Characterization of Bacterial Community Structure in Rhizosphere

Soil of Grain Legumes

S. Sharma1, M.K. Aneja1, J. Mayer2, J.C. Munch1 and M. Schloter1
(1) Institute of Soil Ecology, GSF National Research Center for Environment and Health, PO Box 1129, D-85764 Neuherberg, Germany
(2) Department of Organic Farming and Cropping Systems, University of Kassel, Nordbahnhofstr. 1a, D-37213 Witzenhausen, Germany

Received: 27 February 2004 / Accepted: 14 April 2004 / Online publication: 7 July 2005

Introduction
Abstract
The cultivation of legumes in crop rotation practice is an
Molecular techniques were used to characterize bacterial
important way to improve soil quality in agricultural soil.
community structure, diversity (16S rDNA), and activity
Legumes possess all the characteristic properties of green
(16S rRNA) in rhizospheres of three grain legumes: faba
manures like improving soil by adding large amounts of
beans (Vicia faba L., cv. Scirocco), peas (Pisum sativum
organic material and valuable nutrients to it, preventing
L., cv. Duel) and white lupin (Lupinus albus L., cv.
soil erosion, preventing the leaching away of nutrients,
Amiga). All plants were grown in the same soil under
and improving soils ability to hold water. Most impor-
controlled conditions in a greenhouse and sampled after
tant, however, is their contribution to the overall N
fruiting. Amplified 16S rDNA and rRNA products (using
economy of soil by sequestering atmospheric N through
universal bacterial primers) were resolved by denaturing
symbiotic N2 fixation and through subsoil N retrieval [6].
gradient gel electrophoresis (DGGE). Distinct profiles
Soil immediately surrounding the plant roots and
were observed for the three legumes with most of the
directly influenced by its exudates, called the rhizosphere,
bands derived from RNA being a subset of those derived
is of high importance in maximizing the benefits ob-
from DNA. Comparing the total bacterial profiles with
tained by legumes. This particular zone in agricultural
actinomycete-specific ones (using actinomycete-specific
soil is characterized by a distinct microbial community
primers) highlighted the dominance of this group in the
structure, which differs from bulk soil in its composition
three rhizospheres. 16S PCR and RT-PCR products were
and activity. The rhizosphere has been regarded, there-
cloned to construct libraries and 100 clones from each
fore, as a hot spot for microbial colonization and activity.
library were sequenced. Actinomycetes and proteobacte-
The microbes of rhizosphere and their functions are
ria dominated the clone libraries with differences in the
highly influenced by the plant root [10], and in turn the
groups of proteobacteria. Absence of b-subdivision
bacterial and fungal members in the rhizosphere have a
members in pea and c-subdivision members of proteo-
high impact on plant growth [20]. As variation in
bacteria in faba bean rhizosphere was observed. Plant-
microbial community structure may have effects on
dependent rhizosphere effects were evident from signifi-
ecosystem processes (e.g., nutrient recycling, decompo-
cant differences in the bacterial community structure of
sition of organics) or plantmicrobe interaction (e.g.,
the legume rhizospheres under study. The study gives a
growth of pathogens, release of plant-growth promoting
detailed picture of both residing and active bacterial
rhizobacteria or genetically engineered microorganisms),
community in the three rhizospheres. The high abun-
understanding how community processes affect ecosys-
dance of actinomycetes in the rhizospheres of mature
tem processes is of central interest in ecology [20]. The
legumes indicates their possible role in soil enrichment
amount and composition of organic materials released by
after the legumes are plowed into the soil as biofertilizers.
the plants are the important factors that determine the
structure and function of microbial community in the
rhizospheres [16].
Numerous studies reveal differences in rhizodeposi-
tion by different grain legumes [12, 18], thereby most
likely to influence their respective rhizosphere structure
Correspondence to: S. Sharma; E-mail: shilpi.sharma@gsf.de and function. For example, in a study by Mayer et al. [18]

DOI: 10.1007/s00248-004-0041-7 d Volume 49, 407415 (2005) d
Springer Science+Business Media, Inc. 2005 407
408
considerable differences, both in quality and quantity,
between rhizodeposits of three grain legumes were re-
ported. Studies addressing this effect of legume rhizo-
deposition on bacterial communities are to date limited.
Structural and functional characterization of rhizo-
sphere of legumes has been focused primarily on sym-
bionts, Rhizobium and Bradyrhizobium spp. [15, 28].
However, a deeper understanding of the bacterial com-
munity and their rhizosphere dynamics has been re-
stricted because of the limitations of traditional pure
culture techniques. Therefore, to get a better insight of
the bacterial community structure, methods independent
of cultivation have to be used. rRNA has been proposed
as an appropriate target for assessing changes in active
bacterial populations [30]. We chose to target 16S rRNA
and rDNA extracted directly from rhizosphere soil in
reverse transcriptase PCR (RT-PCR) and PCR analysis,
respectively. DGGE was used for initial screening for
differences or similarities between the samples. In addi-
tion, the PCR and RT-PCR products were cloned to
construct libraries.
