Stable isotope probing analysis of the influence of liming on root exudate utilization by soil microorganisms.

Abstract

Rhizosphere microorganisms play an important role in soil carbon flow, through turnover of root exudates, but there is little information on which organisms are actively involved or on the influence of environmental conditions on active communities. In this study, a 13CO2 pulse labelling field experiment was performed in an upland grassland soil, followed by RNA-stable isotope probing (SIP) analysis, to determine the effect of liming on the structure of the rhizosphere microbial community metabolizing root exudates. The lower limit of detection for SIP was determined in soil samples inoculated with a range of concentrations of 13C-labelled Pseudomonas fluorescens and was found to lie between 10(5) and 10(6) cells per gram of soil. The technique was capable of detecting microbial communities actively assimilating root exudates derived from recent photo-assimilate in the field. Denaturing gradient gel electrophoresis (DGGE) profiles of bacteria, archaea and fungi derived from fractions obtained from caesium trifluoroacetate (CsTFA) density gradient ultracentrifugation indicated that active communities in limed soils were more complex than those in unlimed soils and were more active in utilization of recently exuded 13C compounds. In limed soils, the majority of the community detected by standard RNA-DGGE analysis appeared to be utilizing root exudates. In unlimed soils, DGGE profiles from 12C and 13C RNA fractions differed, suggesting that a proportion of the active community was utilizing other sources of organic carbon. These differences may reflect differences in the amount of root exudation under the different conditions.

Full text

Page 1

Environmental Microbiology (2005) 7(6), 828–838 doi:10.1111/j.1462-2920.2005.00756.x
© 2005 Society for Applied Microbiology and Blackwell Publishing Ltd
Blackwell Science, LtdOxford, UKEMIEnvironmental Microbiology 1462-2912Blackwell Publishing Ltd, 20047 6828838Original ArticleStable isotope probing of soil microbial communitiesJ. I. Rangel-Castro
et al.
Received 8 September, 2004; accepted 2 November, 2004. *For
correspondence. E-mail j.prosser@abdn.ac.uk; Tel. (+44) 0 1224
555848; Fax (+44) 0 1224 555844.
Stable isotope probing analysis of the influence of
liming on root exudate utilization by soil
microorganisms
J. Ignacio Rangel-Castro,1,2 Ken Killham,2 Nick Ostle,3
Graeme W. Nicol,2 Ian C. Anderson,4
Charlie M. Scrimgeour,5 Phil Ineson,6 Andy Meharg2
and Jim I. Prosser1*
1School of Medical Sciences, University of Aberdeen,
Institute of Medical Sciences, Foresterhill, Aberdeen
AB25 2ZD, Scotland, UK.
2School of Biological Sciences, University of Aberdeen,
AB24 3UU, Scotland, UK.
3Centre for Ecology and Hydrology, Lancaster
Environment Centre, Bailrig, Lancaster LA1 4AP, UK.
4The Macaulay Institute, Craigiebuckler, Aberdeen
AB15 8QH, Scotland, UK.
5Plant–Soil Interface, Scottish Crop Research Institute,
Invergowrie, Dundee DD2 5DA, Scotland, UK.
6Department of Biology, University of York, York
YO10 5YW, UK.
Summary
Rhizosphere microorganisms play an important role
in soil carbon flow, through turnover of root exudates,
but there is little information on which organisms are
actively involved or on the influence of environmental
conditions on active communities. In this study, a
13CO2 pulse labelling field experiment was performed
in an upland grassland soil, followed by RNA-stable
isotope probing (SIP) analysis, to determine the effect
of liming on the structure of the rhizosphere microbial
community metabolizing root exudates. The lower
limit of detection for SIP was determined in soil sam-
ples inoculated with a range of concentrations of 13C-
labelled Pseudomonas fluorescens and was found to
lie between 105 and 106 cells per gram of soil. The
technique was capable of detecting microbial commu-
nities actively assimilating root exudates derived from
recent photo-assimilate in the field. Denaturing gradi-
ent gel electrophoresis (DGGE) profiles of bacteria,
archaea and fungi derived from fractions obtained
from caesium trifluoroacetate (CsTFA) density gradi-
ent ultracentrifugation indicated that active communi-
ties in limed soils were more complex than those in
unlimed soils and were more active in utilization of
recently exuded 13C compounds. In limed soils, the
majority of the community detected by standard RNA-
DGGE analysis appeared to be utilizing root exudates.
In unlimed soils, DGGE profiles from 12C and 13C RNA
fractions differed, suggesting that a proportion of the
active community was utilizing other sources of
organic carbon. These differences may reflect differ-
ences in the amount of root exudation under the dif-
ferent conditions.
Introduction
The rhizosphere plays a fundamental role in the formation,
maintenance and turnover of soil organic matter (SOM),
influencing flow, storage and the sink–source balance of
C. Despite this, understanding of the role of rhizosphere
microbial communities in C flow is limited and little is
known of roles of different members of the community in
assimilating root exudate. C in SOM is derived mainly
from rhizosphere C flow and decomposed shoots, roots
and litter (Kuzyakov and Domanski, 2002). Most of
the compounds mediating rhizosphere C flow are low-
molecular weight and polymeric compounds, including
carbohydrates, amino acids and organic acids, which are
either easily assimilated by soil microorganisms or require
production of extracellular enzymes before uptake (Kill-
ham and Yeomans, 2001). This C pool is fundamental as
a driving force in soil microbial processes (Lugtenberg and
Dekkers, 1999) and, consequently, in the interactions
within rhizosphere communities and their physiological
characteristics. The role of microbial communities is also
an important factor when considering applied aspects,
such as control of pathogens, bioremediation processes
and enhancement of plant nutrient supply.
One direct approach to determination of C flow in plant–
soil systems is the use of 13C and 14C isotopes pulsed into
the systems as labelled CO2 (Meharg, 1994; Kuzyakov,
2002). This approach has been restricted to laboratory
systems (Kuzyakov and Domanski, 2000; Kuzyakov et al.,
2001), with the exception of a limited number of pulse

