2014-EMI12595

Full text

Distribution of ether lipids and composition of the
archaeal community in terrestrial geothermal springs:
impact of environmental variables
Wei Xie,1,2 Chuanlun L. Zhang,1,2* Jinxiang Wang,2
Yufei Chen,1Yuanqing Zhu,1,3 José R. de la Torre,4
Hailiang Dong,5,6 Hilairy E. Hartnett,7,8
Brian P. Hedlund9and Martin G. Klotz10**
1State Key Laboratory of Marine Geology, Tongji
University, Shanghai 200092, China.
2Department of Marine Sciences, University of Georgia,
Athens, GA 30602, USA.
3Seismological Bureau of Shanghai, Shanghai 201203,
China.
4Department of Biology, San Francisco State University,
San Francisco, CA 94132, USA.
5State Key Laboratory of Biogeology and Environmental
Geology & Institute of Earth Sciences, China University
of Geosciences, Beijing 100083, China.
6Department of Geology and Environmental Earth
Science, Miami University, Oxford, OH 45056, USA.
7School of Earth and Space Exploration and
8Department of Chemistry and Biochemistry, Arizona
State University, Tempe, AZ 85287, USA.
9School of Life Sciences, University of Nevada Las
Vegas, Las Vegas, NV 89154, USA.
10Department of Biological Sciences, University of North
Carolina, Charlotte, NC 28223, USA.
Summary
Archaea can respond to changes in the environment
by altering the composition of their membrane lipids,
for example, by modification of the abundance and
composition of glycerol dialkyl glycerol tetraethers
(GDGTs). Here, we investigated the abundance and
proportions of polar GDGTs (P-GDGTs) and core
GDGTs (C-GDGTs) sampled in different seasons
from Tengchong hot springs (Yunnan, China), which
encompassed a pH range of 2.5–10.1 and a tempera-
ture range of 43.7–93.6°C. The phylogenetic com-
position of the archaeal community (reanalysed
from published work) divided the Archaea in spring
sediment samples into three major groups that corre-
sponded with spring pH: acidic, circumneutral
and alkaline. Cluster analysis showed correlation
between spring pH and the composition of P- and
C-GDGTs and archaeal 16S rRNA genes, indicating an
intimate link between resident Archaea and the distri-
bution of P- and C-GDGTs in Tengchong hot springs.
The distribution of GDGTs in Tengchong springs was
also significantly affected by temperature; however,
the relationship was weaker than with pH. Analysis of
published datasets including samples from Tibet,
Yellowstone and the US Great Basin hot springs
revealed a similar relationship between pH and GDGT
content. Specifically, low pH springs had higher
concentrations of GDGTs with high numbers of
cyclopentyl rings than neutral and alkaline springs,
which is consistent with the predominance of
high cyclopentyl ring-characterized Sulfolobales and
Thermoplasmatales present in some of the low pH
springs. Our study suggests that the resident
Archaea in these hot springs are acclimated if not
adapted to low pH by their genetic capacity to effect
the packing density of their membranes by increasing
cyclopentyl rings in GDGTs at the rank of community.
Introduction
Archaea constitute one of the three domains of life on
Earth (Woese et al., 1990). A large number of archaeal
species produce specialized cytoplasmic mem branes
composed of isoprenoid glycerol dialkyl glycerol tetrae-
thers (GDGTs) (see Fig. S1 for structures of major
GDGTs), which resist delamination and protect against
proton permeability (van de Vossenberg et al., 1998;
Macalady et al., 2004; Valentine, 2007). The presence of
polar GDGTs (P-GDGTs) is assumed to reflect living
organisms (e.g. Petsch et al., 2001; Biddle et al., 2006;
Lipp et al., 2008) because, upon cell death, the P-GDGTs
may rapidly degrade into more recalcitrant core GDGTs
(C-GDGTs) by cleavage or degradation of the polar head
groups similar to phospholipid fatty acids (White et al.,
1979; Harvey et al., 1986). However, a later report
Received 7 January, 2014; accepted 8August, 2014. For correspond-
ence. *E-mail archaea.zhang@gmail.com; Tel. 86-21-65982012;
Fax 86-21-65988808. **E-mail martin.g.klotz@gmail.com; Tel.
0017046875465; Fax 0017046871488.
bs_bs_banner
Environmental Microbiology (2014) doi:10.1111/1462-2920.12595
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd

indicated that P-GDGTs can be persistent under certain
conditions (Logemann et al., 2011). In contrast, C-GDGTs
can be preserved in the geological record for more than
120 million years (e.g. Kuypers et al., 2001). Another
investigation revealed a similarly extended persistence
of long-chain glycolipids from heterocystous N2-fixing
cyanobacteria (Bauersachs et al., 2010).
The composition of GDGTs is affected by several
physico-chemical factors, which have been investigated
under laboratory conditions and in natural environments.
Because the packing density of the archaeal membrane is
affected by and, potentially, reflective of their growth tem-
perature or pH, the average number of cyclopentane moi-
eties per GDGT lipid (the ‘ring index’, RI) has been used
as a proxy to reflect temperature or pH (Pearson et al.,
2004; Schouten et al., 2007; Boyd et al., 2011). Studies
with Sulfolobus solfataricus (De Rosa et al., 1980),
Thermoplasma acidiphilum (Uda et al., 2001) and
Acidilobus sulfurireducens (Boyd et al., 2011) revealed
that the ring indices of their core GDGTs increased
with incubation temperature. In natural environments,
however, chemical variables other than temperature or pH
appear to affect archaeal community structure and thus
the ring indices as well (Pearson et al., 2008; Li et al.,
2013; Paraiso et al., 2013); for example, the RI of
C-GDGTs was found to correlate with the concentration of
bicarbonate in some hot springs of the US Great Basin
(Pearson et al., 2004).
Tengchong, in central-western Yunnan Province,
Southwestern China, is known for its geothermal features
(Liao and Guo, 1986; Tong and Zhang, 1989; Wang
et al., 2000). The Rehai and Ruidian geothermal fields in
Tengchong are two localities of intense hydrothermal
activity with numerous springs and pools. A wide physico-
chemical diversity of springs (temperature to >97°C; pH
ranges from <1.8 to ≥9.3; granite hosted versus carbon-
ate hosted) provide numerous niches for extremophiles,
including thermophilic and thermoacidophilic Archaea.
These archaeal communities have been studied by tradi-
tional culturing approaches (reviewed in Hedlund et al.,
2012) as well as via culture-independent methods that
provide a deep census (Huang et al., 2010; Jiang et al.,
2010; Song et al., 2010; 2013; Pagaling et al., 2012; Hou
et al., 2013). The Rehai and Ruidian geothermal fields
provide an excellent opportunity to investigate the rela-
tionship between GDGTs, archaeal community com-
position and physico-chemical conditions. Recently, Wu
and colleagues (2013) reported on the distribution of
C-GDGTs from hot springs of Tengchong and other loca-
tions in Yunnan Province as a function of temperature and
pH; however, the study lacked data regarding the pres-
ence and distribution of P-GDGTs as well as phylogenetic
information on the resident archaeal community. This
limited set of data prevented the examination of whether
the resident archaeal community structure correlates
with the abundance and distribution of GDGTs. Further-
more, the global patterning of GDGTs and the underpin-
ning environmental and biological variables are not well
understood in hot springs.
In this study, we show that pH affects the distribution of
GDGTs in terrestrial hot springs at the global scale
whereas temperature affects ether lipid distribution only in
samples with similar phylogenetic composition and within
individual geographical regions. Dissolved oxygen
(limited to Tengchong dataset) also correlated with differ-
ences in lipid distribution, which may be attributed to
differential degradation of GDGTs under different redox
conditions. The lipid composition and the archaeal com-
munity structure correlated only at the local level. The
dominance of pH over temperature and other locally
relevant factors including oxygen concentration and
archaeal community structure is a significant improve-
ment upon our previous finding (Pearson et al., 2004;
2008; Zhang et al., 2006) and could provide additional
insight on the acclimation and, perhaps, adaption of
Archaea to the extreme environments of the geothermal
systems.