In the present study, we report bacterial community
structure in the rhizospheres of three economically
important grain legumes: Vicia faba, Lupinus albus, and
Pisum sativum. These are extensively used as green
manure in Central Europe in organic farming. The main
objectives of the present study were (i) to characterize the
bacterial community in the rhizosphere of the three grain
legumes grown in the same soil, (ii) to compare the
bacterial profiles with group-specific patterns, (iii) to
identify rhizosphere effects with respect to bacterial
diversity, and (iv) to identify the predominantly active
groups as a result of rhizosphere effect.
Methods
Experimental Design and Sampling. Soil samples from
a site in northwest Germany were taken from the top 0
20 cm of a Eutric Cambisol. The field had been cultivated
using organic farming management for 10 years. The soil
was characterized as a sandy loam with 17.3% clay, 30.1%
silt, 52.6% sand, pH (0.01 M CaCl2) 6.0, 1.58% total C,
0.15% total N, 140 mg P kg)1 (CaL), 208 mg K kg)1
(CaL), and 100 mg Mg kg)1 (CaL). The water holding
capacity (10 mm sieved soil) was 309 g H2O kg)1 dry soil.
Fresh soil samples were sieved through a 10 mm mesh,
prior to use.
Ten rectangular boxes per legume (60 45 40 cm)
were prepared with about 70 kg of soil per box. Seeds of
faba beans (Vicia faba L., cv. Scirocco), peas (Pisum
sativum L., cv. Duel), and white lupin (Lupinus albus L.,
cv. Amiga) were inoculated with legume-specific Rhizo-
bium inoculants (Radicin, Radicin Institut, Iserlohn,
Germany) according to manufacturers instructions, and
three seeds were sown in each box. The plants were
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
grown at 18C with 70% humidity and a photoperiod
consisting of 16 h of light and 8 h of darkness. The plants
were watered equally on alternate days. The health of the
plants was monitored during this period by the visual
observation of the color of their leaves, and they appeared
healthy. Sampling was performed after fruiting (5 months
after date of sowing). Plants and soil were removed from
the pots. Subsequently, excess bulk soil was removed
from the roots by shaking, leaving firmly adhering soil,
which was defined as rhizosphere soil. Thirty samples (10
pots per legume, with three plants in each pot) were
taken for each legume, and these were stored at )20C.
Nucleic Acid Extraction. Nucleic acid extraction
from the rhizosphere soil material was performed using
the method of coextraction of DNA and RNA described
by Griffiths et al. [11]. This involved bead beating and
solvent extraction of the nucleic acids. To obtain pure
DNA, incubation at 37C with RNase A (Sigma) at a final
concentration of 100 lg mL)1 for 10 min was performed.
Prior to reverse transcription, DNA was removed from
RNA by treatment with DNase (1 U lL)1; RNase-free;
Promega) according to manufacturers instructions.
PCR and RT-PCR Amplification. PCR and RT-
PCR targeting 16S rDNA and rRNA, respectively, was
performed with F984-GC and R1378 universal bacterial
primers [13], which are specific for highly conserved 16S
rRNA regions of prokaryotes corresponding to E. coli
positions 968 to 984 and 1378 to 1401, respectively.
Amplification of actinomycetes was performed using a
nested PCR approach with actinomycete-specific prim-
ers, F243 and R1378, in the first PCR and universal
bacterial primers, F984-GC and R1378, in the second
PCR to compare the actinomycete population with total
bacterial population [13]. Reverse transcription of 16S
rRNA and subsequent amplification was performed by
using a two-step reaction scheme, as described below.
Two microliters of total-RNA sample was added to an
18-lL RT mixture prepared according to manufacturers
instructions (Omniscript RT Kit, Qiagen, Hilden, Ger-
many). Primer R1378 was used in the reaction. Synthe-
sized cDNA and directly extracted DNA were further
used for PCR according to the following protocol. One
lL of cDNA/DNA was added to a 48.5-lL PCR mixture
that consisted of 29 lL of nuclease-free water, 5 lL of 10
reaction buffer, 5 lL of 2 mM deoxyribonucleoside tri-
phosphate mixture, 1 lL each of 10 lM F984-GC and
R1378 primers, 5 lL of 3% bovine serum albumin, and
2.5 lL of dimethylsulfoxide. The reaction involved hot
start at 95C for 10 min followed by addition of 0.5 lL of
Cloned Pfu DNA Polymerase (2.5 U lL)1, Stratagene,
Amsterdam, The Netherlands). The cycling parameters
were 94C for 1 min, 54C for 1 min, and 72C for 1 min
for 35 cycles, followed by a final extension at 72C for 10
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
min. PCR products were purified using QIAquick PCR
Purification Kit (Qiagen, Hilden, Germany). DNase-
treated nucleic acids, without being subjected to reverse
transcription, were used as controls in PCR to check for
residual DNA in RNA preparations.
