with at least 15 cm of soft black sediment. From the
source pool, the spring flows approximately 20 m as a
stream before discharging into a creek. Further details on
the geochemistry are given in Table 1. Sampling took place
in August 2005. Samples were taken from the soft black
bottom sediment from the source and from purple streamers
that were found attached to the source wall. Fresh samples
were investigated by bright field microscopy using a Leitz
Laborlux D microscope (Leica, Wetzlar, Germany) at the
University of Oklahoma to determine the presence and to
identify tentatively phototrophic community members.
Salinity, pH, conductivity and total dissolved sulfur were
determined using an EXTECH II electrode (ExTech Instru-
ments, Waltham, MA, USA). Redox potential was measured
using a HACH II ORP probe (Hach Corp., Loveland, CO,
USA). Alkalinity and total Fe(II) were detected using com-
mercially available kits (Hach Corp.). Sulfate, chloride, thio-
sulfate, nitrate, nitrite and phosphate were quantified using a
dionex ion chromatograph equipped with an AS4A column
and conductivity detector (Dionex corp, Sunnyvale, CA,
USA). These measurements were performed in May 2010.
Sulfide was determined by using the microsensors in 2005.
Although this measurement is less sensitive than other geo-
chemical methods, the detected value of 10 mM is in the same
rangeasdetermined by Senkoet al.(2004).
Photopigments were analyzed by HPLC at the Max Planck
Institute for Marine Microbiology in Bremen (Germany)
following the procedure described by Jonkers et al. (2003).
In short, samples were extracted in 100% methanol and sub-
sequently separated on a Waters HPLC (2690 Separation
Module) (Waters Corp., Milford, MA, USA) equipped with
a Eurospher-100 C18, 5 Wm Vertex column (Knauer, Ber-
lin, Germany). Absorption spectra of separated compounds
were measured on a Waters 996 Photo Diode Array (PDA)
detector and pigment quantity and purity were checked by
comparing peak areas, peak retention times and absorption
spectra to available pigment standards.
16S rRNA gene-based analyses
Phylogenetic analyses of the source sediment have been pub-
lished by Elshahed et al. (2003). DNA from the purple
streamers was extracted using the UltraClean Soil DNA
extraction kit (MoBio Laboratories, Solana Beach, CA, USA).
Bacterial 16S rRNA genes were amplified using primers
broadly specific for Bacteria (8F: 5’-AGAGTTTGATCMTG-
GC-3’ and 1492R: 5’-TACCTTGTTACGACTT-3’), and a
clone library was constructed using the TOPO TA cloning kit
(Invitrogen, Carlsbad, CA, USA) as described by the manu-
facturer. Colonies were picked, plasmid DNA purified with an
automated plasmid purification system (Gene Machines
RevPrep, Ann Arbor, MI, USA), and a total of 23 clones were
sequenced using the Taq Dyedeoxy Terminator Cycle
Sequencing kit (Applied Biosystems, Foster City, CA, USA)
on an Applied Biosystems 3730XL capillary sequencer at the
Josephine-Bay-Paul Center at the Marine Biological Labora-
tories (Woods Hole, MA, USA). The sequences were loaded
into the 16S rRNA sequence database of the Technical Uni-
versity of Munich (Munich, Germany) using the program
package ARB (Ludwig et al., 2004). The sequence was aligned
automatically by using ARB_ALIGN and then checked and cor-
rected manually, considering the secondary structure of the
rRNA molecule. Phylogenetic analyses were performed using
the neighbor-joiningdistancealgorithmimplemented in ARB.
Table 1 Physical and geochemical characteristics of the Zodletone springwater
Physical orgeochemical characteristic
0.19 ± 0.03 M
0.149 ± 0.029 mM
0.628 ± 0.292 mM
240 mg L)1
0.6 mg L)1
TDS,total dissolved sulfur;BDL,below detectionlimit.
*Determined bymicrosensor measurements in2005.
Carbon cycleinsulfur-rich spring3
? 2011 Blackwell Publishing Ltd
Samples were freeze dried prior to extraction, spiked with
internal standard [1-O-hexadecyl-2-acetyl-sn-glycero-3-phos-
phocholine (PAF) and 2-methyl-octadecanoic acid], and the
IPLs were extracted using the method of Sturt et al. (2004).
The total lipid extracts were saponified as previously described
(Elvert et al., 2003). This method includes a base saponifica-
tion usingKOHinmethanolandsubsequentlya base and acid
extraction of the neutral lipids and the free fatty acids, respec-
tively. The free fatty acids were esterified using BF3-methanol
Compound quantifications and identifications were per-
formed using a Thermo-Finnigan Trace GC coupled to a
FinniganTraceMSplus. A 30 m RestekRtx?-5MS silica capil-
lary column was used (internal diameter: 0.25 mm; film thick-
ness: 0.25 lm). The GC-MS was operated in electron impact
mode at 70 eV with a full scan mass range of 400–800 m⁄z.
The initial oven temperature was held at 60 ?C for 1 min,
increased to 150 ?C with a rate of 10 ?C min)1, then raised to
a temperature of 310 ?C with a rate of 4 ?C min)1and held at
310 ?Cfor20 min.Thecarriergaswashelium with aconstant
flow of 1.0 mL min)1(detection limit: 1 ng C). Analyses of
matograph coupled to an isotope ratio mass spectrometer
(GC-IRMS).The mass spectrometer (MAT 252) isconnected
via a Finnigan Combustion Interface III to a HP 5890 Series
GC, using the same temperature program as for MS measure-
ments. A 30-m Zebron ZB-5 capillary column was used
(internal diameter: 0.25 mm; film thickness: 0.25 lm). The
isotopic compositions of the fatty acid methyl esters were cor-
rected for the methyl group, which was inserted via transeste-
rification. The accuracy of isotope results was monitored
routinelywith standardsand foundto be 0.5& orlower.
For the chromatographic separation of the IPLs, a normal
phase chromatography procedure from Ru ¨tters et al. (2002a)
was adapted according to Lipp et al. (2008). Multistage Mass
Spectrometry (MSn) experiments were performed using a
ThermoFinnigan LCQ Deca XP Plus ion-trap mass spectrom-
eter (ThermoFinnigan, San Jose, CA, USA) with an electro-
spray ionization (ESI) interface. ESI settings were as follows:
capillary temperature 200 ?C, capillary voltage ±11 V, sheath
gas flow 40 (arbitrary units), spray voltage was set at ±5 kV.
During the measurements, the mass spectrometer was config-
ured to rundata dependentiontree experiments,i.e.themole-
cules corresponding to the base peak of each full scan
experiment were trapped and fragmented up to MS3. The full
scan was set to 500–2000 m⁄z. Where available, the resulting
IPL concentrations were corrected by response factors
obtained by using reference standards. The IPLs, where no
standards were available, were corrected by an averaged
response factor. The detection limit is determined for each
individual run using the same approach as shown in Lipp et al.
(2008), varying from 10to 26 ng g)1sediment.
Microsensor measurements were done to investigate oxygen
production and sulfide consumption processes in both the
source sediment and purple streamers attached to the source
walls.The profiles through thestreamers were measured verti-
cally, thus parallel to the wall to which they were attached.