Results
Physical and chemical characteristics of sampling sites
The springs were divided into five major groups based on
visual appearance: (i) large pools with standing water,
which included Dagunguo (Dgg), Gongxiaoshe [including
GxsS (carbonate-rich sediment at bottom) and GxsB
(sinter sample chipped from the sides of the spring)] and
Jinze (Jz); (ii) high-discharge, fast-flowing springs with
small source pools, which included Gumingquan [includ-
ing GmqS (source), GmqC (channel) and GmqP
(streamer pool)] and Jiemeiquan [including JmqR (right
pool) and JmqL (left pool)]; (iii) shallow acidic pools, which
included the Diretiyan (Drty) pools and Zhenzhuquan
(Zzq); (iv) shallow springs with multiple sources, which
included Shuirebaozha (Srbz), Direchi (Drc), left and right
Wumingquan (WmqL, WmqR); and (v) small outflow
channels containing microbial mats at different locations,
which included the upper and lower parts of Shiziwei
(SzwU and SzwD respectively) and the upper, middle and
lower parts of Shiziyao (SzyS, SzyM and SzyD respec-
tively). Springs Gxs and Jz were located in the town of
Ruidian, while all others were located in the Rehai geo-
thermal field, which is about 40 km south of Ruidian (refer
to maps in Hou et al., 2013). The physico-chemical
parameters of all springs are listed in Table S1. The tem-
perature (43.7°C) of the Srbz spring was significantly
lower in July 2012 than it was in January (79.8°C) and
June (72.1°C) of 2011, which may be attributed to an
2W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

unusually shallow water table. Several hot springs, includ-
ing WmqL, WmqR and Shiziyao, were only sampled in
June 2011 because they were dry in January 2011 and
July 2012. In all, the temperatures and pH values were
relatively stable in other springs and ranged from 55.1°C
to 93.6°C and from 2.5 to 9.4 respectively. Soils near the
hot springs were collected in June 2011 to evaluate the
extent to which GDGTs in the hot springs were impacted
by run-off from surrounding soils.
Abundance of archaeal lipids in Tengchong hot springs
and surrounding soils
All hot spring samples contained measurable concentra-
tions of P- and/or C-GDGTs plus P-archaeol (40 out of 46
samples) and C-archaeol (44/46) (Table S1; see Fig. S1
for the structure of archaeol); the latter are mainly pre-
sent in the membranes of Euryarchaeota such as
methanogens. Absolute abundance of P-GDGTs ranged
from 0.4 to 15 912 ng lipid g−1dry mass. Concentra-
tions of C-GDGTs ranged from less than 0.7 to
10 292 ng lipid g−1dry mass. The levels of total P- and
C-GDGTs in hot spring samples from June 2011 were
correlated, as shown by a linear regression analysis of
log-transformed absolute abundances of the P- versus
C-GDGTs (R2=0.73; Fig. S2). To assess whether P-
and/or C-GDGTs in the hot springs may have been intro-
duced from the surrounding soil, pertinent soil samples
were collected in June 2011. While the (log-transformed)
concentrations of P and C-GDGTs in these soil samples
were positively correlated (R2=0.80; Fig. S2), the slope
and y-intercept were different from that of the hot springs
(Fig. S2). The average concentrations of P- and C-GDGTs
were about 22-fold and 25-fold higher, respectively, in hot
spring samples than in the pertinent surrounding soil
samples. In addition, the average concentrations of P-
and C-archaeol were about 16-fold and 13-fold higher,
respectively, in hot spring samples than in the soil
samples (Fig. S3).
Cluster analysis
Cluster analysis was performed to evaluate the distribu-
tion of P- and C-GDGTs from all three sampling periods.
For P-GDGTs, the samples could be clustered into
three major groups that corresponded to the pH rather
than temperature at the sample site (Fig. 1): group 1
(pH =2.8 ±0.3, T =69.3 ±12.2°C; n=11) was character-
ized by relatively high abundance of GDGT-4, GDGT-5
and GDGT-6; group 2 (pH =6.0 ±1.7, T =73.1 ±20.6°C;
n=9) was similar to group 1, but contained less GDGT-6
and GDGT-7 and more GDGT-0 and GDGT-1; and group
3 (pH =8.8 ±0.8, T =78.1 ±11.8°C; n=21) was charac-
terized by high abundance of GDGT-0, GDGT-1, GDGT-2
and GDGT-3 (Fig. 1). In addition, Gxs in January and
June 2011 were clustered together and were character-
ized by high abundance of crenarchaeol. For the
C-GDGTs, the samples also clustered into three major
groups that corresponded with pH and were similar
to clusters formed by analysis of P-GDGTs (Fig. S4):
group 1 (pH =3.0 ±0.6, T =70.7 ±15.2°C; n=14), group
2 (pH =7.5 ±0.8, T =68.1 ±15.2°C; n=8) and group
3 (pH =8.5 ±1.0, T =79 ±11.5°C; n=24). Most sites
sampled over multiple seasons clustered together,
indicating that the distribution of P- and C-GDGTs
in those springs was relatively stable across different
seasons.
While relative abundances of most individual P-GDGT
and C-GDGT values were highly correlated with each
other (Figs 2 and S5), one exception was the proportion of
P-GDGT-0 and C-GDGT-0, which was not significantly
correlated (R2=0.21; Fig. 2). In contrast to samples
with higher dissolved oxygen content (950 ±461 μgl
−1,
n=13), those with lower dissolved oxygen content
(222 ±118 μgl
−1,n=6) had higher relative abundance of
P-GDGT-0 compared with C-GDGT-0, suggesting that
degradation rates of P-GDGT-0 and/or C-GDGT-0 might
be different under very low oxygen conditions (Fig. S5),
although we note that dissolved oxygen concentrations in
these springs are virtually always substantially under
saturated.
Physical or chemical factors controlling the distribution
of P- and C-GDGTs
Redundancy analysis (RDA) of the data was performed to
examine whether the observed GDGT patterns correlated
significantly with environmental variables. RDA1 and
RDA2 explained 90% and 10%, respectively, of the vari-
ance in the archaeal P-GDGTs. The results of this analy-
sis (Fig. 3) showed that the pH vector was highly
significant and nearly parallel to RDA1, indicating that the
RDA1 axis correlated strongly with pH. RDA1 clearly dis-
tinguished the P-GDGT distribution in low pH samples
from those in neutral or high pH samples, and pH was
identified as the most significant environmental factor
(P=0.001). Low pH samples were close to the vectors
for P-GDGT-5 and -6, while neutral or high pH samples
corresponded with the vectors for P-GDGT-0, -1, -2, -3, -4
and crenarchaeol-isomer.
Temperature was identified as another significant
environmental factor (P=0.004) contributing to GDGT
distribution in hot springs (Fig. 3). However, the vector
representing temperature pointed close to RDA2, which
only contributed 10% to the distribution of P-GDGTs. The
results of RDA based on C-GDGT composition were
similar to that of P-GDGTs, indicating that pH is a domi-
nant factor controlling the distribution of P- and C-GDGTs
Archaeal lipids from terrestrial hot springs 3
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

in Tengchong hot springs (Fig. S6). Other factors that
were investigated, including dissolved oxygen, sulfide,
ammonia, nitrate plus nitrite, sulfate and ferrous iron, were
not significantly correlated (P>0.05) with the distributions
of P- or C-GDGTs.
Regression analyses of RI and environmental variables
Multiple regression analysis was performed to quantita-
tively investigate the relationship between P-RI and C-RI
values and environmental variables. Two equations,
including eight environmental variables, best described
the relationship between geochemical data and P-RI or
C-RI respectively. The root mean square errors of Eqs 1
and 2 are 0.19 and 0.29 respectively.
PRI pH T DO
SNH
-=− ×+ ×− ×
[]
+
×
[]
−×
[]
+
−+
41 037 002 024 005
004 0
24
.. . . .
...
..
,.,
003
0 0001 0 01
16 0 94
23
422
2
×+
[]
−×
[]
−×
[]
==
−−
−+
NO NO
SO FE
nR RMMSE P=<
()
019 001., .
(1)
CRI pH T DO
SNH
-=−×− ×+×
[]
−
×
[]
−×
[]
−
−+
4 69 0 3 0 001 0 32 0 01
001
24
.. . . .
.0007
0 0003 0 01
16 0 91
23
422
2
.
..
,.,
×+
[]
+×
[]
−×
[]
==
−−
−+
NO NO
SO FE
nR RMMSE P=<
()
029 001., .
(2)
The coefficients of regression between RI and pH
(0.37 for P-RI and 0.3 for C-RI) as well as RI and dis-
solved oxygen (0.24 for P-RI and 0.32 for C-RI) were
significantly higher than those for correlations of RI with
other environmental factors (<0.1 for both P- and C-RI)
in the equations. The correlation between P- and C-RI
and pH was significant (R2=0.87 and 0.89 respectively;
Fig. 4A), consistent with the cluster analyses and RDA.
The results also indicate that dissolved oxygen corre-
lates significantly with P- and C-RI, even when two
points with unusually high dissolved oxygen values,
Drc and Zzq in July 2012, were considered outliers
and thus excluded (R2=0.58 and 0.73 respectively;
Fig. 4B). The P- or C-RI showed a positive correlation
with dissolved oxygen values within the range of 140–
Fig. 1. Cluster analysis of P-GDGTs from
Tengchong hot spring sediments collected in
January 2011, June 2011 and July 2012.
Sample names are shown on the right of the
figure. GDGTs are colour coded and shown at
the bottom of the figure. Those samples are
clustered into three groups that correspond to
differences in pH: group 1 (pH =2.8 ±0.3,
T=69.3 ±12.2°C; n=11), group 2
(pH =6.0 ±1.7, T =73.1 ±20.6°C; n=9) and
group 3 (pH =8.9 ±0.7, T =78.6 ±12.3°C;
n=19). The samples in corresponding groups
are boxed with dash lines.