Denaturing Gradient Gel Electrophoresis
(DGGE). DGGE was performed using 6% polyacryl-
amide gels (ratio of acrylamide to bisacrylamide, 37:1)
with a 5460% denaturant for total bacterial profiles and
5464% denaturant for profiles of total bacteria com-
pared with profiles specific for actinomycetes (100%
denaturant is defined as 7 M urea plus 40% formamide
[2]). Appropriate volumes containing about 2 lg of the
purified PCR and RT-PCR products, measured by
absorbance at 260 nm, were loaded. The gels were elec-
trophoresed at 60C at 50 V for 17 h using the D-Gene
system (Bio-Rad Laboratories, Munich, Germany). The
gels were silver stained using the protocol described by
Heukeshoven and Dernick [14]. Dried gels were scanned
using an HP Scanjet 7400c. The DGGE profiles obtained
were analyzed by clustering via the unweighted pair
group method with mathematical averages (UPGMA;
Dice coefficient of similarity) using GelCompar II Soft-
ware (Applied Maths, Kortrijk, Belgium). The position
tolerance was set at 1% and background subtraction was
applied. Both strong and weak bands were included in
the analysis, thus taking into account the presence and
absence of bands at specific positions. Cophenetic cor-
relations were calculated using the same software.
Cloning, Sequencing, and Sequence Analy-
sis. Purified amplification product (3 lL) was cloned
into the pCR-Blunt II-TOPO vector of the Zero Blunt
TOPO PCR Cloning Kit (Invitrogen, Karlsruhe, Ger-
many) according to manufacturers instructions. Liga-
tions were transformed into chemically competent One
Shot DH5a-T1R cells provided in the kit. Colonies were
inoculated in LB medium (supplemented with 50 lg
mL)1 kanamycin) and plasmid isolated using Qiagen
Plasmid Mini Kit (Qiagen, Hilden, Germany). Purified
plasmids were tested for inserts by EcoRI (MBI Fermen-
tas, Heidelberg, Germany) digestion using 1 U in 50 lL
reaction.
Inserts were sequenced on an ABI PRISM 310 Ge-
netic Analyser (Applied Biosystems, Foster City, CA,
USA) using CEQ 2000 Dye Terminator Cycle Sequencing
with Quick Start Kit (Beckman Coulter, CA). Sequences
were compared with NCBI BLAST.
Nucleotide Sequence Accession Numbers. The
clone sequences determined in this study have been
submitted to GenBank under accession numbers
AY143694 to AY143793 and AY144121 to AY144124.
409
Results
PCR, RT-PCR, and DGGE. The three rhizosphere soil
samples yielded positive PCR and RT-PCR results using
F984-GC and R1378 primers. DNase-treated nucleic
acids, not subjected to RT, yielded no PCR products,
indicating no residual DNA in the RNA preparation
(data not shown). The same rhizosphere soil sample was
used for both rDNA and rRNA analysis to minimize the
possibility of experimental variability. The products were
subsequently resolved by DGGE. DGGE profiles of 30
plants for each rhizosphere soil type (10 pots per legume
and three plants per pot) were compared to identify
sampling variations for both PCR and RT-PCR products.
More than 95% similarity was observed between the
profiles of the plants in the same pot and also for the 10
different pots for each legume (data not shown). Each of
the three legume rhizospheres compared produced a
distinct molecular profile, which was largely, but not
completely, different from the profile generated by the
other two plant rhizosphere samples (Fig. 1). In Lupinus
36 bands were obtained but the number decreased to 33
in Pisum and 28 in Vicia DNA-derived profiles. High
density of bands was observed in the lower half of the gel
indicating the dominance and high activity of bacteria
with high G+C content. In DGGE profiles for RT-PCR
products, the number of bands decreased to about 20.
In order to compare DGGE patterns, Dice coefficient
was determined and UPGMA was used to create a den-
drogram describing pattern similarities. This analysis
clearly distinguished between the DNA- and RNA-de-
rived DGGE patterns with two distinct clusters, one each
for DNA and RNA samples. The values of cophenetic
correlation, which is a parameter to express the consis-
tence of a cluster, was 93% between the DNA and RNA
clusters, confirming that the data were statistically sig-
nificant. With both 16S rDNA and 16S rRNA, bacterial
communities of Lupinus and Vicia seemed to be more
similar to each other than to Pisum.
To investigate high G+C bacteria in more detail,
actinomycete-specific primers were used in a nested PCR
approach. All three rhizosphere soil samples yielded po-
sitive PCR and RT-PCR results. On comparing the
DGGE profiles of total bacteria with the corresponding
actinomycetes, it was observed that for all legumes, >90%
of bands present in DGGE pattern using only universal
bacterial primers were also present in the profiles for
actinomycetes, thus indicating a dominance of actino-
mycetes in the bacterial populations (Fig. 2).