Custom made glass amperometric Clark-type oxygen and sul-
fide sensors were elongated with glass pipettes to enable in
situ measurementsof oxygen and sulfideconcentrationsin the
source sediment and purple streamers. Measurements were
performed in the morning when light intensity was between
400 and 800 lmol photons m)2s)1. Sensors were calibrated
on site using tap water and source water with defined concen-
trations of oxygen and sulfide [see Ludwig et al. (2005) for
detailed description of sensor calibration procedures]. Micro-
sensors were fixed on site in a micromanipulator mounted on
a heavy stand, allowing vertical movement of microsensors
with depth increments of 0.2 mm. In the early morning, a
green biofilm was visibly present on the surface of source sedi-
ment andthis biofilm retractedinto the firstsurface millimeter
of the sediment during continued illumination. These condi-
tions could also be artificially created during the day. Within
15 min or so after darkening the sediment surface the green
biofilm re-appeared on the surface, indicating the active verti-
cal movement of phototrophic bacteria. Profiles of oxygen
and sulfide concentration were measured vertically from the
overlying water into thesourcesediment orpurple streamers.
13C-labeling experiments anddata analysis
Labeling experiments were performed with13C-bicarbonate
and 1-13C-labeled acetate (99%; Cambridge Isotope Labora-
tories Inc., Andover, MA, USA) to link the phototrophs to
their in situ function and to assess their relative contribution
to photoautotrophy and photoheterotrophy. Labeled bicar-
bonate and acetate were added to source sediment and strea-
mer samples and incubated under different light regimes. For
tions were further varied to distinguish inorganic carbon fixa-
tion in cyanobacteria (only allowing light in the visible range,
VIS-only) from other anoxygenic photosynthetic bacteria
(only allowing light in the near infrared range, NIR-only).
Finally,uptake into fattyacidswasmonitored.
Experiments with the source sediments were performed
in 150-mL glass vials using approximately 50 mL of sedi-
ment and bottles were filled up with spring water. Experi-
ments with the purple streamers were performed in 20-mL
glass vials by addition of approximately 10 g of biomass.
Labeled bicarbonate or acetate was predissolved in source
water and added to the incubation (500 lM final concen-
tration) by injection through the lid. Dark incubations took
place in glass vials wrapped with aluminum foil. For exclu-
sion of visible or infrared light, vials were incubated under
13C-bicarbonate experiments, the experimental condi-
4S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
either a hot or cold mirror, reflecting the respective wave
length (Edmunds Industrial Optics, Barrington, NJ, USA).
The mirrors were operated in 0? angle of incidence. The
cold mirror reflects >90% of light between 425 and
675 nm and allows transmission of >85% of light between
800 and 1200 nm. The hot mirror reflect over 95% of light
between 750 and 1150 nm and allows transmission of
>92% of 425–675 nm light. The experiments were per-
formed under near in situ conditions directly in the source
to mimic natural light and temperature conditions. Experi-
mental duration was 3 h for the source sediment and 2 h
for the purple streamers. Killed controls were prepared
using formaldehyde at a final concentration of 4%.
Carbon isotopic ratios (13C⁄12C) are expressed in the delta
notation (d13C) relative to Vienna Pee Dee Belemnite
Standard (13C⁄12CVPDB= 0.0112372 = RVPDB): d13C (&) =
[(Rsample⁄Rstd) ) 1)] · 1000, where Rsampleand Rstdare the
13C⁄12C of sample and standard, respectively (Craig, 1957).
Incorporation as13C is reflected as excess (above background
samples)13C and is expressed in terms of total uptake (I) as
well as specific uptake (i.e. Dd13C = d13Csample) d13Ccontrol)
after Middelburg et al. (2000). Total uptake I of13C in lipids
was calculated as the product of excess13C (E) and concentra-
tion of the respective compound. E is the difference between
the fraction F of the sample and background: E = Fsample)
F =13C⁄(13C⁄12C) = R⁄(R + 1)
R = (d13C⁄1000 + 1) · RVPDB.
RESULTS AND DISCUSSION
Pigment analysis,microscopic observations and 16SrRNA
Pigment analyses reflect a mixed phototrophic community
in the source sediment (Table 2). Chlorophyll a (Chla),
lutein and zeaxanthin accounted for 7%, 30% and 11% of the
total pigments in the source sediment, respectively, indicating
a substantial population of cyanobacteria in the source sedi-
ment (Jonkers et al., 2003). Bacteriochlorophyll (BChl) c
accounted for 14%, and is probably derived from Chloroflex-
us-like bacteria (CLB; Bachar et al., 2007) or from GSB
(Borrego & Garciagil, 1994). Another indicator for GSB
and CLB is BChle, which accounted for 3% of the pigments.
b-Carotene is a general marker for phototrophic organisms,
which accounted for 8%. A degradation product of Chla
and an unknown carotene-like pigment #1 accounted
for 10% and 17% of the pigments in the source sediment,
By contrast, the purple streamers showed a strong domi-
nance of BChle, which accounted for 77% of all pigments
(Table 2). BChle and isorenieratene are specific biomarkers
for GSB (Borrego & Garciagil, 1994), accounting together
for nearly 80% of pigments, implying that these organisms are
the dominant phototrophs in the purple streamers. Another
abundant pigment in the purple streamers was an unidentified
carotene, which has also been found in cultivated PSB (H.M.
Jonkers, unpubl. data). BChla made up 2% of the pigments,
which has also been shown in PSB (Borrego & Garciagil,
1994). Microscopic observation further showed the presence
of filamentous cyanobacteria in the purple streamers. It is
therefore likely that Chla and zeaxanthin were also present in
the purple streamers, butgotmaskedbycarotene#2.
Furthermore, microscopic observations of the phototro-
phic communities of the source sediment and the purple
streamers revealed an assemblage of Oscillatoria-like filamen-
tous cyanobacteria, motile PSB (Thiocystis-like) and CLB. The
source sediment was dominated by cyanobacteria, whereas
the purple streamers contained mainly PSB and CLB. The
presence of Thiocystis-like PSBin the purple streamers could be
further confirmed based on 16S rRNA analyses (Fig. 1).
Interestingly, the most closely related sequences come from
the chemocline of Fayetteville Green Lake (New York, USA),
Table 2 Composition andputative originofthepigments foundinthesourcesedimentandthepurplestreamers
streamers (%) Putativeoriginforoursetting
Plants, algae, cyanobacteria (Jonkerset al., 2003)
Cyanobacteria(Jonkers et al.,2003)
Plants, algae, cyanobacteria (Jonkerset al., 2003)
GSB(Borrego & Garciagil,1994),CLB(Bacharet al.,2007)
GSB(Borrego & Garciagil,1994),CLB(Bacharet al.,2007)
GSB(Borrego & Garciagil,1994)
PSB (Borrego & Garciagil,1994)
PSB (H.M.Jonkers, unpubl. data)
PSB (H.M.Jonkers, unpubl. data)
Chla, chlorophylla; BChl,bacteriochlorophyll;GSB,green sulfurbacteria;PSB, purplesulfurbacteria;CLB,Chloroflexus-likebacteria;n.d.,notdetected.
Carbon cycleinsulfur-rich spring5
? 2011 Blackwell Publishing Ltd
a meromictic lake that is being used as a modern analog of
euxinia in Earth’s history (Zerkle et al., 2009). Besides PSB,
epsilonproteobacteria dominated the clone library, in particu-
larsequences closelyrelated to epsilonproteobacteria originat-
ing from a biofilm in a sulfidic stream in Frasassi Cave (Italy)
(Fig. 1). The most closely related cultivated organisms are
Sulfurovum species, which are known to be chemolithoauto-
trophs able to oxidize reduced sulfur compounds and hydro-
gen with either oxygen or nitrate (Campbell et al., 2006).