4W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

1000 μgl
−1. Inclusion of the two high value points pro-
duced relatively low P- or C-RI values, which sug-
gests that the impact of dissolved oxygen may not be
positive at high range or that these two measurements
were inaccurate, for example because of atmospheric
contamination of the samples. The absence of samples
with intermediate levels of dissolved oxygen made it
difficult to distinguish these two possibilities. Multiple
linear regression analysis was used to investigate the
relationships among pH, dissolved oxygen and P- or
C-RI, resulting in two multiple regression equations
shown in Fig. 4C and D respectively. The results of the
multiple regression analysis improved the determination
coefficient for P-RI (from 0.87 for pH and 0.58 for
dissolved oxygen to 0.93 for both pH and dissolved
oxygen) and for C-RI (from 0.89 for pH and 0.73 for
dissolved oxygen to 0.93 for both pH and dissolved
oxygen) in the linear regression analysis, indicating the
potential for multiple regression to reveal relationships
between GDGTs-based proxies and multiple environ-
mental variables.
Relationship between archaeal community structure and
the composition of lipids
To link the sources of GDGTs to the composition of
archaeal communities in Tengchong hot springs, we
re-analysed previously published pyrosequencing data
of amplified 16S rRNA gene fragments from samples
collected in January 2011 (Hou et al., 2013). Cluster
analysis based on taxonomic composition divided
samples into three major groups: group 1 (pH =3.9 ±2.1,
T=73.8 ±14.6°C; n=4) was characterized by samples
with low pH values and included Drty1, Drty2, Drty3 and
Zzq; group 2 (pH =8.8 ±0.8, T =85.6 ±5.8°C; n=8) was
characterized by neutral or high pH samples and included
JmqL, JmqR, GmqS, GmqC, GmqP, Dgg, SrbzU and
SrbzD; group 3 (pH =7±0.4, T =77 ±5.7°C; n=2)
included GxsS and Jz, which are located in Ruidian
(Fig. S7).
Archaea in group 1 springs (low pH) consisted mainly
of Sulfolobus (Fig. S7). Because there were just four
Sulfolobus-abundant samples, we could not derive equa-
tions with statistical significance for the regression of pH
and P-RI or C-RI values. However, the P-RI and C-RI
values from Zzq spring (P-RI =3.1, C-RI =3.4; pH 4.6;
temperature 90.7°C; Table S1) were lower than those
of the three Drty hot springs (P-RI =3.8 ±0.2, C-RI =
3.8 ±0.2; pHs around 2.6; temperature 55.1°C, 64.5°C
and 85.1°C; Table S1), despite the lower temperature
of the Drty samples, suggesting that low pH led to
increases in the number of cyclopentyl moieties in
Sulfolobus. Samples from group 2 (neutral or high
pH) were inhabited mainly by Desulfurococcales and
Thermoproteales (Fig. S7), and the pertinent P-RI
and C-RI values were positively correlated with
Fig. 2. X–Y diagrams of the relative abundance of individual P-
and C-GDGTs from hot spring samples with detected dissolved
oxygen (DO) concentrations in June 2011 and July 2012. The
coefficient of determination of individual P- and C-GDGT’s
relationship is shown on the right.
Fig. 3. RDA ordination plots for the first two principal dimensions of
the relationship between the distribution of P-GDGTs and relevant
environmental parameters from Tengchong in January 2011, June
2011 and July 2012. Correlations between environmental variables
and RDA axes are represented by the length and angle of dash
arrows (environmental factor vectors). Solid arrows stand for the
proportion of P-GDGTs.
Archaeal lipids from terrestrial hot springs 5
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

temperature (R2=0.86 and 0.85 for P-RI and C-RI
values respectively; Fig. S8). Samples from group 3
were characterized by high abundance of Thaumar-
chaeota or Aigarchaeota and were located in Ruidian
(Fig. S7). Ruidian hot springs have similar pHs to some
of the Rehai hot springs but are carbonate hosted and
have a distinct chemistry. The relative abundance of
Thaumarchaeota in the sediment from Gxs (GxsS) was
significantly higher than in the sinter from Gxs (GxsB)
(Fig. S7). Likewise, the P-crenarchaeol in GxsS was also
more abundant than in GxsB (Fig. 1), suggesting that the
Thaumarchaeota in the Gxs spring resided mainly in the
sediment rather than the sinter.
Comparison with other geothermal systems
Previously published data exist for the abundance and
distribution of GDGTs and corresponding 16S rRNA
pyrotag sequence data of samples from hot springs in
Tibet (Li et al., 2013; Wang et al., 2013), Yellowstone
National Park (YNP) (Boyd et al., 2013) and the US
Great Basin (Table S1; Cole et al., 2013), which were
re-analysed to compare the pertinent relationships
between archaeal community structure and distribution
of P-GDGTs in hot springs from different regions (Figs 5
and 6; Table S2). The phylogenetic pattern consisted
of six major groups: (i) group 1 (Thaumarchaeota/
Fig. 4. Simple linear regression plots for relationship (A) between P- or C-RI and pH (P-RI =4.39 −0.25 ×pH, R2=0.87;
C-RI =4.81 −0.32 ×pH, R2=0.89), (B) between P- or C-RI and dissolved oxygen [P-RI =2+1.77 ×(DO), R2=0.58;
C-RI =1.59 +2.53 ×(DO), R2=0.73]; curved surfaces show relationships (C) between P-RI and pH and dissolved oxygen
{P-RI =0.27 +0.65 ×pH +6.8 ×[DO] −0.04 ×[pH]2−2.6 ×[(DO)]2−0.8 ×pH ×[DO]} and (D) between C-RI and pH and dissolved oxygen
{C-RI =2.9 +0.25 ×pH +1×[DO] −0.04 ×[pH]2−0.3 ×[(DO)]2−0.02 ×pH ×[DO]}. Samples for analysis were collected in June 2011 and July
2012. January 2011 data are not included because there is no dissolved oxygen data. The determination coefficient and Pvalue are shown on
the top of the each diagram.
6W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Aigarchaeota-abundant group; n=16), (ii) group 2
(Desulfurococcales/Thermoproteales-abundant group;
n=13), (iii) group 3 (‘high-diversity’ group; n=11), (iv)
group 4 (Thermoplasmatales-abundant group; n=2),
(v) group 5 (YLA114-abundant group; n=2) and
(vi) group 6 (Sulfolobales-abundant group; n=4)
(Fig. 5). In contrast to Tengchong hot spring samples,
which were mainly clustered into a Desulfurococcales/
Thermoproteales-abundant group and a Sulfolobales-
abundant group, samples from Tibet, YNP and the US
Great Basin were mainly represented by the ‘high-
diversity’ group and the Thaumarchaeota/Aigarchaeota-
abundant group. On the other hand, clustering based on
archaeal lipid composition demonstrated that samples
from different regions clustered into three major groups
that corresponded with sample pH (Fig. 6): group 1
(pH =3.7 ±1.9; n=14), group 2 (pH =7.7 ±1.6; n=28)
and group 3 (pH =7.6 ±0.4, T =54.4 ±19.3°C; n=4).
The low pH samples (group 1) had high abundance of
GDGT-5 and GDGT-6, while most neutral-high pH
samples (group 2) had relatively high abundance of
GDGT-0, -1, -2, -3 and -4. Group 3 consisted of samples
with abundant crenarchaeol.
RDA of the expanded datasets showed that both pH
and temperature are significantly correlated with the
archaeal phylogenetic composition and corresponding
lipid distribution (Fig. 7A and B respectively). The RDA1 in
Fig. 7A and B explained 64% and 96% of the total vari-
ance respectively. Figure 7A shows that both pH and
temperature correlate significantly with the phylogenetic
composition of the archaeal community at the order level,
whereas Fig. 7B shows a dominant relationship between
pH and archaeal lipid content, with a weaker but still
significant correlation between temperature and archaeal
lipid content.
Spearman’s rho analyses were further used to in-
vestigate the relationships among pH, temperature,
archaeal composition and individual P-GDGTs. The
results (Table S3) indicated that both pH and tempera-
ture were statistically significantly correlated with four of
18 archaeal genera, which is consistent with the RDA
results, demonstrating that pH and temperature equally
contributed to the structure of the archaeal community in
those hot springs. On the other hand, pH was signifi-
cantly correlated with the abundance of all nine individual
P-GDGTs (positively correlated with GDGT-0, -1, -2, -3,
Fig. 5. Cluster analysis based on
phylogenetic composition of Archaea from
Tengchong, Tibet, YNP and the US Great
Basin hot springs at the taxonomic rank of
order. Sample names are shown on the right
of the figure. The orders are colour coded and
shown at the bottom of the figure. Those
samples are majorly clustered into six groups:
group 1 (pH =7.5 ±1.0, T =60.1 ±17.5°C;
n=16), group 2 (pH =7.9 ±1.9,
T=78.8 ±10.3°C; n=13), group 3
(pH =5.2 ±2.9, T =52.3 ±19.4°C; n=11),
group 4 (pH =2.8 ±0.3, T =30.9 ±11.8°C;
n=2), group 5 (pH =7.2 ±0.3,
T=44.5 ±2.1°C; n=2) and group 6
(pH =3.1 ±1.0, T =73.9 ±16.8°C; n=4). The
samples in corresponding groups are boxed
with dash lines. Those data were re-analysed
from Hou and colleagues (2013) (Tengchong),
Wang and colleagues (2013) (Tibet), Boyd
and colleagues (2013) (YNP) and Cole and
colleagues (2013) (US Great Basin). MBGA:
marine benthic group A; MCG: miscellaneous
crenarchaeotic group; pUWA2: an
environmental sequence from acidic hot
spring water (Takai and Sako, 1999); AcrA07:
an environmental sequence of halophilic
crenarchaeota from a hypersaline
endoevaporitic microbial mat (Sorensen et al.,
2005); YLA114: an environmental sequence
of Parvarchaea from Yellowstone Lake (Kan
et al., 2011).