Cloning. Clone libraries were made for each rhi-
zosphere sample (samples used were same for which the
profiles have been shown in Fig. 1) with their respective
PCR and RT-PCR products to gain insight into the res-
ident and active bacterial population. To assess the
410
Figure 2. DGGE profiles of 16S rDNA and 16S rRNA fingerprints
of total bacteria compared with actinomycetes. Lanes 3, 6, 11: 16S
rDNA fragments of total bacteria from Lupinus, Vicia, and Pisum
rhizospheres; lanes 4, 5, 12: 16S rDNA fragments of actinomycetes
from Lupinus, Vicia, and Pisum rhizospheres; lanes 1, 7, 9: 16S
rRNA fragments of total bacteria from Lupinus, Vicia, and Pisum
rhizospheres; lanes 2, 8, 10: 16S rRNA fragments of actinomycetes
from Lupinus, Vicia, and Pisum rhizospheres, respectively.
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
Figure 1. DGGE profiles and UPGMA
tree of 16S rDNA and 16S rRNA
fingerprints. Lanes 1, 2, 3: 16S rDNA
fragments of Lupinus, Vicia, and Pisum
rhizospheres; lanes 4, 5, 6: 16S rRNA
fragments of Lupinus, Vicia, and Pisum
rhizospheres, respectively. UPGMA tree
representing the similarity of the
bacterial community profile obtained
by PCR-DGGE and RT-PCR-DGGE
from the three rhizospheres. Scale
represents percent similarity. Values of
cophenetic correlations are mentioned
at the branches. Arrows indicate the
position for the corresponding clone,
when the clone is run separately on
DGGE.
number of clones sufficient to encompass the bacterial
diversity (defined as a measure of evenness or dominance
of different bacterial groups), increasing numbers of
clones were randomly picked and sequenced from the
16S rDNA clone library for Lupinus. The sequences were
grouped according to NCBI classification into taxonomic
groups (as in Fig. 4). A collectors curve or group
abundance curve (the number of different groups de-
tected plotted versus the number of clones analyzed) was
constructed [23]. A plateau, as expected for full coverage
of library, was obtained after screening 90 clones (Fig. 3).
No new group was observed even when the number of
clones analyzed was increased. Collectors curves were
constructed for the other five clone libraries as well and a
similar plateau could be obtained after analyses of similar
number of clones (data not shown). Therefore, 100
clones were considered to be a representative of the
bacterial community in the rhizosphere soils under
investigation.
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
Figure 3. Collectors curve showing the number of different
groups plotted as a function of number of clones for Lupinus rhi-
zosphere 16S rDNA library. Groups are defined in Fig. 4, following
NCBI nomenclature.
Identification and Distribution of Clones. A se-
quence of 433 bp was used to assign the clones to bac-
terial groups using the NCBI BLAST program. Clones
were not assigned to specific species because of the rel-
atively small fragment under investigation. A total of 600
clones were analyzed (100 each from Lupinus, Pisum and
Vicia 16S rDNA and rRNA libraries). Figure 4 shows the
broad phylogenetic distribution of clones from each li-
brary. Many clones belonged to previously characterized
major groups including actinomycetes and proteobacte-
411
ria (a-subdivision). Recently recognized groups such as
Acidobacter and Verrucomicrobia were also represented.
Phylogenetic Analysis. Firmicutes constituted the
most abundant group in both 16S rDNA and rRNA li-
braries for all the three rhizosphere soil samples with
>30% of the clones in Vicia and Lupinus libraries and
21% of the clones in Pisum DNA library. On the other
hand, firmicutes in 16S rRNA library of Pisum consti-
tuted 42% of the clones; half of these were similar to
Actinobacteridae members (Fig. 5). Other groups such as
Sphaerobacteridae, Rubrobacteridae, the BacillusClos-
tridium group, and uncultured firmicutes constituted 10
55% of clones in the libraries. Pisum 16S rDNA and
rRNA libraries contained clones resembling members of
all the above listed groups.
Proteobacteria constituted the second major group
in all six libraries, with more than 20% representation in
Lupinus and Pisum and 15% in Vicia libraries. Clones
similar to the c-subdivision predominated with >40% of
the proteobacterial clones in 16S rDNA and rRNA li-
braries of Lupinus and Pisum (Fig. 6). In contrast, in
Vicia libraries d-subdivision constituted the major pro-
portion (>50%) of proteobacterial clones while c-sub-
division members were absent. Clones related to the b-
subdivision could only be observed in Lupinus libraries
with one exception in the Vicia DNA library. The a-
subdivision was the second most abundant group with
>30% of proteobacterial clones in all samples. In addi-
tion, uncultured proteobacterial clones not yet assigned
to any of the above-mentioned groups were also present.
Figure 4. Relative distribution of
clones to different phylogenetic groups.
Lu, P, and V stand for Lupinus, Pisum,
and Vicia DNA libraries; Lur, Pr, and
Vr stand for Lupinus, Pisum, and Vicia
cDNA libraries. Distribution of clones
to different groups is in accordance
with NCBI classification.