In addition, in a previous 16S rRNA-based investigation,
representative sequences of all the above-mentioned photo-
trophic bacterial groups were detected in the source sediment
(Elshahed et al., 2003), in line with ourresults.
The IPL distribution of the source sediment is depicted in
Table 3. The IPLs were dominated by the diacylglycerol
phospholipids phosphatidylethanolamine (PE-DAG, 28%),
phosphatidylglycerol (PG-DAG, 20%) and phosphatidylcholine
Fig. 1 16S ribosomal RNA-based neighbor-joining distance tree depicting the phylogenetic relationship of the Thiocystis-like (A) and epsilonproteobacterial (B)
sequences from the purple streamers to select cultured and environmental proteobacterial sequences. In total, 23 clones were obtained of which 4 (PS_Gamma) and
10 (PS_Epsilon) almost identical sequences affiliated with PSB and epsilonproteobacteria, respectively. The remaining clones were only present as single clones and
belonged to Deltaproteobacteria, Epsilonproteobacteria, Bacteroidetes, Spirochaeta and Chloroflexaceae (data not shown). The tree was constructed by using the
phylogeneticsoftware ARB(Ludwigetal.,2004).Thescalebarrepresents0.05 estimatedchanges pernucleotide.
Table 3 Composition andputative originoftheintact polarlipidsfoundinthesource sediment andthepurplestreamers
Source sediment (%)
Purplestreamers (%) Putativeoriginforoursetting
Gly-DAG1 32* Cyanobacteria(Brett& Mu ¨ller-Navarra,1997),anoxygenicphototrophs(Hoelzl &
Doermann, 2007),GSB(Imhoff & Bias-Imhoff, 1995)
Cyanobacteria(Dembitsky, 1996),anaerobic bacteria (Schubotzet al.,2009)
Cyanobacteria(Wada& Murata,1998),GSB(Barridge& Shively,1968; Imhoff&
Bacteria(Lopez-Laraet al., 2003),Gram-negativebacteria (Asselineau,1991),
SRB(Makula& Finnerty,1975),sulfur-oxidizingbacteria (Shively& Knoche, 1969;
Schubotz et al.,2009)
Archaea(Kogaet al., 1998;Lipp et al., 2008;Rossel et al.,2008)
Archaea(Morii& Koga, 1993;Lipp et al., 2008;Rossel et al.,2008)
Gly-DAG,monoglycosyldiacylglycerol;BL,betainelipids; PG,phosphatidylglycerol;PC,phosphatidyl-choline;PE,phosphatidylethanolamine; OL,ornithine lipids;
GlyA-DAG, glucuronic acid diacylglycerol;2Gly-Archaeol,glyceroldialkylarchaeol; 2Gly-GDGT, glyceroldialkylglyceroltetraether;n.d., notdetected.
6 S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
(PC-DAG, 15%), as well as betaine lipids (BL, 19%). Minor
contributions came from ornithine lipids (OL, 7%), and
from monoglycosyldiacylglycerol (Gly-DAG, 1%), as well as
from the archaeal lipids glyceroldialkylglyceroltetraether
archaeol) (Rossel et al., 2008). Of these identified lipids,
Gly-DAG, BL, PG-DAG and PE-DAG may derive from
different types of photosynthetic bacteria (see Table 3). BL
are common constituents in lower plants and marine algae
(Dembitsky, 1996). Due to the high concentration of sul-
fide (10 mM) of the spring water, which is generally
assumed to be toxic for algae and plants, an active algae
community in the source pool is not very likely. However,
plant material from the surrounding vegetation could be
deposited in the source pool and thus contribute to the
observed lipids. Long-chain fatty acids were additionally
observed (data not shown), which are indicative for plant
input (Bianchi et al., 1989). Within the prokaryotes, BL
are known to be present in cyanobacteria (Brett & Mu ¨ller-
Navarra, 1997), which appear to be the most likely source.
However, BL have also been identified in purple non-sulfur
bacteria within the alphaproteobacteria, e.g. Rhodobacter
sphaeroides, where they have been shown to be produced in
response to phosphate limitation (Benning et al., 1995). In
addition, genes encoding for the enzymes involved in the
biosynthesis of betaines have been detected in members of
the alphaproteobacterial subdivision (Lopez-Lara et al.,
2003). PG is, together with PE, the most widespread phos-
pholipid in eukaryotes and bacteria (Dowhan, 1997), and
is present in all photosynthetic membranes (Wada & Mura-
ta, 2007). PG-DAG is found not only in both aerobic and
anaerobic sulfur bacteria, including GSB (Barridge & Shive-
ly, 1968; Imhoff & Bias-Imhoff, 1995), but also in non-
photosynthetic groups such as nitrifying bacteria (Goldfine
& Hagen, 1968), metal-oxidizing bacteria (Short et al.,
1969) and methanotrophic bacteria (Makula, 1978; Fang
et al., 2000b). Other plausible sources for PG in anoxic
waters are sulfate-reducing bacteria (SRB) (Ru ¨tters et al.,
2001; Sturt et al., 2004) or fermenting bacteria (Desiervo
& Reynolds, 1975).
The archaeal biomarkers 2Gly-GDGT (8%) and 2Gly-
archaeol (3%) indicated the presence of archaea in the source
sediment. An assessment of the archaeal diversity in the source
sediment has previously revealed an archaeal community con-
sisting mainly of phylotypes affiliated with the Methanomicro-
biales, Methanosarcinales, Halobacteriales, and diverse groups
of uncultured archaea (Elshahed et al., 2004). Glycosidic
archaeol- and GDGT-based IPLs are typical components of
methanogenic and methanotrophic archaea (e.g. Koga &
Nakano, 2008; Rossel et al., 2008), as well as archaea identi-
fied in deeply buried marine sediments (Lipp et al., 2008).
Both methane-metabolizing archaea and benthic sedimentary
archaea are conceivable sources given the abundantly present
The IPL composition of the purple streamers revealed a
substantially different bacterial community (Table 3), that
was dominated by BL (51%), followed by the glycolipids Gly-
DAG and glucuronic acid diacylglycerol (GlyA-DAG) (32%).
Of the phospholipids detected, PG-DAG was most important
(6%), followed by PE-DAG and PC-DAG(4% and 3%,respec-
tively). OL accounted for a 4%. As already outlined above for
the source sediment, the abundant BL in the purple streamers
probably derive from phototrophic bacteria (Dembitsky,
1996). Together with the other results (pigment analysis,
microscopic observations and 16S rRNA analysis), revealing a
dominance of PSB⁄CLB, we conclude that PSB and⁄or CLB
area likely sourcefor BL in the purple streamers. In contrast to
the source pool sediments, a contribution from allochtonous
sources can be largely excluded as the streamers develop on
the side walls of the pool, preventing the accumulation of any
imported organic matter. Gly-DAG is the most widespread
glyceroglycolipid in nature and is typically found in all photo-
trophic membranes, and is also abundantly present in PSB
(Hoelzl & Doermann, 2007), which is the group that most
likely contributed significantly to the filament’s biomass based
on its purple coloration. OL were also abundantly present in
the purple streamers, and these are assumed to be restricted to
Bacteria (Lopez-Lara et al., 2003). OL are widespread among
Gram-negative bacteria (Asselineau, 1991) and have been
identified in the sulfate-reducing bacterium Desulfovibrio
gigas (Makula & Finnerty, 1975), as well as in some iron- and
sulfur-oxidizing bacteria (e.g. Shively & Knoche, 1969; Ert-
efai et al., 2008; Schubotz et al., 2009). Based on known
sources and abundances of OL in environments with active
sulfur cycling, it has been previously suggested that sulfur-
metabolizing bacteria are the most likely source of OL in
anoxic water bodies (Schubotz et al., 2009). In summary,
both investigated communities are dominated by IPLs origi-
natingfromdiversephototrophic bacteria,withdistinct differ-
encesreflecting variationsin community composition.