Archaeal lipids from terrestrial hot springs 7
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

crenarchaeol and crenarchaeol-isomer and negatively
correlated with GDGT-4, -5 and -6), while temperature
just correlated with two of them (positively correlated with
GDGT-1, -2). The results of Spearman’s rho analyses
were consistent with the results of the RDA, demonstrat-
ing that pH and temperature equally contributed to the
structure of the archaeal community, but pH was the
dominant correlate with the distribution of P-GDGTs.
The Spearman’s rho analyses identified several statisti-
cally significant relationships between the abundance of
certain archaeal genera and P-RI. P-RI was positively
correlated with both presence and abundance of
Sulfolobales,Nitrososphaerales and pUWA2, and nega-
tively correlated with ArcA07 and YLA114.
Discussion
Origin of archaeal lipids in the hot springs
In situ production of archaeal GDGTs in hot springs is
supported by two lines of evidence. Firstly, although sig-
nificant correlations between P- and C-GDGTs were
found for both the hot spring and surrounding soil sites,
the slope and y-intercept were different. Secondly, the
mean concentration of both GDGTs and archaeol were
higher in hot spring sediments than surrounding soils,
suggesting that GDGT and archaeol structures identified
in the hot springs were predominantly autochthonous
rather than the result of exogenous input. The interpreta-
tion of these results is consistent with previous observa-
tions from hot springs in the US Great Basin (Pearson
et al., 2004; Zhang et al., 2006; Paraiso et al., 2013), YNP
(Boyd et al., 2013) and Tibet (He et al., 2012; Li et al.,
2013).
Relative importance of pH, temperature, dissolved
oxygen and community structure in shaping the
distribution of P- and C-GDGTs
The number of cyclopentyl rings in GDGTs, expressed in
the form of the RI, is known to correlate with temperature
(Uda et al., 2001; Boyd et al., 2011; Paraiso et al., 2013)
and pH (Boyd et al., 2011; 2013). In contrast, Pearson and
colleagues (2004) observed that bicarbonate, but not tem-
perature, correlated with the RI in alkaline hot springs in
the US Great Basin while Li and colleagues (2013)
reported that multiple environmental factors could signifi-
cantly impact the RI in Tibet. The analysis of GDGTs
Fig. 6. Cluster analyses of P-GDGTs from
Tengchong hot spring samples in January
2011, from Tibet, YNP and the US Great
Basin. Sample names are shown on the right
of the figure. GDGTs are colour coded and
shown at the bottom of the figure. Those
samples are majorly clustered into three
groups: group 1 (pH =3.7 ±1.9,
T=58.6 ±18.1°C; n=14), group 2
(pH =7.7 ±1.6, T =70.6 ±18.3°C; n=28) and
group 3 (pH =7.6 ±0.4, T =54.4 ±19.3°C;
n=4). The samples in corresponding groups
are boxed with dash lines.
8W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

presented here revealed that the composition of P- and
C-GDGTs was correlated primarily with pH on a global
scale and secondarily correlated with temperature, dis-
solved oxygen and archaeal community structure (see
details below).
pH. The negative correlation between pH and RI was
revealed by RDA and multiple regression analysis for
Tengchong hot spring samples and is consistent with pre-
vious reports on pure cultures (Macalady et al., 2004;
Boyd et al., 2011) and samples from natural environments
(Pearson et al., 2008; Boyd et al., 2013). On the other
hand, our results are in contrast with published work on
the euryarchaeon T. acidiphilum (Shimada et al., 2008)
and with results from hot springs in Tibet (Li et al., 2013).
One reason for the discrepancy between these studies
may be that different Archaea respond to environmental
cues with different changes in the composition of their
lipids. This interpretation is supported by different relation-
ships between GDGT compositions and environmental
factors in Tengchong harbouring different Archaea.
For example, the P-RI and C-RI values in the Desul-
furococcales- and Thermoproteales-dominated group
were correlated with temperature (Fig. S8), whereas the
P-RI and C-RI values in the Sulfolobales-dominated
group were not correlated with temperature. An important
observation based on pooling of hot spring P-GDGT data
from Tengchong, Tibet, YNP and the US Great Basin is
the dominance of the effect of pH on the distribution of
P-GDGTs over any biogeographic patterns or tempera-
ture, as described below.
Temperature. Although it has been demonstrated that the
number of cyclopentyl rings increases with temperature in
some pure cultures (De Rosa et al., 1980; Uda et al.,
2001; Shimada et al., 2008; Boyd et al., 2011) and natural
environments (Schouten et al., 2002; 2003; Li et al., 2013;
Paraiso et al., 2013), temperature showed less impact
than pH on the distribution of P- or C-GDGTs in both
Tengchong hot springs and the expanded dataset. The
impact of temperature on the distribution of P- or
C-GDGTs could be detected in Tengchong hot springs
only in the Desulfurococcales- and Thermoproteales-
dominated group having a narrow pH range (9.0 ±0.9,
n=6). In the expanded dataset, temperature did correlate
with archaeal lipid content but was much less significant
than pH. Similarly, Li and colleagues (2013) only found
a positive correlation between temperature and P-RI
in a subgroup of samples having a narrow pH range
(7.7 ±0.5, n=12) in Tibetan hot springs. Paraiso and
colleagues (2013) found that temperature and pH both
correlated strongly with the relative abundance of lipids
sampled in US Great Basin hot springs that had a pH
range of 6.8–10.1. Lastly, Boyd and colleagues (2013) did
Fig. 7. RDA ordination diagrams of (A) major archaeal community structure and (B) archaeal lipids with environmental variables from hot
springs. Correlations between environmental variables and RDA axes are represented by the length and angle of dashed arrows
(environmental factor vectors). Solid arrows represent the relative abundance of (A) archaeal orders or (B) P-GDGTs. The numbers on the
samples correspond with the sample numbers in Table S2.
Archaeal lipids from terrestrial hot springs 9
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

not find any correlation between temperature and RI in
sediment samples from YNP hot springs having a pH
range of 2.06–9.57. These studies demonstrate that the
influence of temperature on GDGT content decreases
with expanding pH range in terrestrial geothermal
environments.
Dissolved oxygen. Dissolved oxygen significantly corre-
lated with P-RI and C-RI values, indicating that the
oxygen concentration, or redox status, may affect the
distribution of GDGTs in Tengchong hot springs. Paraiso
and colleagues (2013) reported that oxygen correlated
with the distribution of GDGTs in US Great Basin hot
springs, albeit with moderate significance, and suggested
that the relationship may result from a co-variation
between oxygen and temperature. Considering that
there was co-variation between oxygen and pH for the
Tengchong samples, oxygen or redox status may or may
not directly affect the distribution of P- and C-GDGTs in
Tengchong springs. The observed difference between
polar and core GDGT substructures specifically for
C-GDGT-0 and P-GDGT-0 was similar to observations
reported for YNP hot springs (Boyd et al., 2013). In addi-
tion, samples revealing high differences between P- and
C-GDGT-0 lipid species were retrieved from sites charac-
terized by relatively low oxygen concentrations (Figs 2
and S5). The observed differences between P- and
C-GDGT-0 lipid species might be due to different oxygen-
dependent degradation rates of this particular GDGT
structure and, in particular, of the polar moieties. Further
studies are needed to investigate the impact of dissolved
oxygen on the distribution of GDGTs under experimental
conditions.
Relationship with community structure. Archaeal 16S
rRNA gene and lipid profiles from Tengchong hot
springs collected in samples of January 2011 were inves-
tigated to address the effect of community structure
on the archaeal lipid profile. The hot spring samples
from Tengchong were classified into three major groups
based on 16S rRNA pyrosequencing analyses: a Desul-
furococcales- and Thermoproteales-abundant group, a
Sulfolobales-abundant group and a Thaumarchaeota-
Aigarchaeota-abundant group. It is known that
Desulfurococcales produce a high proportion of GDGT-0
(Lanzotti et al., 1989; Blöchl et al., 1997; Huber et al.,
1998; Jahn et al., 2004; Schouten et al., 2007) while
Thermoproteales produce a mixture of GDGT-0, -1, -2, -3
and -4 (Langworthy and Pond, 1986; Thurl and Schäfer,
1988; Trincone et al., 1992; Völkl et al., 1993; Itoh et al.,
1999; 2003; Schouten et al., 2007). Consequently,
RI values from the Desulfurococcales- and Thermo-
proteales-abundant group of springs were relatively low.