412
Proteobacterial clones in 16S rRNA libraries were ob-
served to be a subset of clones of respective DNA li-
braries.
Clones belonging to Verrucomicrobia and Fibrobact-
er/Acidobacter groups present in 16S rDNA libraries were
not found in the corresponding 16S rRNA libraries, with
Vicia being the only exception. Few clones related to
Green Nonsulfur bacteria and Nitrospira groups were also
observed among Vicia rhizosphere clones (rDNA).
Highly variable proportions (1842%) of clones from all
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
Figure 5. Relative distribution of
clones to different phylogenetic groups
of firmicutes. For nomenclature, see
legend to Fig. 4.
the libraries were most closely related to uncultured/
unidentified bacteria as grouped in NCBI.
Correlation between DGGE and Cloning. Selected
clones that were characteristic in Lupinus clone libraries
(Lur3, Lur36, and Lur44 belonging to firmicutes and
clones Lur43, Lur49, and Lu53 belonging to proteobac-
teria) were PCR amplified as described before and re-
solved by DGGE. This was done to find any existing
correlation between dominant bands in DGGE and the
Figure 6. Relative distribution of
clones to different phylogenetic groups
of proteobacteria. For nomenclature,
see legend to Fig. 4.
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
repetitive sequences in the clone libraries. All the selected
six clones had corresponding bands in the DGGE profiles
(Fig. 1).
Discussion
DNA-based detection assays do not discriminate between
dormant and active groups. DNA obtained from envi-
ronmental samples could originate from dormant cells,
dead cells, or even from free DNA. Adsorption of DNA at
mineral surfaces, especially in soils, could harbor more or
less intact nucleic acids a long time after lysis of the
source organism [17]. To obtain information about the
presence of metabolically active bacteria in the environ-
ment, 16S rRNA was targeted in RT-PCR and the
resulting DGGE profiles compared with those obtained
by DNA target analysis. Inherent limitations of PCR
amplification of 16S rDNA/rRNA from complex micro-
biota, such as differential PCR amplification, formation
of chimeric molecules, or formation of deletion mutants
[29], cannot be completely ruled out. Such errors were
minimized by using replicate soil samples, multiple nu-
cleic acid extractions, and subsequent PCR.
To rule out the limitations of DGGE [20, 27] and
also to identify the organisms, cloning of rDNA and
rRNA PCR products was performed and clones analyzed.
Taxonomic groups, considered as operational taxonomic
units (OTU), when plotted to form a collectors curve,
formed a plateau-shaped graph after analysis of 90 clones.
In several studies analyzing bulk soil microbial commu-
nities, no repetitive sequences have been observed even
with a higher number of sequences screened [4, 22, 31].
This is in contrast to the rhizosphere soil where plant
roots directly influence the survival, growth, and activity
of bacteria in their vicinity [3]. Rhizosphere effect results
in selection of specific bacterial population in the sur-
rounding of the plant root, and this selection is plant
species specific [27]. In fact, the rhizosphere has been
recognized as an oligotrophic environment that contains
minute cells whose growth is limited by lack of substrates
[5]. A collectors curve reaching a stationary phase after
90 clones indicates the ongoing selection process in the
plant rhizosphere. In addition, sequence analysis of a
comparatively short stretch of 16S sequence (in the
present study, 433 bp) is not sufficient to affiliate se-
quences to particular bacterial species, and so grouping of
the sequences according to their taxonomic affiliations
was performed. McCaig et al. [19] constructed collectors
curve by clustering the clones into OTUs at a level of
sequence similarity >97% to quantify diversity in the
rhizospheres of grassland ecosystem.
A large number of environmental factors influence
selection of specific bacterial communities in the rhizo-
sphere. Crop species is a crucial factor for the supply of
413
energy and carbon to the heterotrophic microbial com-
munity by producing root exudates [20, 27]. Study by
Mayer et al. [18] revealed differences in the quality and
quantity of rhizodeposits from the legumes chosen for
this study. Our study aimed at comparing the bacterial
community in the rhizospheres of three commercially
important grain legumes and identifying plant-specific
rhizosphere effects.
DGGE Analysis. Each DGGE profile revealed about
2036 distinct bands. A similar number of bands using
the same primer pair in studies of rhizosphere soil has
earlier been reported [3, 20, 27]. Although several com-
mon bands occurred in DGGE patterns of the three le-
gume rhizospheres, a plant-dependent diversity of
bacterial pattern could be shown. There were differences
in the profiles with respect to the number and position of
bands, highlighting the dominance of different bacterial
groups in different plant rhizospheres. This is in accor-
dance with a previous study by Smalla et al. [27], in
which rhizospheres of strawberry, oilseed rape, and po-
tato were compared and a distinct plant-dependent rhi-
zosphere effect was reported for the three plants.
