Download Bacterial community composition in the rhizosphere of a transgenic

FEMS Microbiology Ecology 40 (2002) 29^37
www.fems-microbiology.org
Bacterial community composition in the rhizosphere of a transgenic,
herbicide-resistant maize (Zea mays) and comparison to its
non-transgenic cultivar Bosphore
Achim Schmalenberger, Christoph C. Tebbe
Institut fu«r Agraro«kolologie, Bundesforschungsanstalt fu«r Landwirtschaft (FAL), Bundesallee 50, 38116 Braunschweig, Germany
Received 13 September 2001; received in revised form 8 January 2002; accepted 8 January 2002
First published online 4 March 2002
Abstract
Bacterial communities in rhizospheres of transgenic maize (Zea mays, with the pat-gene conferring resistance to the herbicide
glufosinate; syn. L-phosphinothricin) were compared to its isogenic, non-transgenic cultivar. Total DNA was extracted from bacterial cell
consortia collected from rhizospheres of plants grown in an agricultural field. With the use of three different primer pairs binding to
evolutionarily conserved regions of the bacterial 16S rRNA gene, partial sequences were amplified by polymerase chain reaction (PCR).
The PCR products were subjected to single-strand conformation polymorphism (SSCP) to generate genetic profiles which corresponded to
the diversity of the amplified sequences. Genetic profiles of rhizospheres consisted of 40^60 distinguishable bands depending on the
chosen primer pairs, and the variability between independent replicates was very low. Neither the genetic modification nor the use of the
herbicide Liberty (syn. Basta; active ingredient: glufosinate) affected the SSCP profiles as investigated with digital image analysis. In
contrast, PCR^SSCP profiles of bacterial communities from rhizospheres of sugar beet, grown in the same field as a control crop, were
clearly different. A less pronounced but significant difference was also observed with rhizosphere samples from fine roots of maize plants
collected 35 and 70 days after sowing. Sequencing of the dominant 30 products from one typical SSCP profile generated from transgenic
maize rhizospheres indicated the presence of typical soil and rhizosphere bacteria: half of the bands could be attributed to Proteobacteria,
mainly of the K- and L-subgroups. Other SSCP bands could be assigned to members of the following phylogenetic groups: Cytophaga^
Flavobacterium^Bacteroides, Chlamydiales^Verrucomicrobium, Planctomyces, Holophaga and to Gram-positive bacteria with a high G+C
DNA content. = 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
Keywords : Bacterial diversity ; Polymerase chain reaction^single-strand conformation polymorphism; Genetic pro¢le; Glufosinate ; Transgenic Zea mays;
Rhizosphere
1. Introduction
Genetic engineering techniques extend the possibilities
to develop crops with improved properties. In this context,
transgenic crops like maize, rape and sugar beet were developed which are resistant to the herbicidal compound
glufosinate (syn. L-phosphinothricin), an L-amino acid
that inhibits glutamine synthetase [1,2]. It is phytotoxic
by interfering with photophosphorylation through accumulation of intracellular NHþ
4 [3,4] but its toxicity to
mammals is rather low [5,6]. Some microbial activities,
including growth of several soil bacteria [7] and of Esche-
* Corresponding author. Tel. : +49 (531) 596 2553 ;
Fax : +49 (531) 596 2699.
E-mail address : [email protected] (C.C. Tebbe).
richia coli [8] or the nodulation of Lucerne (alfalfa) roots
by Sinorhizobium meliloti [9] are sensitive to concentrations of 1 mM or less under laboratory conditions. In
contrast, it has been reported that many bacteria in surface or subsurface soil are resistant to glufosinate
[7,8,10,11]. In soil, glufosinate is quickly degraded by deamination and decarboxylation to a non-toxic intermediate, which is further mineralised [12^15]. Laboratory experiments have indicated that microbial activities in soil
will not dramatically be a¡ected by glufosinate at concentrations which are relevant for agricultural practice
[9,11,16]. However, the application of herbicides like Liberty or Basta, which contain glufosinate, in the context of
the cultivation of transgenic plants (‘Liberty Link’) has
not to our knowledge been studied. Independent of the
herbicide, the transgenic plant may have unintended modi¢ed properties. The expression rates of a recombinant gene
0168-6496 / 02 / $22.00 = 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII : S 0 1 6 8 - 6 4 9 6 ( 0 2 ) 0 0 1 9 6 - 4
FEMSEC 1336 13-5-02
30
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
can vary dramatically depending on the chromosomal insertion site [17] and other genes, e.g. those in proximity to
the insertion site, may also be a¡ected. Should such unintended e¡ects have any impact on soil microbial activity, it
would likely be seen ¢rst in the rhizosphere.