Page 2

Stable isotope probing of soil microbial communities829
© 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 7, 828–838
labelling field experiments (Johnson et al., 2002; Staddon
et al., 2003; Olsrud and Christensen, 2003; Rangel-
Castro et al., 2004, 2005). Rangel-Castro and colleagues
(2005) demonstrated that the flux of recently photosynthe-
sized 13C to soil microbial biomass occurs within hours
(i.e. less than 24 h) of incorporation of 13CO2 into shoot
biomass. Calculation of exponential decay kinetics for 13C
incorporated indicated that the half-life of the isotope in
the microbial biomass was 4.7 days. These data, coupled
with results of Ostle and colleagues (2003), who found
maximum incorporation of 13C into microbial RNA within
4–8 days of pulse labelling, demonstrate rapid incorpora-
tion and turnover of 13C by the soil microbial biomass.
Tracking 13C in cellular components (e.g. lipids and
nucleic acids) (stable isotope probing) has recently been
used to determine which functional groups actively assim-
ilate 13C-labelled substrates (Boschker et al., 1998; Rada-
jewski et al., 2000; Manefield et al., 2002a; Butler et al.,
2003; Lueders et al., 2004a,b; Treonis et al., 2004). Stable
isotope probing of DNA and RNA has been used in micro-
cosm systems to detect active groups in assimilation of a
range of organic compounds (Padmanabhan et al., 2003),
phenol degradation (Manefield et al., 2002a), and meth-
ane and methanol assimilation (Radajewski et al., 2000;
Lueders et al., 2004b). However, a recent attempt to iden-
tify soil microbes active in assimilation of rhizosphere
C flow, in pots planted with grassland turfs, was un-
successful (Griffiths et al., 2004). 13C dilution in nucleic
acid samples was suggested as a factor preventing
adequate separation of labelled and unlabelled RNA after
centrifugation.
Land management strategies aimed at improving plant
production, including application of lime, are known to
affect microbial activities such as C and N transformations
(Hopkins, 1997; Neale et al., 1997). In a previous study,
liming was found to affect C flow in plant shoots and roots
of upland grassland plots (Rangel-Castro et al., 2004)
and turnover of C in microbial biomass of limed soil plots
was faster than in unlimed plots (Rangel-Castro et al.,
2005). This could result from differences in active rhizo-
sphere microbial communities in limed and unlimed soils.
The main objective of this study was therefore to use
RNA-SIP to determine if differences in C turnover in limed
and unlimed soils were due to differences in the active
microbial community structure. In addition, lower limits for
detection of 13C-labelled cells in soil, using SIP, were
determined.
Results
SIP optimization
Preliminary experiments were carried out using
labelled Escherichia coli and Pseudomonas fluorescens
13C-
cells to determine the location of 13C and 12C RNA follow-
ing centrifugation and to determine lower detection limits
of 13C-labelled cells. Results were similar for both organ-
isms and only P. fluorescens analyses are presented.
Figure 1 illustrates DGGE profiles from fractions derived
from soil containing 105 and 106 P. fluorescens cells per
gram of soil. Banding patterns obtained using the latter
inoculum indicated that maximum 13C levels were found
in fractions 4–6. This was consistent with observations of
Manefield and colleagues (2002b) in their successful
application of this approach to separate 13C and 12C
RNA from a 13C-labelled Pseudomonas putida phenol-
degrading strain. Therefore, fraction 7 was not considered
for further analyses and fractions 8–10 were analysed as
representative of 12C RNA fractions.
The 12C community showed the complex DGGE band-
ing patterns typical of soil bacterial communities, with one
band corresponding also to P. fluorescens, although this
was not expected because of its likely low abundance in
comparison with a total cell concentration of approxi-
mately 109 cells per gram of soil. Denaturing gradient gel
electrophoresis profiles derived from 13C RNA (fractions
4–6) were dominated by the P. fluorescens band, although
sequences representative of the total soil community were
detected as faint bands in some fractions. This indicates
that some 12C sequences could have migrated to the 13C
zone during centrifugation and were amplified during the
polymerase chain reaction (PCR). Denaturing gradient gel
electrophoresis profiles derived from 12C extracted from