In conclusion, the analysis of the community composition
based on microscopy, IPLs, and pigments of the source sedi-
ment and the purple streamers, revealed that these habitats are
composed of a variety of phototrophs, namely cyanobacteria,
PSB, GSB and CLB, with the cyanobacteria being the domi-
nant phototrophs in the source sediments, and GSB and PSB
dominating the purple streamers. Our results further suggest
that PSBand⁄orGSBaresynthesizing BL.
Pigment,IPLand microscopicobservationsshowed abundant
cyanobacteria in the examined samples. The observation of
active cyanobacteria in a highly sulfidic environment brought
up the question if they were able to produce oxygen under
these conditions, with obvious implications for the interpreta-
tion of the geological record. In both investigated Zodletone
habitats, the production of oxygen was not detectable under
Carbon cycleinsulfur-rich spring7
? 2011 Blackwell Publishing Ltd
light conditions, suggesting that the observed cyanobacteria
1991). In addition, a substantial drop in H2S concentration
was observed in the illuminated green surface of the source
sediment (Fig. 2), a layer dominated by Oscillatoria-like fila-
mentous cyanobacteria. The drop in H2S concentration prob-
ably reflects substantial sulfide consumption in this particular
zone. Zodletone spring has sulfide concentrations of 10 mM
which exceeds those found in most fresh water and marine
sulfidic environments including microbial mats (Stal, 1995).
In some cyanobacteria, mere exposure to low redox potentials
may drastically inhibit oxygen evolution. Under these condi-
tions, some cyanobacteria are able to switch from oxygenic
photosynthesis to anoxygenic photosynthesis (Arnon et al.,
1961; Castenholz et al., 1991). Sulfide-dependent anoxygen-
ic photosynthesis among cyanobacteria, driven by PS I alone,
was first described for Oscillatoria limnetica from Solar Lake,
Sinai (Cohen et al., 1975). Four types of adaptations of
cyanobacteria to high sulfide concentration have been
described (Cohen et al., 1986). In our setting, we are most
likely dealing with the fourth type: sulfide-sensitive oxygenic
photosynthesis replaced by sulfide-dependent anoxygenic
photosynthesis. PS II in these cyanobacteria is sensitive to
sulfide, and as a result oxygenic photosynthesis is already
completely inhibited at relatively low sulfide concentrations.
Anoxygenic photosynthesis operated optimally in O. limneti-
ca at sulfide concentrations between 0.2 and 1 mM (Cohen
et al., 1986). At these concentrations PS II was fully inhibited
and PS I fully induced, allowing efficient CO2photoassimila-
tion rates that are comparable to oxygenic photosynthesis in
the absenceof sulfide. Yet,even afterprolonged growth under
high sulfide levels, during which PS II is completely blocked,
O. limnetica still maintained the capacity for oxygenic photo-
synthesis and switched readily to oxygenic photosynthesis
upon sulfide removal (Cohen et al., 1986). In addition to
their high tolerance for sulfide, this metabolic flexibility gives
the cyanobacteria an additional selective advantage over both
eukaryotic phototrophs such as diatoms and photosynthetic
bacteria, which are limited either to an exclusively oxygenic
(diatoms) or anoxygenic (GSB, PSB) mode of phototrophic
tion (Fig. 3A), reaching a total uptake into all investigated
fatty acids of 21 ng g)1sediment [dry weight (dw)]. Experi-
ments with ambient light showed an uptake of 8 ng g)1sedi-
ment (dw) and for the NIR-only incubation 6 ng g)1
sediment (dw) were determined in total. In the absence of
light, the13C uptake ofbicarbonate waslower,buta consider-
able uptake of 3 ng g)1sediment (dw) took still place. No
incorporationtook placein thepoisoned control.
Under all illuminated experimental conditions, uptake was
mainly into C14:0, C16:1and C16:0. No uptake into C14:0was
detectable for the dark treatment in the source sediment, indi-
cating that C14:0is exclusively produced by phototrophs in
this system. C16:1and C16:0became labeled under all tested
experimental condition, showing their ubiquity in this envi-
ronment. iC16:0and C18:2were only labeled or mainly labeled
in VIS-only, respectively, indicating a possible cyanobacterial
origin for these fatty acids. In the absence of light, uptake was
mainly into C16:1. Some uptake under IR-only may be due
to the activities of cyanobacteria, which might be able to use
IR-lightorthelowamountsofVIS-light(£10%) that stillpen-
etrates the mirrors. Characteristic was the exclusively dark
incorporation, albeit low, into aiC15:0as well as predominant
dark incorporation into C18:0. aiC15:0has been described as
being indicative for sulfate-reducing- (Ru ¨tters et al., 2002a)
and⁄or Gram-positive bacteria (Villanueva et al., 2004),
13C-bicarbonate-labeling experiments with the source
Fig. 2 (A) Oxygen microprofiles from source sediment and purple streamers
and (B) total sulfide (H2S + HS)+ S2)) microprofile from illuminated source
sediment. Oxygen concentration remains below detection limit (1 lM) even in
the illuminated green biofilm present in the 0–1 mm depth surface layer of
the source sediment, indicating that oxygenic photosynthesis does not occur.
The dip in the sulfide profile at the position of the green biofilm in the source
sediment indicatessulfideconsumption, most likelybyphototrophs.
8 S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
suggesting chemoautotrophy or cross-feeding by SRB and⁄or
Gram-positive bacteria. Another possibility is the uptake of
CO2by heterotrophs viaanapleroticreactions.
The13C-bicarbonate incubations with the purple streamers
yielded a total uptake into fatty acids up to 100 times higher
than for the incubations with source sediment (Fig. 3B).
This higher uptake is most likely due to the higher contribu-
tion of active biomass relative to thetotal sampleweightin the
purple streamers compared to the source sediment. Highest
uptake was found for VIS-only incubations again, reaching
2200 ng g)1sediment (dw) in total, followed by ambient
light incubation with 1700 ng and NIR-only with 1600 ng
g)1sediment (dw). The comparable values reached in all three
light treatments in thepurple streamers indicate a more impor-
tant role of photoautotrophy by PSB⁄CLB⁄GSB compared to
the source sediment. Uptake of inorganic carbon in these
experiments was mainly channeled into C18:1and C16:0, two
fatty acids with low chemotaxonomic specificity. The fatty
acids aiC17:0and C17:0were highly labeled in the VIS-only
treatment. These fatty acids are mainly described as being
indicative for SRB (e.g. Bu ¨hring et al., 2005) or other hetero-
trophs (Boschker & Middelburg, 2002), and the specific
uptake of label under light (especially strong under VIS-only)
conditions into this compound suggests a fast transfer of label
intoSRB,possiblyvia cyanobacterial exudates.C16:1exhibited
highest incorporation in the NIR-only treatment, which indi-
cates a predominant uptake by PSB or CLB, as both groups
are known to produce high amounts of monounsaturated C16
fatty acids (Kenyon & Gray, 1974; Pfennig et al., 1997). The
lower uptake into this compound under ambient light condi-
tions may be due to light inhibition. However, the differences
were less pronounced in the purple streamers compared to the
source sediment, indicating an adaptation of this community to
higher light levels, which is in line with their exposed position
at the wall of the source. Uptake under dark conditions was
negligible indicating the virtual absence of chemoautotrophic
activity in the purple streamers. This was somewhat unex-
pected as some PSB are known to be able to switch between
photoautotrophy and chemoautotrophy (Overmann & Pfen-
nig, 1992; Thar & Ku ¨hl, 2001). However, chemoautotrophy
in the dark by PSB is coupled to the presence of oxygen
(Overmann & Pfennig, 1992; Thar & Ku ¨hl, 2001), which is
absent in this case. It further suggests that the detected epsi-
lonproteobacteria dominating the 16S rRNA clone library
(Fig. 1) may not grow chemoautotrophically under the
provided conditions, possibly due to a shortage of suitable
Fig. 3 Total uptake into fattyacids afterincubationwith13C-bicarbonate;(A) sourcesediment and (B)purplestreamers (n.d.,notdetected).