The composition of microbial communities in rhizospheres is governed mainly by the quality and quantity
of carbon sources that are released as root exudates
[18,19]. Thus, an altered composition of root exudates
may select a di¡erent community of rhizosphere microorganisms. Even small modi¢cations, as may exist between
di¡erent cultivars of the same plant species, can result in
the selection of di¡erent microbial communities in the rhizosphere [20,21]. In studies with rape (canola ; Brassica
napus), the composition of rhizosphere bacteria of a transgenic cultivar could be distinguished from other non-engineered cultivars [22,23]. In another study with potatoes,
however, the composition of microbial communities in
rhizospheres was not signi¢cantly a¡ected by genetic engineering [24,25]. It can be suspected that, generally, unintended modi¢cations of rhizosphere-inhabiting communities are possible but that the degree of variation will be
in£uenced by both the plant species and the type of modi¢cation.
Here we report results of a ¢eld study in which transgenic, glufosinate-resistant maize and isogenic, non-engineered control plants were grown under conditions common for agricultural practice regarding the soil
management, crop rotation, seeding, fertilisers, and use
of pesticides. Our objective was to compare the composition of rhizosphere microbial communities of both cultivars. In order to distinguish between e¡ects caused by the
genetic modi¢cation and those caused by di¡erent herbicide treatments, we also analysed the transgenic cultivar
treated with conventional herbicides. Since there is ample
evidence that soil and rhizospheres harbour a much larger
diversity of microorganisms than those which can be cultivated in the laboratory on growth media [26,27], community analysis was carried out using cultivation-independent methods. Partial 16S rRNA genes were ampli¢ed
from bacterial community DNA by PCR, using primers
which bind to evolutionarily conserved regions within this
gene [28]. The diversity of PCR-ampli¢ed products was
transformed to genetic pro¢les (‘¢ngerprints’) using the
community single-strand conformation polymorphism
(PCR^SSCP) approach [29].
plants or treatments. Maize plants were collected from
three treatments: transgenic maize (KX8445 ; transformed
with the modi¢ed bacterial pat-gene [30] encoding for
phosphinothricin-acetyltransferase [1]) treated either with
conventional herbicides (Artett, active ingredients: terbuthylazine and bentazon; and Motivell, active ingredient:
nicosulfuron) or Liberty (active ingredient: glufosinate),
and isogenic maize (KX6345; registered in France under
the name ‘Bosphore’) treated with conventional herbicides
(Artett and Motivell). Conventional herbicides were obtained from Syngenta Agro (Maintal, Germany). The
maize seeds and the herbicide Liberty were obtained
from Aventis Crop Science (Frankfurt, Germany). The
¢eld plots were sown in the year 1999 with maize on
May 5, whereas sugar beet plants used as a control crop
were sown on April 19.
2.2. Sampling and extraction of bacterial cells
If not otherwise stated, maize plants were collected
7 days after herbicide application corresponding to
35 days after sowing. Shoot heights at the time of sampling were approx. 25 cm. In addition, some samples were
also taken 70 days after sowing when shoot heights were
approx. 140 cm. Sugar beet plants were collected 74 days
after sowing, in the same month as the 70-day sampling of
maize took place. The plants were taken from random
positions of a plot. For each ¢eld plot, three composite
samples were collected. Each of the 35-day-old samples
consisted of eight individual plants. For samples taken
after 70 days, six plants were su⁄cient. Each treatment
consisted of three independent plots (replicates) which
were all sampled. Thus, a total of nine samples were analysed for each treatment.
Roots were detached from the green plant material in
the laboratory and dipped into tap water to remove larger
soil aggregates. The root-adhering water was carefully removed with autoclaved paper towels and ¢ne roots were
carefully collected with sterile forceps and transferred into
sterile tubes (total volume : 50 ml; Sarstedt, Nu«mbrecht,
Germany). A total of 20 ml sterile saline (0.85% NaCl in
water) was added to each tube and the tubes were incubated for 30 min at 4‡C and 20 rpm in an overhead shaker
(KH, Guwina-Ho¡mann, Berlin, Germany). Equal
amounts of the suspension were transferred to two fresh
tubes and the bacterial cells were collected by centrifugation at 4100Ug for 30 min at 4‡C. The supernatants were
discarded and the pellets were stored at 370‡C.