In contrast, GDGT-5, -6, -7 and -8 are frequently found in
Sulfolobales (De Rosa and Gambacorta, 1988;
Takayanagi et al., 1996; Sturt et al., 2004; Schouten et al.,
2007), resulting in high RI values in the Sulfolobales-
dominated group of springs. Thaumarchaeota synthe-
size crenarchaeol and GDGTs decorated with 0–3
cyclopentyl moieties (Schouten et al., 2008; Pitcher
et al., 2010; 2011), which result in low RI values
for the Thaumarchaeota-Aigarchaeota-abundant group of
springs. The significant correlation between the taxo-
nomic structure of the archaeal community and the GDGT
lipid composition suggests that the composition of
archaeal lipids in Tengchong hot springs reflects the
resident community structure. Although Briggs and
colleagues (2013) reported significant temporal change in
archaeal taxonomic compositions because of high rainfall
influx during the monsoon season, lipid samples collected
from most sites over multiple seasons clustered together,
indicating that the distribution of P- and C-GDGTs in those
springs was relatively stable and not affected by archaeal
community dynamics. Whereas only pH had a significant
influence on the structure of the archaeal community and,
consequently, the distribution of GDGTs in Tengchong hot
springs, pH and temperature appeared to control the
structure of the archaeal community equally when the
Tengchong data were extended to include data from Tibet,
YNP and the US Great Basin hot springs. Only five in 18
genera were correlated with P-RI, indicating a more
complex relationship between archaeal community struc-
ture and the distribution of P-GDGTs. Nevertheless, the
distribution of P-GDGTs still exhibited a dominant corre-
lation with pH, suggesting that the resident Archaea in
these hot springs are acclimated if not adapted to low pH
by their genetic capacity to effect the packing density of
their membranes by increasing cyclopentyl rings in
GDGTs at the rank of community.
Conclusions
Comparative analysis of the abundance of P- and
C-GDGTs in hot springs and their surrounding soils
revealed that archaeal lipids in both fractions were
predominantly produced in situ. Cluster analyses of P-
and C-GDGTs were similar to phylogenetic clusters, sug-
gesting that environmental factors control both the
phylogenetic and GDGT compositions that are intimately
linked in Tengchong hot springs. pH appears to be
the dominant factor affecting the distribution of P- or
C-GDGTs in terrestrial hot springs at the global scale
whereas temperature appears to affect lipid distribution
when the pH and/or phylogenetic composition are similar.
Furthermore, dissolved oxygen or redox status may affect
GDGT profiles because of differential degradation of indi-
vidual GDGT species under different redox conditions,
which is also secondary control on GDGT composition.
10 W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Experimental procedures
Field measurements and sample collection
Field measurements and sample collection in Tengchong
and the US Great Basin were conducted as follows. Water
temperature and pH were measured with portable meters
(LaMotte, MD, USA). Hot spring water was sampled
without disrupting the sediment as to not affect the bulk
water chemistry. Total ammonia, nitrate plus nitrite, sulfate,
total sulfide, ferrous iron, dissolved oxygen were measured
colourimetrically using Hach test kits (Hach Chemical, IA,
USA) in the field following procedures described previously
(Zhang et al., 2008; Dodsworth et al., 2012). Briefly, water
samples for measurements of total ammonia, sulfate, nitrate
plus nitrite and ferrous iron were filtered (0.20 μm
polyethersulfone membrane filters, Pall Corp., NY) and
allowed to cool to ambient temperature prior to analysis.
Ammonia was determined using salicylate oxidation
method. Nitrate plus nitrite was determined by cadmium
reduction of nitrate and subsequent diazotization of nitrite.
Sulfate was determined by barium reaction. Ferrous iron
was measured with 1,10-phenanthroline indicator. Dissolved
oxygen and sulfide were measured immediately after sam-
pling. Dissolved oxygen measurements were made using
the HRDO Accuvac ampule method (Hach, high range) or
the Indigo Carmine method (Hach, low range), avoiding
contacting with the atmosphere. Sulfide was measured with
the Pomeroy methylene blue method to prevent heat inac-
tivation of reagents and to allow rapid analysis prior to
oxidation.
A total of 46 samples from 15 geothermal springs were
collected with sterile spatulas and spoons; each sample was
homogenized in a pre-sterilized aluminum pan before alloca-
tion in different containers. Another 10 samples were col-
lected from surrounding soils in June 2011. Four sediment
samples from Great Boiling Spring in the US Great Basin
were collected in February 2010. All the samples for lipid and
DNA analyses were immediately frozen on dry ice and kept at
−80°C in the laboratory.
Lipid extraction and fractionation
All the samples were freeze-dried and ground to fine powder.
For samples from January 2011, about 5 g of the powder
were extracted at room temperature using a sonication
method (Schouten et al., 2007; Wei et al., 2011) with metha-
nol (MeOH) (twice), MeOH-dichloromethane (DCM) (1:1, v/v;
twice) and DCM (twice). The total lipid extract (TLE) was
fractionated over a pre-activated silica gel chromatography
into apolar and polar (C-GDGTs and P-GDGTs) fractions
using hexane-DCM (9:1, v/v) and DCM-MeOH (1:1, v/v) as
eluents respectively. The polar fraction was amended with a
known amount of a C46 internal GDGT standard, divided into
two equal aliquots and evaporated under N2. One aliquot was
re-dissolved in hexane-isopropanol (99:1, v/v) and filtered
with a 0.45 μm membrane to obtain the C-GDGTs and then
for high-performance liquid chromatography-atmospheric
pressure chemical ionization-mass spectrometry (HPLC-
APCI-MS) analysis. The other aliquot was acid hydrolysed by
refluxing it in 2 ml of 5% HCl in MeOH for 3 h at 70°C. The
solution was cooled to room temperature, adjusted to pH 5
with KOH-MeOH (1:1, v/v) and partitioned by addition of
bidistilled water and DCM to reach a ratio of 1:1 H2O-MeOH.
The mixture was washed two more times with DCM. All the
DCM phases were dried under N2and processed as
described above for the C-GDGTs. The abundance of
P-GDGTs was calculated as the difference between hydro-
lysed and non-hydrolysed fractions.
For samples from Tengchong in June 2011 and July 2012
and samples from the US Great Basin in February 2010, we
used a modified Bligh and Dyer method (Sturt et al., 2004)
using a mixture of MeOH, DCM and phosphate buffer at pH
7.4 (2:1:0.8 v/v) for four times. After sonication for 5–10 min,
additional DCM and buffer were added to the mixture to
achieve a final MeOH/DCM/buffer ratio of 1:1:0.8. The
phases were separated and the extraction repeated three
more times. The extracts were separated into an organic
phase and an aqueous phase by centrifugation at 2500 r.p.m.
for 10 min; the bottom DCM phase was collected using a
glass pipette, which was repeated three times. The TLEs
were dried under a stream of nitrogen gas at about 37°C. A
known amount of C46 compound was added as internal stand-
ard to the TLEs. The TLEs were equally divided into two
parts. One half, named F1, was dried under a stream of
nitrogen gas at about 37°C and dissolved in hexane/
isopropanol (99:1, v/v) and filtered using a 0.45 μm
polytetrafluoroethylene (PTFE) filter for further analysis. The
other half, named F2, was acid hydrolysed as described
above. The hydrolysed fractions were dissolved in hexane/
isopropanol (99:1, v/v) and filtered using a 0.45 μm PTFE
filter for further analysis. Core lipid in the sediment samples
were directly quantified from F1. Polar lipid was indirectly
quantified by the difference between F2and F1.
Analysis and quantification of GDGTs
P- and C-GDGTs samples were analysed using published
methods (Hopmans et al., 2000). Analyses were per-
formed using an Agilent 1200 HPLC device connected to an
Agilent 6460 triple-quadrupole mass spectrometer. Separa-
tion was achieved on an Alltech Prevail Cyano column
(150 ×2.1 mm, 3 μm) maintained at 30°C. Injection volume
was 10 μl. For the first 5 min, GDGTs were eluted
isocratically with 99% A (hexane) and 1% B (isopropanol),
followed by a linear gradient to 1.8% B in 45 min with a
constant flow rate of 0.2 ml min−1. After each analysis,
the column was cleaned by back-flushing hexane-propanol
(9:1, v/v) at 0.2 ml min−1for 10 min. Quantification of
archaeol (m z−1653), GDGTs (m z−11302, 1300, 1298,
1296, 1294, 1292, 1290 and 1288) and a C46 GDGT internal
standard (m z−1744) was achieved by integration of the
extracted peak areas. GDGTs were quantified by integration
of peak area from the extracted ion chromatogram. The
GDGT-4 peak area from extracted ion chromatogram was
corrected by subtracting 46% of crenarchaeol (Hopmans
et al., 2000).