Dice coefficients were used to measure the similarity
between patterns. This measure takes into account the
matches between the profiles based on presence or ab-
sence of bands. The intensity of bands was not taken into
consideration, as silver staining is not quantitative for
comparison. More than 95% similarity between the
DGGE profiles for all plants from each legume provided
evidence of very little variation between different plants
of the same legume. Comparison of DNA and RNA
profiles for each rhizosphere revealed marked dissimi-
larities indicating a difference between the total and
metabolically active bacterial communities. RNA-derived
bands were mostly a subset of the bands detected in the
DNA-based analysis with few additional bands in RNA
profiles. This could be due to higher ribosomal RNA per
active cell, which increases the template amount for
subsequent RT-PCR [30]. Similar to a previous report by
Duineveld et al. [3], wherein a reduced number of bands
in RNA-generated profiles were reported from the rhiz-
ospheres of chrysanthemum as compared to DNA-based
profiles, our study also shows a reduction of bands in
RNA profiles when compared to the respective DGGE
profile generated by DNA. However, the extent of
reduction was lower in the present study.
Cluster analysis grouped the DNA profiles of each of
the three legumes into one cluster while RNA fingerprints
formed a separate cluster (Fig. 1). Similarity levels in the
DNA clusters were higher than that in the RNA cluster. The
responses at the rRNA level are more rapid and have greater
amplitude than those at the rDNA level, thus resulting in
greater variation among rRNA-derived profiles [3]. The
closer similarity between the profiles of Vicia and Lupinus,
414
as compared to Pisum, can be attributed to similarities with
respect to rhizodeposition reports in the former two. A
study by Mayer et al. [18] showed that 20 and 22% of the
nitrogen derived from rhizodeposition could be traced in
microbial biomass in Vicia and Lupinus rhizospheres,
respectively. In the case of Pisum, this value was reduced to
only 8%. This indicates that rhizodeposits are a major
factor in the selection of bacterial community.
The present study, by comparing profiles of total
bacteria with their respective actinomycete profile, con-
firmed a dominance of actinomycetes in the rhizospheres.
The actinomycete profiles had many dominant bands
that were common to the bands in the total bacterial
profiles. This confirms actinomycetes to be the most
dominant active group of the bacterial population.
Clone Analysis. The rhizosphere effect was very
pronounced when the distribution of different proteo-
bacterial subdivisions in the six libraries were compared.
Vicia rhizosphere libraries lacked c-subdivision members,
whereas b-subdivision members were absent from Pisum
rhizosphere libraries. The Lupinus rhizosphere was the
most diverse of the three with respect to distribution
across the different subdivisions of the proteobacteria.
This confirmed observations made by DGGE that Lupi-
nus could be the most diverse of the three plant rhizo-
spheres compared.
Sequencing of libraries revealed a proportion of
dominant population in the rhizosphere belonging to
high G+C Gram-positive bacteria, as earlier indicated in
the DGGE profiles. This group included clones similar to
some ecologically important members such as nitrogen
fixers Frankia and Clostridium and antifungal actinomy-
cetes such as Actinomadura and Geodermatophilus.
Clones similar to Arthrobacter and Microbacterium, which
are known to possess biodegradation properties, were
also present as active members. The presence of actino-
mycetes in the rhizosphere was earlier reported by Du-
ineveld et al. [3], where they were detected only at DNA
level. Dominance of actinomycetes in rRNA clone li-
braries in the present study provides strong evidence for
their presence and probable activity in the system under
study. There is increasing evidence that Gram-positive
bacteria may be more dominant in the rhizosphere than
previously supposed. McCaig et al. [19] reported that in a
clone library obtained from grass rhizospheres, Actino-
myces species were the second most abundant group after
the most frequently found a-proteobacteria. Arthrobacter
species were also found as the dominant population in
the molecular fingerprints of 16S rDNA fragments
amplified from the rhizosphere of maize [8], Medicago
sativa, and Chenopodium album [25]. Such a high pro-
portion of dominant population belonging to a diverse
range of high G + C Gram-positive bacteria, as found in
the present study, has not yet been reported from le-
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
gumes. Similar observations were reported earlier by
Smalla et al. [27] in the rhizospheres of strawberry, oil-
seed rape, and potato. In an earlier cultivation-dependent
study by Scott and Knudsen [26] on pea rhizosphere,
presence of actinomycetes and Gram-negative bacteria
was reported. Dominance of actinomycetes in the rhi-
zosphere could be attributed to the stage of sampling of
the plants [24]. The rhizosphere soil samples were taken
after the fruiting stage of the legumes when only the K-
strategic organisms would be present because of reduc-
tion in root exudation and initiation of root degradation.
The analysis was performed with mature plants. This
stage is particularly important to identify the rhizosphere
population present at the time when these legumes are
plowed into soil to enrich soil nitrogen.