2. Materials and methods
2.3. DNA extraction
2.1. Field site and agricultural practice
The experimental ¢eld was located in BraunschweigVo«lkenrode. The total area of the ¢eld was 3.5 ha and
consisted of three independent replicates, each composed
of 36 di¡erent ¢eld plots (each 10U12.5 m) with di¡erent
Sterile lysis bu¡er (12 ml; 0.05 M Tris^HCl, 0.01 M
Na2 EDTA, 0.05 M NaCl, pH 8.0) was directly pipetted
onto the frozen pellets and the suspensions were subjected
to ¢ve cycles of freeze^thawing, each consisting of 5-min
freezing in liquid nitrogen, 5 min at 65‡C, and vortexing
FEMSEC 1336 13-5-02
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
for 10 s at the highest setting (VF2, IKA Labortechnik,
Stau¡en, Germany). Afterwards, proteinase K (0.28 mg
ml31 ; Roche, Mannheim, Germany) digestion was conducted at 65‡C for 1 h. Crude DNA was extracted by
phenol^chloroform extraction as previously described
[29] and stored at 320‡C in a ¢nal volume of 120 Wl TE
[31] sample31 . Crude DNA was further puri¢ed with the
Wizard DNA Puri¢cation Kit (Promega, Mannheim, Germany) with 10^30 Wl DNA per tube and eluted with 40 Wl
75‡C pre-warmed 10 mM Tris^HCl bu¡er (pH 8.0) [31].
2.4. PCR ampli¢cations of partial 16S rRNA genes
PCR was performed with the thermal cycler Primus 96
(MWG-Biotech, Ebersberg, Germany). Ampli¢cations
from environmental samples were processed in a ¢nal volume of 100 Wl containing 5 U Platinum-Taq-Polymerase
(Gibco-Lifetech, Invitrogen, Karlsruhe, Germany), onefold PCR-bu¡er supplied by the manufacturer, 1.5 mM
MgCl2 , 0.5 WM of each primer, 200 WM of each desoxynucleotide (Amersham Pharmacia Biotech, Freiburg, Germany) and 2 Wl of puri¢ed DNA extracted from rhizospheres.
Three primer pairs were chosen for the ampli¢cation of
di¡erent partial sequences of the 16S rRNA gene (for
numbering of variable regions, see [32]): (i) Com1 and
Com2 corresponding to E. coli positions 519^537 and
907^926 [29], bordering the variable regions 4 and 5; (ii)
f986 and r1346 corresponding to E. coli positions 968^986
and 1330^1346 ([33]; but without a GC-clamp) generating
PCR products which included variable regions 6, 7, and 8;
(iii) f120 and r518 corresponding to E. coli positions 101^
120 and 518^537 with variable regions 2 and 3 [34]. All
reverse primers were phosphorylated at the 5P-end for subsequent single-strand digestion (see next paragraph). Primers were synthesised by Gibco Lifetech (Invitrogen) or
MWG Biotech (Ebersberg, Germany). The PCR conditions were as follows : 95‡C for 3 min, followed by 35
cycles of 1 min 95‡C, 50‡C for 1 min, 72‡C for 70 s
and, ¢nally, 72‡C for 5 min.
2.5. Genetic pro¢ling by SSCP
PCR products were converted to single-stranded products by removal of the reverse phosphorylated strands
with lambda-exonuclease, as described by Schwieger and
Tebbe [29,35], except that incubation times were reduced
to 45 min per sample. Precipitated DNA samples were
resuspended in 8 Wl denaturing loading bu¡er (95% formamide, 10 mM NaOH, 0.025% bromophenol blue and
xylene cyanol). Samples were incubated for 2 min at
95‡C and then chilled on ice. Electrophoresis was conducted in non-denaturing polyacrylamide gels (MDE,
FMC Bioproducts, Rockland, ME, USA), under conditions which have been described elsewhere [34], and
DNA was visualised by silver staining [36].
31
For cloning and sequencing, single bands of community
pro¢les were excised with razor blades and the singlestranded DNA was eluted from the gel material by ‘crush
and soak’ [31]. Eluted DNA was precipitated with ethanol,
centrifuged, dried and ¢nally resuspended in 10 Wl of 10
mM Tris^HCl bu¡er (pH 8.0). For reampli¢cation, 2 Wl of
the eluted products was used as a template in PCR, conducted under conditions described before for community
analysis, except that the ¢nal volume was 50 Wl with 1.25 U
of Platinum-Taq-polymerase. Half of the PCR products
were converted to single-stranded DNA by lambda-exonuclease, as described above, in order to con¢rm the purity
and identity on SSCP gels. The remaining half of the PCR
products was used for DNA cloning and sequencing, as
described elsewhere [34,37].