Calculation of RI
RI was calculated according to and equation described by
Uda and colleagues (2004) as follows:
Archaeal lipids from terrestrial hot springs 11
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

RI
GDGT GDGT GDGT GDGT
GDGT GDGT G
=
+× +× +×
+× +× +×
-- - -
--
12 2 3 3 4 4
55667DDGT
GDGT GDGT GDGT GDGT
GDGT GDGT GDGT GDG
-
----
---
7
0123
456
++ +
++++
TT -7 (3)
The structures of GDGT-0 to -7 are referred to Fig. S1.
Pyrosequencing data
To examine the phylogenetic composition of microbial com-
munities associated with the lipid analyses, previously pub-
lished, pyrosequenced amplified 16S rRNA gene datasets
were downloaded from the Short Read Archive database at
NCBI: Tengchong (SRA056421; Hou et al., 2013), Tibet
(SRA061437; Wang et al., 2013), YNP (SRR648329; Boyd
et al., 2013) and US Great Basin (SRX201239, SRX201242,
SRX201245 and SRX201246; Cole et al., 2013). To ensure
that all analyses were comparable, we re-analysed all the
datasets using a single pipeline. Pyrosequencing data were
converted to sequence reads using MOTHUR software
(Schloss et al., 2009) and were analysed using the QIIME
standard pipeline (Caporaso et al., 2010). Sequence reads
were first filtered by removing low-quality or ambiguous
reads. Sequences that passed the initial quality control were
checked with CHIMERASLAYER (Haas et al., 2011). The chi-
meric sequences were excluded for further analysis. The
remaining high quality 16S rRNA sequences were clustered
into OTUs using UCLUST (Edgar, 2010) with 97% sequence
identity threshold. Taxonomy was assigned using the Ribo-
somal Database Project classifier 2.2 (minimum confidence
of 80%) (Cole et al., 2009). Then all the archaeal taxonomies
at the rank of order were chosen to recalculate the proportion.
Statistical methods
Clustering of samples by lipid profiles and archaeal commu-
nity composition was performed using the base program
in R 2.12.1 [freeware available at http://cran.r-project.org/
(Maindonald, 2007)]. The relative abundances of P-GDGTs
and C-GDGTs and identified archaeal taxons from hot spring
sediments were imported into R 2.12.1. A hierarchical clus-
tering tree was generated by Heatmap command in R
package (Maindonald, 2007). Relationship between distribu-
tion of P- and C-GDGTs and environmental factors were
explored with RDA using the software CANOCO [version 4.5;
Microcomputer Power (Ter Braak and Smilauer, 2002)]. Vari-
able significance was examined by Monte-Carlo significance
test with permutation number of 999. Spearman’s rho values
were calculated in SPSS Statistics 16.1 (SPSS, Chicago, IL,
USA) to identify correlative relationships among community
structure, distribution of GDGTs and environmental variables.
Equations to predict relationship between environmental
parameters and P- or C-RI based on the distribution of
GDGTs was derived using multiple regression analysis
(Peterse et al., 2012) by MATLAB software (freeware avail-
able at http://www.mathworks.com/, Version 7.11.0, The
Mathworks, Natwick, MA, USA). The units of the regression
equations are as Table S1 shown; one exception is the dis-
solved oxygen, which is transferred to milligram per litre).
Samples having all the eight measured variables were
included into the multiple regress analyses (four exceptions
were the Gxs in June 2011 and July 2012 having high abun-
dance of crenarcheaol, Drc and Zzq in July 2012 having
outliers DO values).
Acknowledgements
We thank the Tengchong Partnership for International
Research and Education team and the staff from the Yunnan
Tengchong Volcano and Spa Tourist Attraction Development
Corporation for their assistance. We also wish to thank two
anonymous reviewers and the handling editor for their con-
structive and helpful comments. This study was funded by the
National Natural Science Foundation of China Grants
#41373072 and #40972211, the US National Science Foun-
dation Grants #ETBC-1024614 and #OISE-0968421, the
‘1000 Talents Program’ and the State Key Laboratory of
Marine Geology at Tongji University. M. G. K. was supported
by Visiting Scholar Funds of Tongji University and Research
Incentive Funds from the University of North Carolina at
Charlotte. B. P. H. was supported by a generous donation
from Greg Fullmer through the UNLV Foundation.
References
Bauersachs, T., Speelman, E.N., Hopmans, E.C., Reichart,
G.-J., Schouten, S., and Damsté, J.S.S. (2010) Fossilized
glycolipids reveal past oceanic N2fixation by heterocystous
cyanobacteria. Proc Natl Acad Sci USA 107: 19190–
19194.
Biddle, J.F., Lipp, J.S., Lever, M.A., Lloyd, K.G., Sorensen,
K.B., Anderson, R., et al. (2006) Heterotrophic Archaea
dominate sedimentary subsurface ecosystems off Peru.
Proc Natl Acad Sci USA 103: 3846–3851.
Blöchl, E., Rachel, R., Burggraf, S., Hafenbradl, D.,
Jannasch, H.W., and Stetter, K.O. (1997) Pyrolobus
fumarii, gen. and sp. nov., represents a novel group of
archaea, extending the upper temperature limit for life to
113°C. Extremophiles 1: 14–21.
Boyd, E.S., Pearson, A., Pi, Y., Li, W.J., Zhang, Y.G., He, L.,
et al. (2011) Temperature and pH controls on glycerol
dibiphytanyl glycerol tetraether lipid composition in
the hyperthermophilic crenarchaeon Acidilobus sulfurire-
ducens.Extremophiles 15: 59–65.
Boyd, E.S., Hamilton, T.L., Wang, J.X., He, L., and Zhang,
C.L. (2013) The role of tetraether lipids in the adaptation of
thermophilic archaea to acidity. Front Microbiol 4: 62.
doi:10.3389/fmicb.2013.00062.
Briggs, B.R., Brodie, E.L., Tom, L.M., Dong, H.L., Jiang, H.C.,
Huang, Q.Y., et al. (2013) Seasonal patterns in microbial
communities inhabiting the hot springs of Tengchong,
Yunnan Province, China. Environ Microbiol 16: 1579–
1591.
Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K.,
Bushman, F.D., Costello, E.K., et al. (2010) QIIME allows
analysis of high-throughput community sequencing data.
Nat Methods 7: 335–336.
Cole, J.K., Peacock, J.P., Dodsworth, J.A., Williams, A.J.,
Thompson, D.B., Dong, H.L., et al. (2013) Sediment micro-
bial communities in Great Boiling Spring are controlled by
temperature and distinct from water communities. ISME J
7: 718–729.
12 W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Cole, J.R., Wang, Q., Cardenas, E., Fish, J., Chai, B., Farris,
R.J., et al. (2009) The Ribosomal Database Project:
improved alignments and new tools for rRNA analysis.
Nucleic Acids Res 37: D141–D145.
De Rosa, M., and Gambacorta, A. (1988) The lipids of
archaebacteria. Prog Lipid Res 27: 153–175.
De Rosa, M., Esposito, E., Gambacorta, A., Nicolaus, B., and
Bu’Lock, J.D. (1980) Effects of temperature on ether lipid
composition of Caldariella acidophila.Phytochemistry 19:
827–831.
Dodsworth, J.A., McDonald, A.I., and Hedlund, B.P. (2012)
Calculation of total free energy yield as an alternative
approach for predicting the importance of potential
chemolithotrophic reactions in geothermal springs. FEMS
Microbiol Ecol 81: 446–454.
Edgar, R.C. (2010) Search and clustering orders of magni-
tude faster than BLAST. Bioinformatics 26: 2460–2461.
Haas, B.J., Gevers, D., Earl, A.M., Feldgarden, M., Ward,
D.V., Giannoukos, G., et al. (2011) Chimeric 16S rRNA
sequence formation and detection in Sanger and 454-
pyrosequenced PCR amplicons. Genome Res 21: 494–
504.
Harvey, H.R., Fallon, R.D., and Patton, J.S. (1986) The effect
of organic matter and oxygen on the degradation of bac-
terial membrane lipids in marine sediments. Geochim
Cosmochim Acta 50: 795–804.
He, L., Zhang, C.L., Dong, H.L., Fang, B., and Wang, G.H.
(2012) Distribution of glycerol dialkyl glycerol tetraethers in
Tibetan hot springs. Geosci Front 3: 289–300.
Hedlund, B.P., Cole, J.K., Williams, A.J., Hou, W., Zhou, E.,
Li, W., et al. (2012) A review of the microbiology of the
Rehai geothermal field in Tengchong, Yunnan Province,
China. Geosci Front 3: 273–288.