Carbon cycleinsulfur-rich spring9
? 2011 Blackwell Publishing Ltd
electron acceptors (e.g. O2, NO32), S0, Fe3+). It could further
indicate that the activities of chemoautotrophic microbes,
such as the epsilonproteobacteria, are directly coupled to the
activities of phototrophs that produce suitable electron accep-
tors, e.g. S0andsulfate.
13C-acetate-labeling experiments under ambient light and
dark conditions were performed as well and results are sum-
marized in Fig. 4. Experiments with the source sediment
(Fig. 4A) under light conditions showed a total uptake of
135 ng g)1, which was more than 16 times higher than under
dark conditions [8 ng g)1sediment (dw)], providing a clear
indication for photoheterotrophic activity. Under light condi-
tions, uptake was mainly into C14:0and C16:1, followed by
C16:0. Most of the uptake under dark conditions was into
C16:1. C14:0was also labeled using
light conditions, indicating an origin either from different
phototrophic organisms or from a single organism that can
perform photoautotrophy as well as photoheterotrophy, a
phenomenon known for CLB (Van Der Meer et al., 2007)
ble of complementary uptake of organic molecules (Church
et al., 2006), and Paerl et al. (1993) suggested photohetero-
trophy in microbial mats as a supplementary mechanisms for
photosynthesis to help assimilate trace concentrations of
DOM in oligotrophic waters. The strong uptake of label from
acetate into C14:0under light conditions, a fatty acid that
became predominantly labeled in the bicarbonate experiment
under VIS-only conditions, may indicate a cyanobacterial
origin and therewith point to photoheterotrophic activity by
cyanobacteria. Some acetate could also have been respired to
CO2 and subsequently been incorporated by autotrophs,
although this is likely to be minimal within the timeframe of
the incubation. In this context, it is noteworthy that photo-
trophic purple non-sulfur bacteria that are able to utilize
acetate have been found to use the Calvin cycle to re-fix most
of the CO2being produced during acetate oxidation to main-
tain redox balance (McKinlay & Harwood, 2010). Similar
processes might apply to Zodletone spring, which would fur-
ther minimize distribution of the label to other community
members (see also below).
13C-acetate uptake of the purple streamers under light con-
ditions was 10 times higher than under dark conditions
[37 800 ng g)1compared with 3800 ng g)1sediment (dw),
respectively]. For both experiments, uptake was mainly in the
Fig. 4 Total uptake into fattyacids afterincubationwith13C-acetate;(A)source sediment and(B)purplestreamers (n.d.,notdetected).
10S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
taxonomically unspecific fatty acids C18:1and C16:0(Fig. 4B).
Several scenarios could explain the relatively high acetate
uptake in the purple streamers under dark condition. On one
hand, acetate uptake indicates an active heterotrophic com-
munity, which could be due to anaerobic heterotrophs, like
SRB. However, SRB as the main drivers of heterotrophic
uptake are not very likely, because relatively low uptake was
observed for SRB-specific fatty acids, like branched C15:0and
C17:0fatty acids. By contrast, uptake occurred mainly into
fatty acids that were also labeled under light conditions. Thus,
the labeling of C18:1and C16:0probably reflects heterotrophic
activity of phototrophic organisms in the dark performing
either fermentation or anaerobic respiration. In addition,
re-fixing of CO2being produced during acetate metabolism
couldhave taken place(McKinlay& Harwood, 2010).
In Zodletone spring, we found indications for a syntrophically
operating phototrophic community with potentially ancient
roots. Thus, the results of this study will contribute not only
to a better understanding of the functioning of this particular
system but also to similar systems that might have existed in
Earth’s history. The bicarbonate-labeling experiments clearly
indicate photoautotrophy in the source and the purple stream-
ers as the dominant processes. The different light treatments
revealed the overall importance of cyanobacteria in this pro-
cess. However, the NIR-only illumination showed consider-
able uptake as well. Principally, this experimental setup should
favor photoautotrophy of PSB, CLB and GSB (e.g. Bachar
et al., 2007). For Chloroflexus, the potential for autotrophy
has previously been shown (Van Der Meer et al., 2005), but
only during low-light conditions (e.g. early morning).
Another possibility is that cyanobacteria may be able to use
light in the NIR range for autotrophic growth, as cyanobacte-
ria growing under anoxygenic conditions have been shown to
uselight with a wavelength up to 720 nm (Oren et al., 1977).
Possibly cyanobacteria are able to channel light with longer
(<700 nm) wavelength to PS I under high sulfide concentra-
tions. In addition, some VIS-light still passed the filter, and
could thus support some growth of cyanobacteria that are
adapted to low-light conditions.
The highly elevated uptake of acetate into lipids under light
conditions uncovered photoheterotrophy as one key driver of
this system. Interestingly, although total incorporation of
labeled acetate was substantially higher in the light in both
systems, much lower, but still significant incorporation into
the same type of lipids occurred under dark conditions. This
phenomenon points either to the incorporation of the
label into unspecific lipids of both phototrophic and non-
phototrophic community members, or dark incorporation of
acetate by heterotrophically growing phototrophic bacteria.
Photoheterotrophy and also heterotrophy in general is a phe-
nomenon often described for anoxygenic phototrophs (e.g.
Boomeret al.,2000;VanDerMeeret al.,2005,2007;Hana-
da & Pierson, 2006), and also for cyanobacteria (Stal, 1991;
Zotina et al., 2003). Indications for photoautotrophy by
PSB⁄CLB⁄GSB were found in both environments, but as the
totaluptake under NIR-only conditions in the purplestreamer
incubations was similar to uptake under conditions favoring
the cyanobacteria (VIS-only incubations), we can conclude
that carbon fixation by this group is of greater importance in
this habitat compared to the source sediment community. It is
also possible that PSB are still photosynthetically active due to
the presence of carotenoids that can function as antenna
pigments by channeling light in the VIS range to the photo-
synthetic reaction center. Another difference is that chemoau-
totrophy appears to be absent in the dark in the purple
streamers, most likely due to a shortage of suitable electron
acceptors. Also, crossfeeding by SRB via photosynthetic
exudates seems to be more pronounced in the purple stream-
ers, possibly due to close spatial associations between commu-
nity members. The microbial community of the purple
streamers also appeared to be adapted to higher light condi-
tions, indicated by a less pronounced drop in uptake under
ambient light conditions.