soil containing 105 cells P. fluorescens per gram showed
similar complexity and a very faint band, corresponding
to P. fluorescens, was observed. Products representing
P. fluorescens were not detected in the 13C fractions, indi-
cating that the detection limit lies within the range of 105-
106 cells per gram for this organism. Again, faint bands
representative of sequences from the soil bacterial com-
munity were observed in fractions 5 and 6, but none was
detected in fraction 4.
Microbial incorporation of root exudates
To confirm that density gradient centrifugation led to frac-
tionation of 12C and 13C RNA, and that 13C root exudates
had been assimilated by soil microorganisms, fractions
along the gradient were analysed for d13C. The data, illus-
trated in Fig. 2, show d13C levels significantly greater than
natural abundance values (see Table 1) in fractions 4–6,
as expected, but label was also detected in other fractions,
indicating that separation was not completely efficient.
However, the proportion of 13C RNA in fractions 8–10 will
be negligible, in comparison with 12C, and will not have
compromised analysis of these fractions. High levels of
13C were also detected in fraction 3 but amplification of
sequences was less reliable and consistent, and occa-

Page 3

830 J. I. Rangel-Castro et al.
© 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 7, 828–838
sionally represented only high GC content sequences.
Nevertheless, future studies should examine this fraction,
which may contain sequences that are not present in
fractions 4–6 but which potentially are highly 13C-labelled.
Table 1 shows d13C values of RNA extracted from soils
collected 3 h and 5 days after completion of pulse label-
ling. 13C enrichment was highest in samples collected
after 3 h of labelling and were greater than natural abun-
dance values. Denaturing gradient gel electrophoresis
profiles generated by reverse transcription polymerase
chain reaction (RT-PCR) amplification, using bacterial
primers, of fractions sampled along the 12C–13C gradient
are illustrated in Fig. 3. For limed soils, no difference could
be detected by visual examination of profiles in different
fractions. Denaturing gradient gel electrophoresis profiles
reflect relative abundance of different sequence types but
provide no information on absolute numbers or biomass,
or abundance of sequence types. Similar amounts of RT-
PCR product were added to each lane, leading to similar-
ities in total intensity within lanes for each fraction, rather
than differences in levels in soil samples of 12C and 13C
RNA. Similarities in DGGE profiles in each lane therefore
indicate similarities in relative abundance of different
sequence types in 12C- and 13C RNA. Denaturing gradient
gel electrophoresis profiles from 12C RNA fractions are
equivalent to those from standard RNA-DGGE analyses,
which are considered to represent active members of the
community. Similarity between these profiles and those
from 13C RNA DGGE therefore indicates that these organ-
isms were actively incorporating root exudates. In con-
trast, DGGE profiles derived from 12C and 13C fractions
differed for unlimed soils. For example, bands marked I,
II and III increased in relative intensity in 13C fractions,
while those marked A, B and C decreased in relative
intensity. Other bands showed no detectable difference in
M 456 8 9 10 M 4 5 6 8 9 10 M
105Cells g-1
106Cells g-1
P. fluorescens
band
Fig. 1. Denaturing gradient gel electrophoresis
banding pattern of bacterial SSU RNA partial
sequences of fractions 4–6 (13C) and 8–10 (12C)
obtained by ultracentrifugation of CsTFA den-
sity gradients of RNA extractions from unlimed
grassland soils inoculated with Pseudomonas
fluorescens at final cell concentrations of 105
and 106 cells per gram of soil. M represents
markers consisting of ammonia oxidizer clones
amplified with the same bacterial primers.
Fig. 2. d13C values of fractions obtained after separation of labelled
and non-labelled RNA from limed (?) and unlimed (?) soil samples
by CsTFA density gradient separation.
-19.0
-18.5
-18.0
-17.5
-17.0
-16.5
-16.0
2345678910 11 12
Fraction
13C
Table 1. d13C values of RNA extractions in samples collected 3 h and
5 days after completion of pulse labelling in the field.
Time after completion of pulse labelling Sample
d13C
3 hUnlimed
Limed
Unlimed
Limed
Unlimed
Limed
-14.34
-13.85
-20.00
-24.06
-23.90
-22.80
5 days
Natural abundance