2.6. Digital image analysis of SSCP pro¢les
The GelCompar programme package (version 4.1; Applied Maths, Kortrijk, Belgium) was used to analyse the
similarity of SSCP patterns on each gel. Pro¢les were normalised using species standards on each gel as a reference
and the recommended background subtraction procedure
was then applied to increase the comparability of the single pro¢les. Calculation of the similarity matrix was based
on Pearson correlation coe⁄cients. The clustering method
was UPGMA (unweighted pair group method with arithmetic averages). The similarity was also analysed by bandbased coe⁄cients to evaluate the stability of clusters.
2.7. Phylogenetic analyses and nucleotide sequence
accession numbers
Nucleotide sequences were edited with the AlignIR 1.1
programme (Li-Cor, Lincoln, NE, USA). Primer sequences were removed and consensus sequences generated in
this study were deposited in the GenBank database (accession numbers are listed in Table 1). For phylogenetic
placement the sequences were loaded into the ARB database (W. Ludwig, Munich, Germany; www.arb-home.de).
Distance matrices were calculated with 50% conservation
¢lter for each phylogenetic group [38].
3. Results
3.1. SSCP pro¢les of partial sequences
A total of 27 genetic pro¢les was generated corresponding to (i) three pro¢les ampli¢ed from DNA extracted
from three composite samples taken from each plot, to
analyse the variation of the extraction e⁄ciency, etc., (ii)
three plots for each treatment to analyse the variation of
replicates, and (iii) three treatments. The number of samples exceeded the number that could be analysed on the
same gel. Therefore, the pro¢les were analysed on three
FEMSEC 1336 13-5-02
32
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
Fig. 1. SSCP pro¢les of PCR-ampli¢ed partial 16S rRNA genes. Products include variable regions 2 and 3 (a), 6, 7 and 8 (b), and 4 and 5 (c). The pro¢les were derived from the rhizosphere of the non-transgenic maize cultivar Bosphore treated with conventional herbicides (A), the corresponding transgenic cultivar treated with conventional herbicides (B), and with the herbicide Liberty (glufosinate) (C). Each variable region was analysed on three independent SSCP gels. The three single gels were loaded with samples from separate replicate ¢eld plots. Each gel is bordered by standards consisting of
single-stranded DNA ampli¢ed from the following species (top to bottom) by PCR with Com primers (see Section 2): Bacillus licheniformis (not seen in
panel b), Rhizobium trifolii, Flavobacterium johnsonae, Rhizobium radiobacter (double band).
di¡erent gels, run under the same conditions, with each gel
carrying samples of all treatments and all replicate plots.
With primers amplifying products with variable regions 2
and 3, pro¢les were generated consisting of about 50 dis-
tinct bands (Fig. 1a). The pro¢les (patterns) of all samples
on one gel were highly similar. Digital image analysis
could not di¡erentiate them according to their treatments
(data not shown). In addition, Fig. 1a shows that the
FEMSEC 1336 13-5-02
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
33
Fig. 2. SSCP-pro¢le analyses of partial 16S rRNA gene sequences with variable regions 4 and 5 ampli¢ed by PCR and evaluated by digital image analysis. a: Bacterial communities from rhizospheres of ¢ne roots collected from maize plants 35 days (1) and 70 days (2) after sowing; b: comparison of
rhizosphere communities from sugar beet (S) and maize plants (M). A, B, and C: Treatments as described in the legend of Fig. 1.
pro¢les on di¡erent gels were also highly similar, indicating that the conditions for SSCP (gel matrix, electrophoresis) were highly reproducible. Most of the variation between pro¢les was caused by di¡erent intensities of silverstained bands, which was most likely a result of slightly
di¡erent amounts of total DNA loaded into each lane.
For the ampli¢cation of products containing variable
regions 6, 7, and 8, the selected primer pairs generated
SSCP pro¢les consisting of more than 40 distinguishable
bands (Fig. 1b). As found for partial sequences with variable regions 2 and 3, no treatment e¡ect or other signi¢cant di¡erences, e.g. according to the location of the single
plots in the ¢eld, could be detected. The variation of patterns from di¡erent gels was also low. Digital image analysis clustered group A on the left gel of Fig. 1b into a
separate group (data not shown) but this result was obviously triggered by the lower amount of total DNA from
PCR products loaded onto the gel. Other group-A replicates did not separate out into uniform clusters.