From the microsensor measurements, we could not
obtain evidence that the cyanobacteria produce oxygen in
this highly sulfidic environment, suggesting that cyanobacte-
ria as well as the anoxygenic phototrophs GSB⁄CLB⁄PSB
use sulfide as electron donor, which is abundantly present in
the spring water. The phototrophic oxidation of sulfide to
sulfur and ultimately sulfate further stimulates the activities
of SRB, sulfur-reducers and sulfur⁄thiosulfate-disproportio-
nationators, all of which have been detected in this spring
(Elshahed et al., 2003), completing an anaerobic sulfur
cycle. In addition, dark sulfur respiration by cyanobacteria
could also take place (Oren & Shilo, 1979). Under ambient
light conditions, the incubations were probably light
inhibited, a phenomenon also described by Garcia-Pichel &
Castenholz (1994). This is probably a result of our experi-
mental setup. We placed the organisms in glass vials on top
of the sediment, possibly resulting in over-exposure of the
low-light-adapted organisms, as evidenced by the migration
of the phototrophs into the sediment and on top of the
sediment surface during the day and upon darkening,
Finally, the Zodletone spring ecosystem can be viewed as a
modern analog of ancient anoxic microbial mat systems in
which cyanobacteria were present, but might not have
performed oxygenic photosynthesis. At present, it is not
known if these cyanobacteria actually do have the capacity for
oxygenic photosynthesis, not have had the ability in the first
place (like protocyanobacteria), or have secondarily lost the
ability, similar to what has recently been described for oceanic
N2-fixing cyanobacteria (Zehr et al., 2009). Mulkidjanian
et al. (2006) investigated several cyanobacterial strains for
their genome core and enzymes involved in photosynthesis
? 2011 Blackwell Publishing Ltd
and proposed that the first phototrophs were anaerobic ances-
tors of cyanobacteria (procyanobacteria), which conducted
anoxygenic photosynthesis using a PS I-like reaction center.
Themost ancientform ofanoxygenic photosynthesismaywell
have used sulfur compounds as electron donors for CO2fixa-
tion (Olson & Pierson, 1986), and also as electron acceptor
(elemental sulfur or partially oxidized sulfur compounds) for
dark⁄anoxic respiration for energy generation during the dark
period. The observation of active cyanobacteria in the pres-
ence of 10 mM sulfide is remarkable and will provide fertile
The authors thank J. S. Lipp, F. Schubotz and H. Fredricks
for helping with the HPLC measurements and interpretation
of mass spectral data, K. Savage for field work assistance,
C. Harms for help in the laboratory, and A. M. Spain for assis-
tance in sampling and anion analyses. We further thank
S. Golubic and two anonymous referees for useful comments
that significantly improved the manuscript. This work was
supported through NSF Grant MCB-0240683. S.I.B. was
funded by a Marie Curie Outgoing International Fellowship
(MOIF-CT-2004-509865) from the European Community
and a MARUM postdoctoral fellowship. S.M.S. also kindly
acknowledges support through a fellowship from the Hanse
Wissenschaftskolleg in Delmenhorst, Germany (http://
Arnon DI, Losada M,NozakiM,Tagawa K (1961)Photoproduction
ofhydrogen, photofixation ofnitrogen and a unifiedconceptof
photosynthesis. Nature 190, 601–606.
AsselineauJ (1991)Bacterial lipids containingamino acids orpeptides
linkedbyamide bonds. Fortschritte der Chemie Organischer
Naturstoffe 56, 1–85.
Bachar A,Omoregie E,DeWit R,JonkersHM(2007)Diversity and
function ofChloroflexus-likebacteriaina hypersaline microbialmat:
phylogenetic characterization and impact onaerobicrespiration.
Appliedand Environmental Microbiology 73, 3975–3983.
BarridgeJK, Shively JM(1968) Phospholipids ofthiobacilli. Journal
of Bacteriology 95, 2182.
Bekker A,HollandHD, WangPL, RumbleD,SteinHJ,HannahJL,
Coetzee LL, BeukesNJ(2004) Dating the rise ofatmospheric
BenningC,HuangZH, GageDA(1995) Accumulation ofanovel
glycolipidanda betainelipidincellsofRhodobacter sphaeroides
grownunder phosphatelimitation. Archives of Biochemistry and
Biophysics 317, 103–111.
Beukes N (2004) Biogeochemistry – earlyoptionsinphotosynthesis.
Bianchi G,AvatoP,Scarpa O,MurelliC,AudisioG,RossiniA(1989)
Compositionand structureofmaizeepicuticular wax esters. Phyto-
chemistry 28, 165–171.
Boomer SM, PiersonBK, Austinhirst R,CastenholzRW(2000)
ments fromalkaline hot springsinYellowstoneNationalPark.
Archives of Microbiology 174, 152–161.
Borrego CM, GarciagilLJ(1994) Separationofbacteriochlorophyll
homologs fromgreenphotosyntheticsulfur bacteriabyreversed-
phaseHPLC. Photosynthesis Research 41, 157–164.
Boschker HTS,MiddelburgJJ(2002) Stable isotopesandbiomarkers
inmicrobialecology. FEMS Microbiology Ecology 40, 85–95.
BrettMT,Mu ¨ller-Navarra DC(1997) The roleofhighlyunsaturated
fatty acidsinaquaticfoodwebprocesses. Freshwater Biology 38,
Brocks JJ, Love GD, Summons RE, Knoll AH, Logan GA,
Bowden SA (2005) Biomarker evidence for green and purple
sulphur bacteria in a stratified Palaeoproterozoic sea. Nature
Bu ¨hring SI,Elvert M,WitteU (2005) The microbialcommunity
structure ofdifferentpermeable sandysediments characterisedby
the investigation ofbacterial fattyacids andfluorescence insitu hy-
bridisation.Environmental Microbiology 7, 281–293.
Campbell BJ, Engel AS,PorterML, TakaiK(2006)Theversatileepsi-
lon-proteobacteria: key players insulphidic habitats. Nature
Reviews Microbiology 4, 458–468.
CanfieldDE(1998) A newmodelforProterozoic oceanchemistry.
Nature 396, 450–453.
Castenholz RW, JørgensenBB, Damelio E,BauldJ (1991)Photosyn-
thetic and behavioralversatilityofthe cyanobacteriumOscillatoria
boryana ina sulfide-richmicrobialmat.FEMS Microbiology Ecology
Caumette P(1986) Phototrophic sulfur bacteriaand sulfate-reducing
bacteriacausingred waters ina shallowbrackischcoastal lagoon
(PrevostLagoon, France). FEMS Microbiology Ecology 38, 113–
ChurchMJ, DucklowHW,Letelier RM, Karl DM(2006) Temporal
tionin the subtropicalNorthPacificOcean. Aquatic Microbial
Ecology 45, 41–53.
CohenY,JørgensenBB, Padan E,Shilo M (1975)Sulfide-dependent
ca. Nature 257, 489–492.
CohenY,JørgensenBB, Revsbech NP, PoplawskiR (1986) Adapta-
tiontohydrogen-sulfide ofoxygenicand anoxygenicphotosynthe-
sisamongcyanobacteria. Applied and Environmental Microbiology
CraigH(1957) Isotopicstandardsforcarbonand oxygenandcorrec-
Geochimicaet Cosmochimica Acta 12, 133–149.
Crowe SA, Jones C, Katsev S, Magen C, O’Neill AH, Sturm A,
Canfield DE, Haffner GD, Mucci A, Sundby B, Fowle DA
(2008) Photoferrotrophs thrive in an Archean Ocean analogue.