Page 4

Stable isotope probing of soil microbial communities 831
© 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 7, 828–838
relative intensity along the gradient. These data indicate
that a proportion of the active community, as determined
by 12C RNA DGGE profiles, was active through utilization
of root exudates but that other members were active
through utilization of other forms of soil carbon. Those
sequence types derived from organisms utilizing root exu-
dates will have reduced relative band intensity in the 12C
fractions, while those utilizing other forms of soil carbon
will show greater or similar relative band intensity.
Archaeal DGGE profiles (Fig. 4) showed a similar pat-
tern in limed soils, with no detectable change in relative
intensity of bands in DGGE profiles in fractions taken
along the 12C–13C gradient. Again, this indicates that the
activity of archaea detected by targeting RNA was due to
utilization and incorporation of root exudates. In unlimed
soils, however, no band was detected in 13C fractions,
indicating that any archaea in these soils utilizing root
exudates were below the limit of detection of the tech-
nique and/or that activity was due to utilization of other
forms of organic carbon. Archaeal communities have
been characterized previously at Sourhope and are dom-
inated by group 1.1b and 1.1c crenarchaea (Nicol et al.,
2003). The large difference in the percentage of GC con-
tent of 16S rRNA genes of these two lineages results in
group 1.1c sequences typically migrating further than
group 1.1b sequences in DGGE analysis. Comparison
with the marker lane containing sequences representative
of different crenarchaeal lineages including group 1.1b
and 1.1c indicates that both these lineages may be
present and utilize root exudates (Fig. 4).
Denaturing gradient gel electrophoresis profiles gener-
ated using fungal 18S rRNA gene primers for amplification
of RNA from limed and unlimed soils showed similar
behaviour to bacterial profiles (Fig. 5). In limed soils,
DGGE profiles generated from 12C and 13C fractions were
similar, while differences were observed for unlimed soils.
This provides further evidence for use of alternative
organic carbon substrates in unlimed soils, while the
majority of the active fungal community in limed soils were
utilizing root exudates. In addition, visual observation of
DGGE profiles indicates that the communities in limed
soils were more complex than in unlimed soils.
Sequencing of clones unique to 13C fractions in unlimed
soils
Sequences were determined for three clones of bacterial
origin and two of fungal origin containing 16S rRNA and
18S rRNA gene inserts similar to those which increased
in relative abundance in
13C fractions derived from
Fig. 3. Denaturing gradient gel electrophoresis banding pattern of
bacterial SSU RNA partial sequences of fractions 4–6 (13C) and 8–
10 (12C) obtained by ultracentrifugation of CsTFA density gradients of
RNA extractions from limed and unlimed grassland soils labelled in
the field with 13CO2. M represents markers consisting of ammonia
oxidizer clones amplified with the same bacterial primers.
M 4 5 6 8 9 10 M
Limed soil
Unlimed soil
M 4 5 6 8 9 10 M
II
I
III
B
A
C
Fig. 4. Denaturing gradient gel electrophoresis banding pattern of
archaeal SSU RNA partial sequences of fractions 4–6 (13C) and 8–
10 (12C) obtained by ultracentrifugation of CsTFA density gradients of
RNA extractions from limed and un-limed grassland soils labelled in
the field with 13CO2. M represents a marker consisting of cloned 16S
rRNA gene sequences representative of different group 1 Crenarcha-
eota lineages including 1.1b and 1.1c.
M 4 5 6 8 9 10 M
Limed soil
Unlimed soil
M 4 5 6 8 9 10 M