Hopmans, E.C., Schouten, S., Pancost, R.D., van der Meer,
M.T., and Sinninghe Damsté, J.S. (2000) Analysis of intact
tetraether lipids in archaeal cell material and sediments by
high performance liquid chromatography/atmospheric
pressure chemical ionization mass spectrometry. Rapid
Commun Mass Spectrom 14: 585–589.
Hou, W.G., Wang, S., Dong, H.L., Jiang, H.C., Briggs, B.R.,
Peacock, J.P., et al. (2013) A comprehensive census of
microbial diversity in hot springs of Tengchong, Yunnan
Province China using 16S rRNA gene pyrosequencing.
PLoS ONE 8: e53350.
Huang, Q.Y., Jiang, H.C., Zhang, C.L., Li, W.J., Deng, S.C.,
Yu, B.S., et al. (2010) Abundance of ammonia-oxidizing
microorganisms in response to environmental variables of
hot springs in Yunnan Province, China. Wei Sheng Wu Xue
Bao 50: 132–136.
Huber, R., Dyba, D., Huber, H., Burggraf, S., and Rachel, R.
(1998) Sulfur-inhibited Thermosphaera aggregans sp.
nov., a new genus of hyperthermophilic archaea isolated
after its prediction from environmentally derived 16S rRNA
sequences. Int J Syst Bacteriol 48: 31–38.
Itoh, T., Suzuki, K., Sanchez, P.C., and Nakase, T. (1999)
Caldivirga maquilingensis gen. nov., sp. nov., a new genus
of rod-shaped crenarchaeote isolated from a hot spring in
the Philippines. Int J Syst Bacteriol 49: 1157–1163.
Itoh, T., Suzuki, K., Sanchez, P., and Nakase, T. (2003)
Caldisphaera lagunensis gen. nov., sp. nov., a novel
thermoacidophilic crenarchaeote isolated from a hot spring
at Mt Maquiling, Philippines. Int J Syst Evol Microbiol 53:
1149–1154.
Jahn, U., Summons, R., Sturt, H., Grosjean, E., and Huber,
H. (2004) Composition of the lipids of Nanoarchaeum
equitans and their origin from its host Ignicoccus sp. strain
KIN4/I. Arch Microbiol 182: 404–413.
Jiang, H.C., Huang, Q.Y., Dong, H.L., Wang, P., Wang, F.P.,
Li, W.J., et al. (2010) RNA-based investigation of
ammonia-oxidizing archaea in hot springs of Yunnan Prov-
ince, China. Appl Environ Microbiol 76: 4538–4541.
Kan, J., Clingenpeel, S., Macur, R.E., Inskeep, W.P., Lovalvo,
D., Varley, J., et al. (2011) Archaea in Yellowstone Lake.
ISME J 5: 1784–1795.
Kuypers, M.M., Blokker, P., Erbacher, J., Kinkel, H., Pancost,
R.D., Schouten, S., et al. (2001) Massive expansion of
marine archaea during a mid-Cretaceous oceanic anoxic
event. Science 293: 92–95.
Langworthy, T.A., and Pond, J.L. (1986) Archaebacterial
ether lipids and chemotaxonomy. Syst Appl Microbiol 7:
253–257.
Lanzotti, V., Trincone, A., Nicolaus, B., Zillig, W., Rosa, M.D.,
and Gambacorta, A. (1989) Complex lipids of Pyrococcus
and AN1, thermophilic members of archaebcteria belong-
ing to Thermococcales.Biochim Biophys Acta 1004:
44–48.
Li, F.Y., Zhang, C.L., Dong, H.L., Li, W.J., and Williams, A.
(2013) Environmental controls on the distribution of
archaeal lipids in Tibetan hot springs: insight into the appli-
cation of organic proxies for biogeochemical processes.
Environ Microbiol Rep 5: 868–882.
Liao, Z., and Guo, G. (1986) Geology of the Tengchong
geothermal field and surrounding area, west Yunnan,
China. Geothermics 15: 339–345.
Lipp, J.S., Morono, Y., Inagaki, F., and Hinrichs, K.U. (2008)
Significant contribution of Archaea to extant biomass in
marine subsurface sediments. Nature 454: 991–994.
Logemann, J., Graue, J., Köster, J., Engelen, B., Rullkötter,
J., and Cypionka, H. (2011) A laboratory experiment of
intact polar lipid degradation in sandy sediments.
Biogeosciences 8: 2547–2560.
Macalady, J.L., Vestling, M.M., Baumler, D., Boekelheide, N.,
Kaspar, C.W., and Banfield, J.F. (2004) Tetraether-linked
membrane monolayers in Ferroplasma spp: a key to sur-
vival in acid. Extremophiles 8: 411–419.
Maindonald, J.H. (2007) Data Analysis and Graphics Using
R: An Example-Based Approach. New York, USA: Cam-
bridge University Press.
Pagaling, E., Grant, W.D., Cowan, D.A., Jones, B.E., Ma,
Y., Ventosa, A., et al. (2012) Bacterial and archaeal
diversity in two hot spring microbial mats from the geo-
thermal region of Tengchong, China. Extremophiles 16:
607–618.
Paraiso, J.J., Williams, A.J., Huang, Q., Wei, Y., Dijkstra, P.,
Hungate, B.A., et al. (2013) The distribution and abun-
dance of archaeal tetraether lipids in US Great Basin
hot springs. Front Microbiol 4: 247. doi:10.3389/
fmicb.2013.00247.
Pearson, A., Huang, Z.Y., Ingalls, A.E., Romanek, C.S.,
Wiegel, J., Freeman, K.H., et al. (2004) Nonmarine
crenarchaeol in Nevada hot springs. Appl Environ
Microbiol 70: 5229–5237.
Archaeal lipids from terrestrial hot springs 13
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

Pearson, A., Pi, Y.D., Zhao, W.D., Li, W.J., Li, Y.L., Inskeep,
W., et al. (2008) Factors controlling the distribution of
archaeal tetraethers in terrestrial hot springs. Appl Environ
Microbiol 74: 3523–3532.
Peterse, F., van der Meer, J., Schouten, S., Weijers, J.W.H.,
Fierer, N., Jackson, R.B., et al. (2012) Revised calibration
of the MBT–CBT paleotemperature proxy based on
branched tetraether membrane lipids in surface soils.
Geochim Cosmochim Acta 96: 215–229.
Petsch, S.T., Eglington, T.I., and Edwards, K.J. (2001) 14C-
dead living biomass: evidence for microbial assimilation of
ancient organic carbon during shale weathering. Science
292: 1127–1131.
Pitcher, A., Rychlik, N., Hopmans, E.C., Spieck, E., Rijpstra,
W.I., Ossebaar, J., et al. (2010) Crenarchaeol dominates
the membrane lipids of Candidatus Nitrososphaera
gargensis, a thermophilic group I.1b Archaeon. ISME J 4:
542–552.
Pitcher, A., Wuchter, C., Siedenberg, K., Schouten, S., and
Sinninghe Damsté, J.S. (2011) Crenarchaeol tracks winter
blooms of ammonia-oxidizing Thaumarchaeota in the
coastal North Sea. Limnol Oceanogr 56: 2308–2318.
Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R.,
Hartmann, M., Hollister, E.B., et al. (2009) Introducing
mothur: open-source, platform-independent, community-
supported software for describing and comparing microbial
communities. Appl Environ Microbiol 75: 7537–7541.
Schouten, S., Hopmans, E.C., Schefuß, E., and
Sinninghe Damsté, J.S. (2002) Distributional variations in
marine crenarchaeotal membrane lipids: a new tool for
reconstructing ancient sea water temperatures? Earth
Planet Sci Lett 204: 265–274.
Schouten, S., Wakeham, S.G., Hopmans, E.C., and
Sinninghe Damsté, J.S. (2003) Biogeochemical evidence
that thermophilic archaea mediate the anaerobic oxidation
of methane. Appl Environ Microbiol 69: 1680–1686.
Schouten, S., van der Meer, M.T., Hopmans, E.C., Rijpstra,
W.I., Reysenbach, A.L., Ward, D.M., et al. (2007) Archaeal
and bacterial glycerol dialkyl glycerol tetraether lipids in
hot springs of Yellowstone National Park. Appl Environ
Microbiol 73: 6181–6191.
Schouten, S., Hopmans, E.C., Baas, M., Boumann, H.,
Standfest, S., Konneke, M., et al. (2008) Intact membrane
lipids of ‘Candidatus Nitrosopumilus maritimus,’ a culti-
vated representative of the cosmopolitan mesophilic
group I Crenarchaeota. Appl Environ Microbiol 74: 2433–
2440.
Shimada, H., Nemoto, N., Shida, Y., Oshima, T., and
Yamagishi, A. (2008) Effects of pH and temperature on the
composition of polar lipids in Thermoplasma acidophilum
HO-62. J Bacteriol 190: 5404–5411.
Song, Z.Q., Chen, J.Q., Jiang, H.C., Zhou, E.M., Tang, S.K.,
Zhi, X.Y., et al. (2010) Diversity of Crenarchaeota in terres-
trial hot springs in Tengchong, China. Extremophiles 14:
287–296.
