25th International Meeting on Organic Geochemistry IMOG 2011

P-177
Rhizoliths in loess: evidence for the heterotrophic lifestyle of
branched GDGT-producing bacteria
Arnaud Huguet 1 , Guido L.B. Wiesenberg 2 , Martina Gocke 2 , Céline Fosse 3 , Sylvie
Derenne 1
1 BioEMCo, CNRS/UPMC UMR 7618, Paris, France, 2 Department of Agroecosystem Research, BayCEER,
University of Bayreuth, Bayreuth, Germany, 3 Chimie ParisTech (ENSCP), Laboratoire de Spectrométrie de
Masse, Paris, France (corresponding author:arnaud.huguet@upmc.fr)
Over the last years, an increasing number of
studies have focused on glycerol dialkyl glycerol
tetraethers (GDGTs), which are complex lipids of high
molecular weight (>1000 Da) present in membranes
of archaea and some bacteria. Isoprenoid GDGTs
with acyclic or ring containing dibiphytanyl chains are
known to be synthesized by archaea. In soil, another
type of GDGTs, which can be distinguished from
tetraethers of archaeal origin by way of the branched
nature of the alkyl chain, was discovered recently.
Branched GDGTs were subsequently detected in a
large variety of environments (peats and soils, hot
springs, lacustrine and marine sediments) and are
synthesized by still unknown bacteria. The relative
distribution of branched GDGTs can be related mainly
to mean annual air temperature (MAAT) and soil pH.
The cyclisation ratio of branched tetraethers (CBT),
quantifying the relative abundance of cyclopentyl
rings of branched GDGTs, correlates rather well with
soil pH, whereas the methylation index of branched
tetraethers (MBT), expressing the degree of
methylation of branched GDGTs, depends on MAAT
and, to a lesser extent, on soil pH. The MBT and CBT
can be used as proxies for paleotemperatures and
paleo soil pH. The aim of the present work was to
examine the distribution and abundance of GDGTs in
rhizoliths (calcified roots) and surrounding loess
collected from a loess-palaeosol sequence in
Nussloch (SW Germany). For two rhizoliths from a
depth between 2.2 m and 2.6 m below present
surface, loess transects were sampled from the
former root towards root-free loess at distances of 0-
2.5 cm and 2.5-5 cm (rhizoloess samples). Two
reference loess samples without visible root remains
were taken at a distance of 50-70 cm from the
rhizoliths.
Branched GDGTs were much more abundant in the
rhizoliths than in the rhizoloess and almost absent in
reference loess samples. The very high abundance of
branched GDGTs in the rhizoliths suggests that
branched GDGT source organisms feed on root
remains. In larger distances to root surfaces, nutrient
and energy supply decreases, likely leading to the
lower abundance of branched GDGT source
microorganisms observed in the rhizoloess and
reference loess compared to the rhizoliths. Branched
GDGT distribution patterns were different for
individual sample types. The MBT was lower in the
rhizoliths (0.23-0.26) than in the rhizoloess (0.37-
0.47) and loess samples (0.39-0.41). Branched
GDGTs in the rhizoliths and loess might have been
produced during different time intervals, since
radiocarbon dating of one rhizolith revealed its
Holocene age (3150 years b.p.), in contrast to late
Pleistocene age of surrounding loess (17-20 ka). In
rhizoliths, branched GDGTs were very likely
generated during and/or after plant death (i.e. about
3000 years ago), whereas in loess they were probably
formed after sedimentation and/or during earlier
pedogenesis of the modern soil. Consequently, the
age of branched GDGTs in the rhizoliths and
reference loess may differ, which could explain the
differences in GDGT composition between the former
roots and the loess.
MAAT and pH were reconstructed using the MBT
and CBT indices. Consistent results were obtained for
each sample type (rhizolith, rhizoloess, loess). Thus,
MAAT estimates derived from the two rhizoliths (2.9–
3.7 °C) were lower than those derived from rhizoloess
(9.1–14.6 °C) and loess (10.1–10.8 °C) samples. At
the time of the formation of rhizoliths, temperature
might have been like present MAAT (10 °C), implying
that MBT/CBT-derived MAAT estimates for the
rhizoliths (ca. 3 °C) are likely lower than the ―real‖
MAAT value. In contrast, MAAT estimates based on
branched GDGT distribution in the loess seem too
high (ca. 10 °C). Indeed, loess was deposited during
cold periods, with assumed temperatures close to
zero. The reasons for the deviation between real and
reconstructed MAAT values remain unclear,
suggesting that caution should be exercised when
applying and interpreting temperature estimates
derived from bacterial GDGTs in loess-paleosol
sequences, considered as important terrestrial
archives for studying Quaternary climate. In contrast
to temperature, CBT-derived pH values were in the
same range (7.7–8.1) for the three sample types and
are consistent with the actual pH of reference loess
(8.1). To the best of our knowledge, this is the first
time that the heterotrophic lifestyle of branched
GDGT-producing bacteria is evidenced solely based
on GDGT abundance.
317