Identification of bacterial populations that are pres-
ent and active in rhizospheres of legumes might help
focus future biocontrol efforts. Actinomycetes, such as
Streptomyces, act as antagonists to many different phy-
topathogenic fungi, including different pathogenic races
of the Fusarium wilt pathogen [7], Verticillium wilt [1],
and potato scab [21]. Detection of chitinolytic activity of
actinomycetes [9] also makes them promising candidates
as biocontrol agents for diverse fungal diseases. Domi-
nance of actinomycetes in the three rhizospheres opens
the possibility of exploring them to identify members
with antagonistic properties toward some economically
hazardous legume pathogens, e.g., rust fungus (Uromyces
fabae) and common root rot (Aphanomyces euteiches).
Studies addressing such potential of actinomycetes are
under investigation.
The present study revealed a hitherto unknown
diversity of rhizospheric bacteria associated with grain le-
gumes. With our experimental setup, using the same soil
material but three different legumes and a uniform inoc-
ulation with Rhizobium sp., we have clearly shown that
plant roots contribute to the development of bacterial
communities in the rhizosphere and this effect is plant
dependent. The extent of rhizosphere effect could vary in
natural field conditions, as the present study was per-
formed under controlled conditions in the greenhouse
using soil from an agricultural site. Studies addressing the
same aspects in field conditions are under investigation.
Extraction and in situ analysis of rRNA has enabled iden-
tification of active taxa in the present study. However,
studies dealing with structural diversity do not necessarily
lead to an improved functional understanding. Therefore,
these studies should be combined with more investigation
into the functions of these microbial communities in the
rhizosphere. Detection of total RNA provides an insight
into the expression of key functional genes, and studies
investigating the same are also being conducted using the
same rhizosphere soil types. Besides, analysis of fungal
population in these legume rhizospheres and the study of
younger plant rhizospheres are essential to gain a complete
S. SHARMA ET AL. BACTERIAL COMMUNITY STRUCTURE IN RHIZOSPHERE OF LEGUMES
understanding of the population dynamics in the system
under study. Similar studies on different cultivars also need
to be performed to establish the consistence of dominant
actinomycetes in the legume rhizospheres.
Acknowledgment
The study was supported by a research grant MU 831/10-
1 from the Deutsche Forschungsgemeinschaft (DFG),
Bonn, Germany.
References
1. Abd-Allah, EF (2001) Streptomyces plicatus as a model biocontrol
agent. Folia Microbiol (Praha) 46: 309314
2. Abrams, ES, Stanton, VP (1992) Use of denaturing gradient gel
electrophoresis to study conformational transitions in nucleic
acids. Methods Enzymol 212: 71104
3. Duineveld, BM, Kowalchuk, GA, Keijzer, A, Van Elsas, JD, Van
Veen, JA (2001) Analysis of bacterial communities in the rhizo-
sphere of chrysanthemum via denaturing gradient gel electro-
phoresis of PCR-amplified 16S rRNA as well as DNA fragments
coding for 16S rRNA. Appl Environ Microbiol 67: 172178
4. Felske, A, Wolterink, A, Van Lis, R, Akkermans, ADL (1998)
Phylogeny of the main bacterial 16S rRNA sequences in Drenste A
grassland soils (The Netherlands). Appl Environ Microbiol 64:
871879
5. Foster, RC (1988) Microenvironments of soil microorganisms. Biol
Fertil Soils 6: 189203
6. Gathumbi, SM, Ndufa, JK, Giller, KE, Cadisch, G (2002) Do mixed
species improved fallows increase above- and below-ground re-
sources capture? Agron J 94: 518526
7. Getha, K, Vikineswary, S (2002) Antagonistic effects of Strepto-
myces violaceusniger strain G10 on Fusarium oxysporum f.sp.cu-
bense race 4: indirect evidence for the role of antibiosis in the
antagonistic process. J Ind Microbiol Biot 28: 303310
8. Gomes, NCM, Heuer, H, Schonfeld, J, Costa, R, Hagler-Mendo-
nca, L, Smalla, K (2001) Bacterial diversity of the rhizosphere of
maize (Zea mays) grown in tropical soil studied by temperature
gradient gel electrophoresis. Plant Soil 232: 167180
9. Gomes, RC, Semedo, LT, Soares, RM, Alviano, CS, Linhares, LF,
Coelho, RR (2000) Chitinolytic activity of actinomycetes from a
cerrado soil and their potential in biocontrol. Lett Appl Microbiol
30: 146150
10. Griffiths, BS, Ritz, K, Ebblewhite, N, Dobson, G (1999) Soil
microbial community structure: effects of substrate loading rates.