Proceedings of the National Academy of Sciences of the USA 105,
Dembitsky VM(1996) Betaineether-linkedglycerolipids:chemistry
andbiology. Progress in Lipid Research 35, 1–51.
Desiervo AJ, ReynoldsJW(1975) Phospholipidcompositionand
cardiolipinsynthesis infermentativeand nonfermentativemarine
bacteria. Journal of Bacteriology 123, 294–301.
sity:whyaretheresomanylipids?Annual Reviewof Biochemistry
Elshahed MS, SenkoJM, NajarFZ, KentonSM,Roe BA, DewersTA,
SpearJR, KrumhozLR(2003) Bacterial diversityand sulfurcycling
ina mesophilicsulfide-richspring. Applied and Environmental
Microbiology 69, 5609–5621.
Elshahed MS, Najar FZ, Roe BA, Oren A,Dewers TA, KrumholzLR
(2004) Survey ofarchaealdiversityrevealsanabundanceofhalo-
12 S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd
philicArchaea ina low-salt,sulfide-andsulfur-rich spring.Applied
and Environmental Microbiology 70, 2230–2239.
ofspecificmembrane fattyacidsaschemotaxonomic markersfor
ane.Geomicrobiology Journal 20, 403–419.
ErtefaiTF, Fisher MC, FredricksHF, LippJS,PearsonA, Birgel D,
UdertKM,CavanaughCM, GschwendPM, HinrichsKU(2008)
Vertical distributionofmicrobial lipids and functionalgenes in
chemicallydistinctlayers ofa highly polluted meromicticlake.
Organic Geochemistry 39, 1572–1588.
FangJ,Barcelona MJ, NogiY,Kato K(2000a)Biochemicalimplica-
tionsandgeochemicalsignificance ofnovelphospholipids ofthe
extremelybarophilicbacteriafrom the MarianasTrench at
11,000 m.Deep Sea Research Part I 47, 1173–1182.
FangJS,Barcelona MJ, SemrauJD(2000b) Characterizationofmet-
hanotrophicbacteria onthe basisofintact phospholipidprofiles.
Fems Microbiology Letters 189, 67–72.
FischerWW(2008) Biogeochemistry– life beforetherise ofoxygen.
Garcia-Pichel F, Castenholz RW (1994) On the significance of solar
ultraviolet radiation for the ecology of microbial mats. In Microbial
Mats. Structure, Development and Environmental Significance (eds
Stal LJ, Caumette P). Springer Verlag, Heidelberg, pp. 77–84.
GoldfineH(1984) Bacterial-membranes and lipidpacking theory.
Journalof Lipid Research 25, 1501–1507.
Phospholipidsofhyphomicrobia. Journal of Bacteriology 95, 367.
Grotzinger JP,Knoll AH(1999)StromatolitesinPrecambriancar-
bonates:evolutionary mileposts orenvironmentaldipsticks?
Annual Review of Earth and Planetary Sciences 27, 313–358.
Hanada S,PiersonBK(2006) The family Chloroflexaceae. InThe
Prokaryotes: An Evolving Electronic Resource for the Microbiological
Community. (ed Dworkin M). Springer, Heidelberg, Germany,
Herman EK, KumpLR(2005) Biogeochemistryofmicrobial mats
under Precambrianenvironmental conditions: a modellingstudy.
Geobiology 3, 77–92.
Heubeck C (2009) An early ecosystem of Archean tidal microbial
mats (Moodies Group, South Africa, ca. 3.2 Ga). Geology 37,
HoelzlG,DoermannP (2007)Structure and function,of glycoglycer-
olipidsinplantsandbacteria. Progress in Lipid Research46, 225–
ImhoffJF,Bias-ImhoffU (1995) Lipids,quinones and fattyacidsof
anoxygenicphototrophicbacteria. InAnoxygenic Photosynthetic
Bacteria (eds Blankenship RE, Madigan MT, Bauer CE). Kluwer
AcademicPublishers, Dordrecht,The Netherlands,pp. 179–205.
Johnston DT, Wolfe-SimonF,PearsonA,KnollAH(2009) Anoxy-
Earth’smiddleage. Proceedings of the National Academy of Sciences
of the USA 106, 16925–16929.
JonkersHM, Ludwig R,DeWit R,Pringault O,Muyzer G,Niemann
H, FinkeN,DeBeer D (2003)Structuralandfunctionalanalysis of
a microbial mat ecosystem from a uniquepermanenthypersaline
BatesS,Anbar A,ArnoldGL, GarvinJ,Buick R(2007) Late
Science 317, 1900–1903.
KenyonCN, Gray AM(1974)Preliminary analysis oflipids and fatty
acidsofgreenbacteriaand Chloroflexus aurantiacus. Journalof
Bacteriology 120, 131–138.
Koga Y,NakanoM (2008) A dendrogramofarchaea basedonlipid
componentpartscomposition and itsrelationshiptorRNAphylog-
eny.Systematic and AppliedMicrobiology 31, 169–182.
Koga Y,MoriiH,Akagawa-MatsushitaM,Ohga I (1998) Correlation
ofpolar lipidcompositionwith 16SrRNAphylogenyinmethano-
gens. Further analysisoflipidcomponentparts. Bioscience Biotech-
nology and Biochemistry 62, 230–236.
KumpLR(2008) The rise ofatmospheric oxygen. Nature 451,
butionofArchaea toextant biomassinmarine subsurface sedi-
ments.Nature 454, 991–994.
Lopez-Lara IM,SohlenkampC,Geiger O (2003) Membrane lipids in
plant-associatedbacteria: their biosynthesesandpossiblefunctions.
Molecular Plant–Microbe Interactions 16, 567–579.
thesis-controlledcalcification ina hypersalinemicrobial mat.
Limnology and Oceanography 50, 1836–1843.
Makula RA(1978)Phospholipid composition ofmethane-utilizing
bacteria. Journal of Bacteriology 134, 771–777.
Makula RA, FinnertyWR (1975)Isolation andcharacterizationofan
ornithine-containinglipidfromDesulfovibriogigas. Journal of Bac-
teriology 123, 523–529.
McKinlay JB,HarwoodCS(2010) Carbondioxidefixationasa
centralredoxcofactor recycling mechanisminbacteria.
Proceedings of the National Academy of Sciencesof the USA 107,
MiddelburgJJ, Barranguet C,Boschker HTS,Herman PMJ,Moens
T,Heip CHR (2000)The fate ofintertidal microphytobenthos:an
insitu13Clabellingstudy. Limnology and Oceanography 45, 1224–
arithmicgrowthphase ofMethanobacterium thermoautotrophicum
incompensationforthe decreaseofdietherlipids.FEMS Microbiol-
ogy Letters 109, 283–287.
Mulkidjanian AY, KooninEV, Makarova KS, MekhedovSL,Sorokin
A,Wolf YI,Dufresne A,Partensky F,BurdH, KaznadzeyD,Has-
elkornR,GalperinMY(2006) The cyanobacterial genomecore
andthe originofphotosynthesis. Proceedings of the National Acad-
emy of Sciences of the USA 103, 13126–13131.
Noffke N, Beukes N, Bower D, Hazen RM, Swift DJP (2008) An
actualistic perspective into Archean worlds-(cyano-)bacterially
induced sedimentary structures in the siliciclastic Nhlazatse
Section, 2.9 Ga Pongola Supergroup, South Africa. Geobiology 6,
OlsonJM, PiersonBK(1986)Photosynthesis3.5thousand million
yearsago. Photosynthesis Research 9, 251–259.