Song, Z.Q., Wang, F.P., Zhi, X.Y., Chen, J.Q., Zhou, E.M.,
Liang, F., et al. (2013) Bacterial and archaeal diversities in
Yunnan and Tibetan hot springs, China. Environ Microbiol
15: 1160–1175.
Sorensen, K.B., Canfield, D.E., Teske, A.P., and Oren, A.
(2005) Community composition of a hypersaline
endoevaporitic microbial mat. Appl Environ Microbiol 71:
7352–7365.
Sturt, H.F., Summons, R.E., Smith, K., Elvert, M., and
Hinrichs, K.U. (2004) Intact polar membrane lipids
in prokaryotes and sediments deciphered by high-
performance liquid chromatography/electrospray ioniza-
tion multistage mass spectrometry-new biomarkers for
biogeochemistry and microbial ecology. Rapid Commun
Mass Spectrom 18: 617–628.
Takai, K., and Sako, Y. (1999) A molecular view of archaeal
diversity in marine and terrestrial hot water environments.
FEMS Microbiol Ecol 28: 177–188.
Takayanagi, S., Kawasaki, H., Sugimori, K., Yamada, T.,
Sugai, A., Ito, T., et al. (1996) Sulfolobus hakonensis sp.
nov., a novel species of acidothermophilic archaeon. Int J
Syst Bacteriol 46: 377–382.
Ter Braak, C., and Smilauer, P. (2002) Canoco for Windows
Version 4.5. Wageningen, the Netherlands: Biometris-Plant
Research International.
Thurl, S., and Schäfer, W. (1988) Lipids from the
sulphur-dependent archaebacterium Thermoproteus
tenax.Biochim Biophys Acta 961: 253–261.
Tong, W., and Zhang, M. (1989) Tengchong Geothermics.
Beijing, China: Science Press.
Trincone, A., Nicolaus, B., Palmieri, G., De Rosa, M.,
Huber, R., Huber, G., et al. (1992) Distribution of
complex and core lipids within new hyperthermophilic
members of the archaea domain. Syst Appl Microbiol 15:
11–17.
Uda, I., Sugai, A., Itoh, Y.H., and Itoh, T. (2001) Variation in
molecular species of polar lipids from Thermoplasma
acidophilum depends on growth temperature. Lipids 36:
103–105.
Uda, I., Sugai, A., Itoh, Y.H., and Itoh, T. (2004) Variation
in molecular species of core lipids from the order
Thermoplasmales strains depends on the growth tempera-
ture. J Oleo Sci 53: 399–404.
Valentine, D.L. (2007) Adaptations to energy stress dictate
the ecology and evolution of the Archaea. Nat Rev
Microbiol 5: 316–323.
van de Vossenberg, J.L., Driessen, A.J., and Konings, W.N.
(1998) The essence of being extremophilic: the role of the
unique archaeal membrane lipids. Extremophiles 2: 163–
170.
Völkl, P., Huber, R., Drobner, E., Rachel, R., Burggraf, S.,
Trincone, A., et al. (1993) Pyrobaculum aerophilum sp.
nov., a novel nitrate-reducing hyperthermophilic archaeum.
Appl Environ Microbiol 59: 2918–2926.
Wang, F., Peng, Z.C., Chen, W.J., Wang, Z.R., Yang, J.W.,
Zhang, Z.L., et al. (2000) High-precision thermal ionization
mass spectrometry dating of young volcanic rocks by using
U-series method. Chin Sci Bull 45: 83–87.
Wang, S., Hou, W.G., Dong, H.L., Jiang, H.C., Huang, L.Q.,
Wu, G., et al. (2013) Control of temperature on microbial
community structure in hot springs of the Tibetan Plateau.
PLoS ONE 8: e62901.
Wei, Y.L., Wang, J.X., Liu, J., Dong, L., Li, L., Wang, H., et al.
(2011) Spatial variations in archaeal lipids of surface water
and core-top sediments in the South China Sea: implica-
tions for paleoclimate studies. Appl Environ Microbiol 77:
7479–7489.
14 W. Xie et al.
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology

White, D., Davis, W., Nickels, J., King, J., and Bobbie, R.
(1979) Determination of the sedimentary microbial
biomass by extractible lipid phosphate. Oecologia 40:
51–62.
Woese, C.R., Kandler, O., and Wheelis, M.L. (1990) Towards
a natural system of organisms: proposal for the domains
Archaea, Bacteria, and Eucarya. Proc Natl Acad Sci USA
87: 4576–4579.
Wu, W.Y., Zhang, C.L., Wang, H.Y., He, L., Li, W.J., and
Dong, H.L. (2013) Impacts of temperature and pH on the
distribution of archaeal lipids in Yunnan hot springs,
China. Front Microbiol 4: 312. doi:10.3389/fmicb.2013
.00312.
Zhang, C.L., Pearson, A., Li, Y.L., Mills, G., and Wiegel, J.
(2006) Thermophilic temperature optimum for crenar-
chaeol synthesis and its implication for archaeal evolution.
Appl Environ Microbiol 72: 4419–4422.
Zhang, C.L., Ye, Q., Huang, Z.Y., Li, W.J., Chen, J.Q., Song,
Z.Q., et al. (2008) Global occurrence of archaeal amoA
genes in terrestrial hot springs. Appl Environ Microbiol 74:
6417–6426.
Supporting information
Additional Supporting Information may be found in the online
version of this article at the publisher’s web-site:
Fig. S1. Structures of GDGTs presented in the terrestrial
geothermal springs and the surrounding soil. GTGT-0 is
glycerol trialkyl glycerol tetraether without any cyclopentyl
ring.
Fig. S2. Correlations of abundances of P-GDGTs and
C-GDGTs (in log values) from Tengchong hot spring sedi-
ments and surrounding soils in June 2011.
Fig. S3. Comparison of concentrations of archaeol and
isoprenoid lipids from hot springs and surrounding soils in
June 2011.
Fig. S4. Cluster analysis of C-GDGTs from Tengchong
hot spring sediments in January 2011, June 2011 and
July 2012. Sample names are shown on the right of the
figure; GDGTs are colour coded and shown at the bottom
of the figure. Samples are clustered into three groups:
group 1 (pH =3.0 ±0.7, T =70.7 ±15.2°C; n=14), group 2
(pH =7.5 ±0.8, T =68.1 ±15.2°C; n=8) and group 3
(pH =8.5 ±1.0, T =79.0 ±11.5°C; n=24). Samples in corre-
sponding groups are boxed with dash lines.
Fig. S5. Difference (in %) between P-GDGTs and corre-
sponding C-GDGTs from Tengchong hot spring samples
having measurable dissolved oxygen (DO) concentrations.
Sample names are shown on the right, followed by corre-
sponding dissolved oxygen (μgl
−1).
Fig. S6. RDA ordination plots for the first two principal
dimensions of the relationship between the distribution
of C-GDGTs and relevant environmental parameters from
Tengchong in January 2011, June 2011 and July 2012. Cor-
relations between environmental variables and RDA axes are
represented by the length and angle of dash arrows (envi-
ronmental factor vectors). Solid arrows stand for the propor-
tion of P-GDGTs.
Fig. S7. Cluster analysis based on archaeal phylogenetic
composition of Tengchong hot springs in January 2011 at
the rank of order. MCG: miscellaneous crenarchaeotal
group. Sample names are shown on the right of the figure.
Samples are clustered into three groups that correspond to
pH: group 1 (pH =3.1 ±1.0, T =73.9 ±16.8°C; n=4), group
2 (pH =8.8 ±0.8, T =85.6 ±5.8°C; n=8) and group 3
(pH =7.0 ±0.4, T =77.1 ±5.7°C; n=2). Groups are deline-
ated by dashed lines.
Fig. S8. Regression plots for temperature and P- or C-RI
in group 2 from Tengchong hot springs in January 2011
[see Fig. S8; Desulfurococcales- and Thermoproteales-
dominated group; P- or C-RI of SrbzU and SrbzD did not fit
the equation, which might be because Srbz pool was sourced
by surface run-off (Hou et al., 2013), which may deliver more
soil contamination than other hot springs].
Table S1. Lipid and chemistry data for isoprenoidal GDGTs
from Tengchong hot springs in January 2011, June 2011, July
2012 and US Great Basin hot spring in October 2010.
Table S2. Composition and relative abundance (%) of
Archaea and distribution of P-GDGTs from Tengchong hot
springs in January 2011, YNP hot springs in June through
August of 2008, Tibet hot springs in June 2010 and US Great
Basin hot springs in October 2010.
Table S3. The non-parametric correlation results of archaeal
composition, P-GDGTs (relative abundance) and environ-
mental variables from Tengchong hot springs in January
2011, YNP hot springs in June through August of 2008, Tibet
hot springs in June 2010 and US Great Basin hot springs in
October 2010.
Archaeal lipids from terrestrial hot springs 15
© 2014 Society for Applied Microbiology and John Wiley & Sons Ltd, Environmental Microbiology