Soil Biol Biochem 31: 145153
11. Griffiths, RI, Whiteley, AS, ODonnell, AG, Bailey, MJ (2000)
Rapid method for co-extraction of DNA and RNA from natural
environments for analysis of ribosomal DNA- and rRNA-based
microbial community composition. Appl Environ Microbiol 66:
54885491
12. Haynes, RJ, Martin, RJ, Goh, KM (1993) Nitrogen fixation,
accumulation of soil nitrogen and nitrogen balance for some field-
grown legume crops. Field Crop Res 35: 8592
13. Heuer, H, Smalla, K (1997) Application of denaturing gradient gel
electrophoresis and temperature gradient gel electrophoresis for
studying soil microbial communities. In: Van Elsas, JD, Trevors,
JT. Wellington, EMH (Eds.) Modern Soil Microbiology, Marcel
Dekker, New York, pp 353373
14. Heukeshoven, J, Dernick, R (1986) Neue Ergebnisse zum Mecha-
nismus der Silberfarbung. In: Radola, BJ (Ed.) Electrophorese
Forum86, Technische Universitat Munchen, Munich, pp 2227
415
15. Hynes, RK, Jans, DC, Bremer, E, Lupwayi, NZ, Rice, WA, Clayton,
GW, Collins, MM (2001) Rhizobium population dynamics in the
pea rhizosphere of rhizobial inoculant strain applied in different
formulations. Can J Microbiol 47: 595600
16. Jaeger III, CH, Lindow, SE, Miller, W, Clark, E, Firestone, MK
(1999) Mapping of sugar and amino acid availability in soil around
roots with bacterial sensors of sucrose and tryptophan. Appl
Environ Microbiol 65: 26852690
17. Lorenz, M, Wackernagel, W (1987) Adsorption of DNA to sand
and variable degradation rates of adsorbed DNA. Appl Environ
Microbiol 53: 29482952
18. Mayer, J, Buegger, F, Jensen, ES, Schloter, M, He, J (2003) Esti-
mating N rhizodeposition of grain legumes using a 15N in situ
labelling method. Soil Biol Biochem 35: 2128
19. McCaig, AE, Glover, LA, Prosser, JI (1999) Molecular analysis of
bacterial community structure and diversity in unimproved and
improved upland grass pastures. Appl Environ Microbiol 65:
17211730
20. Miethling, R, Wieland, G, Backhaus, H, Tebbe, CC (2000) Varia-
tion of microbial rhizosphere communities in response to crop
species, soil origin and inoculation with Sinorhizobium meliloti
L33. Microb Ecol 41: 4356
21. Neeno-Eckwall, EC, Kinkel, LL, Schottel, JL (2001) Competition
and antibiosis in the biological control of potato scab. Can J
Microbiol 47: 332340
22. Nusslein, K, Tiedje, JM (1998) Characterization of the dominant
and rare members of a young Hawaiian soil bacterial community
with small-subunit ribosomal DNA amplified from DNA frac-
tionated on the basis of its guanine and cytosine composition.
Appl Environ Microbiol 64: 12831289
23. Odum, EP (1971) Principles and concepts pertaining to organi-
zation at the community level. In: Fundamentals of Ecology.
Saunders College Publishing, Philadelphia, pp 140161
24. Srensen, J (1997) The rhizosphere as a habitat for soil organisms
In: Van Elsas, JD, Trevors, JT, Wellington, EMH (Eds.) Modern
Soil Microbiology, Marcel Dekker, New York, pp 2145
25. Schweiger, F, Tebbe, CC (2000) Effect of field inoculation with
Sinorhizobium meliloti L33 on the composition of bacterial com-
munities in rhizospheres of target plant (Medicago sativa) and a
non-target plant (Chenopodium album)-linking of 16S rRNA gene-
based single-strand conformation polymorphism community
profiles to the diversity of cultivated bacteria. Appl Environ
Microbiol 66: 35563565
26. Scott, JS, Knudsen, GR (1999) Soil amendment effects of rape
(Brassica napus) residues on pea rhizosphere bacteria. Soil Biol
Biochem 31: 14351441
27. Smalla, K, Wieland, G, Buchner, A, Zock, A, Parzy, J, Kaiser, S,
Roskot, N, Heuer, H, Berg, G (2001) Bulk and rhizosphere soil
bacterial communities studied by denaturing gradient gel electro-
phoresis: plant-dependent enrichment and seasonal shifts revealed.
Appl Environ Microbiol 67: 47424751
28. Tan, Z, Hurek, T, Vinuesa, P, Muller, P, Ladha, JK, Reinhold-
Hurek, B (2001) Specific detection of Bradyrhizobium and Rhizo-
bium strains colonizing rice (Oryza sativa) roots by 16S-23S
ribosomal DNA intergeneric spacer-targeted PCR. Appl Environ
Microbiol 67: 36553664
29. von Wintzingerode, F, Gobel, UB, Stackerbrandt, E (1997) Fate and
activity of microbial diversity in environmental samples: pitfalls of
PCR-based rRNA analysis. FEMS Microbiol Rev 21: 213229
30. Wagner, R (1994) The regulation of ribosomal RNA synthesis and
bacterial cell growth. Arch Microbiol 161: 100106
31. Zhou, J, Davey, E, Figueras, JB, Rivkina, E, Gilichinsky, D, Tiedje,
JM (1997) Phylogenetic diversity of a bacterial community deter-
mined from Siberian tundra soil DNA. Microbiology 143: 3913
3919