Page 5

832 J. I. Rangel-Castro et al.
© 2005 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 7, 828–838
unlimed soil samples. Sequences of archaeal bands were
not determined, as no archaeal band was preferentially
enriched in 13C RNA. To confirm reliability of sequences,
three different colonies with the same insert were
sequenced. Final analyses of the sequences were based
on sequences of approximately 600 bp. Sequences were
submitted to BLAST search and closest matches with
sequences from GenBank are listed in Table 2. Bacterial
sequences were most closely related to Sphingomonas
and Mycobacterium isolates and an isolate from rhizo-
sphere soil from a grazed pasture (Joseph et al., 2003).
The two fungal sequences were most closely related to
Sistotrema eximum and Rhodotorula glutinis.
Discussion
Previous studies employing SIP have employed laboratory
systems in which incorporation is optimized by application
of relatively high levels of label, repeated applications and
minimization of losses of labelled substrates. Field appli-
cation of label reduces the sensitivity of the technique and
a seeding experiment, with 13C-labelled P. fluorescens,
was therefore carried out to determine the minimum level
of detection of the method, which was found to lie between
105 and 106 cells per gram of soil. Sensitivity will be
affected by a number of factors, including the efficiency of
cell lysis, RNA extraction and PCR amplification, all of
which may vary with target organism, soil type, primers
and protocols used. In addition, inoculated cells were
labelled by growth on fully labelled glucose, maximizing
labelling of RNA, and the detection limit will be higher for
partially labelled cells. This experiment also allowed
assessment of the efficiency of separation of 12C- and 13C-
labelled RNA and of potential generation of random PCR
products at low levels of target DNA, as observed in other
studies. Results indicate successful separation of 12C and
13C RNA but with low level cross-contamination of 12C and
13C RNA. For example, a number of bands derived from
the indigenous (unlabelled) soil community were detected
in fractions 5 and 6, suggesting some migration of 12C
RNA to the 13C fractions during centrifugation, while the
P. fluorescens band was detectable in 12C fractions. How-
ever, there was no significant carry-over of 12C to fraction
4. Lueders and colleagues (2004a) found that caesium
trifluoroacetate (CsTFA) separation of labelled RNA was
most efficient in gradients containing RNA from a single
bacterial species, with reduced efficiency in gradients
derived from two or more species. Lueders and col-
leagues (2004a) and Manefield and colleagues (2002a)
considered reduced efficiency to be related to the pres-
ence of rRNA from species with similar buoyant densities,
and is the most likely explanation for cross-contamination
in the complex rhizosphere community investigated here.
Table 2. Closest database matches of sequences corresponding to bacterial and fungal bands with increased relative abundance in 13C RNA
fractions from unlimed soils.
Band number and origin Accession numberClosest match % Similarity
I Bacteria
II Bacteria
III Bacteria
I Fungi
II Fungi
AY705870
AY705871
AY705872
AY705873
AY705874
Sphingomonas sp. CSSB-2 AB167382
Bacterium Ellin5080a AY234497
Mycobacterium sp. M175 AY004157
Sistotrema eximum AF334935
Rhodotorula glutinis AB018403
99
97
99
98
99
a. Strain isolated from a rotationally grazed pasture (Joseph et al., 2003).
Fig. 5. Denaturing gradient gel electrophoresis banding pattern of
fungal SSU RNA partial sequences of fractions 4–6 (13C) and 8–10
(12C) obtained by ultracentrifugation of CsTFA density gradients of
RNA extractions from limed and unlimed grassland soils labelled in
the field with 13CO2. M represents a marker consisting of fungal 18S
rRNA gene clones.
M 4 5 6 8 9 10 M
Limed soil
Unlimed soil
M 4 5 6 8 9 10 M
II
I