OrenA,Shilo M (1979) Anaerobicheterotrophicdarkmetabolism in
the cyanobacteriumOscillatorialimnetica– sulfurrespirationand
lactate fermentation.Archives of Microbiology 122, 77–84.
OrenA,PadanE,AvronM (1977)Quantumyieldsfor oxygenicand
ca. Proceedings of the National Academy of Sciencesof the USA 74,
OvermannJ,PfennigN (1992)Continuous chemotrophicgrowth
tions. Archives of Microbiology 158, 59–67.
Carboncycle insulfur-richspring 13
? 2011 Blackwell Publishing Ltd
PaerlHW, Bebout BM, Joye SB, MaraisDJD (1993) Microscalechar- Download full-text
acterizationofdissolvedorganicmatter production and uptake in
marine microbialmat communities.Limnology and Oceanography
perigen nov,spec nov,a newphototrophic Proteobacterium ofthe
alpha group.Archives of Microbiology 168, 39–45.
Kranendonk Mj(2007) EarlyArchaeanmicroorganisms preferred
elementalsulfur, not sulfate. Science 317, 1534–1537.
RasmussenB,FletcherIR,Brocks JJ, KilburnMR(2008)Reassessing
the firstappearanceofeukaryotesandcyanobacteria. Nature 455,
ReinhardCT, Raiswell R,ScottC,Anbar AD, Lyons TW(2009) A
the continents. Science 326, 713–716.
RosselPE, LippJS,FredricksHF, Arnds J,BoetiusA,Elvert M,
HinrichsKU(2008) Intact polar lipids ofanaerobicmethanotroph-
icarchaea and associatedbacteria. Organic Geochemistry 39, 992–
Ru ¨ttersH,SassH,Cypionka H,Rullko ¨tter J (2001) Monoalkylether
phospholipidsin the sulfate-reducingbacteriaDesulfosarcinavaria-
bilisandDesulforhabdus amnigenus. Archives of Microbiology 176,
Ru ¨ttersH,SassH,Cypionka H,Rullko ¨tter J (2002a)Microbial com-
munities ina WaddenSeasediment core– clues from analyses of
intact glyceridelipids,and releasedfattyacids. Organic Geochemis-
try 33, 803–816.
Ru ¨ttersH,SassH,Cypionka H,Rullko ¨tter J (2002b)Phospholipid
analysis asa tool tostudycomplexmicrobial communities inmarine
sediments.Journal of Microbiological Methods 48, 149–160.
SchubotzF, WakehamSG, Lipp JS, FredricksHF, HinrichsKU
(2009)Detection ofmicrobial biomassbyintactpolar membrane
lipidanalysisin the watercolumnandsurfacesedimentsofthe Black
Sea. Environmental Microbiology 11, 2720–2734.
ScottC,Lyons TW, BekkerA, ShenY,PoultonSW, Chu X,Anbar
ocean. Nature 452, 456–459.
SenkoJM, Campbell BS, Henriksen JR, ElshahedMS, DewersTA,
KrumholzLR(2004) Baritedepositionresulting from phototroph-
icsulfide-oxidizingbacterial activity. Geochimicaet Cosmochimica
Acta 68, 773–780.
SessionsAL, Doughty DM, WelanderPV, Summons RE, Newman
DK(2009)The continuingpuzzleofthe greatoxidationevent.
Current Biology 19, R567–R574.
Shen Y, Buik R, Canfield DE (2001) Isotopic evidence for microbial
sulphate reduction in the early Archaean era. Nature 410,
Shively JM, Knoche HW (1969) Isolation of an ornithine-contain-
ing lipid from Thiobacillus thiooxidans. Journal of Bacteriology 98,
Short SA, White DC, AleemMIH (1969)Phospholipid metabolism
inFerrobacillus ferrooxidans. Journal of Bacteriology 99, 142.
StalLJ(1991) The metabolicversatilityofthe mat-building
matsand othercommunities. New Phytologist131, 1–32.
Sturt HF, SummonsRE, SmithK,ElvertM,HinrichsKU(2004)
Intact polar membrane lipids inprokaryotes and sediments deci-
ionizationmultistagemassspectrometry – newbiomarkersforbio-
geochemistry and microbialecology. Rapid Communications in
Mass Spectrometry 18, 617–628.
Summons RE, JahnkeLL,HopeJM,Logan GA(1999) 2-Methylh-
opanoidsasbiomarkersforcyanobacterial oxygenic photosynthesis.
Nature 400, 554–557.
TharR,Ku ¨hlM (2001) Motility ofMarichromatiumgracilein
response tolight,oxygen, andsulfide. Appliedand Environmental
Microbiology 67, 5410–5419.
Tice MM, LoweDR(2004) Photosyntheticmicrobial matsinthe
3,416-Myr-oldocean. Nature 431, 549–552.
Tice MM, Lowe DR (2006) Hydrogen-based carbon fixation in
the earliest known photosynthetic organisms. Geology 34, 37–
Van Der Meer MTJ, Schouten S, Bateson MM, Nu ¨bel U, Wieland
A, Ku ¨hl M, De Leeuw JW, Damste JSS, Ward DM (2005) Diel
variations in carbon metabolism by green nonsulfur-like bacteria
in alkaline siliceous hot spring microbial mats from Yellowstone
National Park. Applied and Environmental Microbiology 71,
VanDer MeerMTJ,Schouten S,Damste JSS, WardDM(2007)
andgreennon-sulfur-likebacteriainhabitinga microbialmat from
analkalinesiliceoushot spring inYellowstoneNationalPark
(USA).Environmental Microbiology 9, 482–491.
Villanueva L, NavarreteA, Urmeneta J,White DC, GuerreroR
(2004) Combinedphospholipidbiomarker-16SrRNA genedena-
turing gradientgelelectrophoresisanalysisofbacterial diversity and
physiologicalstatus inanintertidalmicrobial mat. Appliedand
Environmental Microbiology 70, 6920–6926.
WadaH,MurataN (1998)Membrane lipids inCyanobacteria.
InLipids inPhotosythesis: Structure,Function andGenetics (eds
Siegenthaler PA, Murata N). Kluwer Academic Publishers,
Dordrecht⁄Boston⁄London, pp. 65–81.
WadaH,MurataN (2007)The essential roleofphosphatidylglycerol
inphotosynthesis.Photosynthesis Research 92, 205–215.
Waldbauer JR, ShermanLS,SumnerDY,Summons RE(2009) Late
Archeanmolecular fossils fromthe Transvaal Supergrouprecord
the antiquity ofmicrobial diversityand aerobiosis. Precambrian
Research 169, 28–47.
ZehrJP, BenchSR,CarterBJ,Hewson I,Niazi F,Shi T,TrippHJ,
Affourtit JP(2009) Globallydistributeduncultivatedoceanic
N2-fixing cyanobacterialackoxygenicphotosystems II. Science
ZerkleAL, KamyshnyA,KumpLR, Riccardi AL, ArthurMA,Farqu-
harJ (2009) Biogeochemical sulfurcyclinginmeromictic Fayette-
ville GreenLake,NY. Geochimicaet Cosmochimica Acta 73,
ZotinaT,Koster O,Juttner F(2003) Photoheterotrophy andlight-
dependentuptake oforganicand organicnitrogenouscompounds
byPlanktothrixrubescens under low irradiance. Freshwater Biology
14S. I. BU¨HRING et al.
? 2011 Blackwell Publishing Ltd