Each SSCP pro¢le generated with the use of the primer
set amplifying the variable regions 4 and 5 consisted of
about 60 distinguishable bands (Fig. 1c). As reported for
the other two partial rRNA gene products, all patterns
were highly similar and could not be distinguished according to their treatment or any other ¢eld-associated factor.
In addition to sampling 35 days after sowing we also collected maize plants 70 days after sowing. The patterns of
the samples looked similar to those generated from the 35day-old samples, but with digital image analysis, two clusters became distinguishable in accordance to the sampling
dates (Fig. 2a). The patterns of the treatments within each
cluster looked highly similar, indicating that also after 70
days no treatment e¡ect was detectable and that the rhizo-
sphere communities were very similar. Data from the 70day-old samples are shown in Fig. 2a with only one replicate, but all three replicates were analysed on separate
SSCP gels.
As a control for the sensitivity of SSCP, we compared
pro¢les of maize rhizospheres to those obtained from sugar beet plants grown in the same ¢eld in separate plots.
Rhizosphere communities of 74-day-old sugar beet plants
were compared to those of 70-day-old maize plants. Patterns of similar complexity generated from sugar beet rhizospheres were clearly di¡erent from those found with
maize (Fig. 2b). As indicated by the scale bars, the di¡erences between pro¢les were more pronounced than those
between maize plants of di¡erent age.
3.2. Identi¢cation of SSCP bands
In order to identify which organisms were responsible
for the SSCP patterns, dominant bands of a single pro¢le
with partial sequences carrying the variable regions 4 and
5 were selected. Since pro¢les of all treatments were highly
similar, only one single pro¢le generated from transgenic
maize rhizosphere was selected for band identi¢cation.
Dominant bands were cut out (Fig. 3) and reampli¢ed in
order to con¢rm the position of the targeted band and, in
addition, to test whether some bands were metastable conformers, i.e. that they actually contained the same sequence, but in a di¡erent conformation [34]. In fact,
band numbers 6, 7, 13, and 14 (Fig. 3) resulted in identical
double bands and, thus, were considered to be metastable
conformers.
Reampli¢cation of band numbers 10 and 15 resulted for
each band in two separately reampli¢able bands, one in
FEMSEC 1336 13-5-02
34
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
Table 1
Identi¢cation of PCR-ampli¢ed partial 16S rRNA gene sequences isolated from SSCP pro¢les generated from bacterial communities extracted from rhizosphere of a transgenic (pat) ¢eld-grown maize cultivar
Phylogenetic group
Clonea
Closest relative, GenBank
accession number
K-Proteobacteria
29
11
18
Blastobacter sp., Z23157
K-Proteobacterium, AF236001
Rhizosphere soil bacterium,
AJ252702
Methylobacterium rhodinum,
D32229
Caulobacter henricii,
AJ227758
Uncultured bacterium,
AF234063
Humic substances enrichment
clone, AF231448
[Alcaligenes] latus, D88007
Agricultural soil clone, AJ252648
Herbaspirillum seropedicae,
AJ238361
Telluria mixta, X65589
Uncultured eubacterium, AF047645
Q-Proteobacterium, U15111
Cellvibrio mixtus, AJ289160
Serratia plyuthica, AJ233433
Cytophaga sp., AF260716
Uncultured bacterium, AF087043
Cytophagales strain, AB022889
Rhizosphere soil bacterium,
AJ252690
Agricultural soil bacterium,
AJ252610
Potato plant root bacterium, AJ2522723
Agricultural soil bacterium,
AJ252615
Uncultured bacterium, AF234138
Agricultural soil bacterium,
AJ252626
Rhizosphere soil bacterium,
AJ252689
Uncultured bacterium, AF083615
Planctomycete strain, AJ231182
Uncultured acidobacterium, AF200696
7EZ9517, Z95724
Zea mays chloroplast, J01422
16
28
5
L-Proteobacteria
2
3
25
14ab
L/Q-Proteobacteria
Q-Proteobacteria
CFB
22
17
10ac
1
9
26
15ac
24
19
12
15bc
20
High G+C Gram-positive
Chlamydiales^Verrucomicrobium
8
27
23
Planctomyces
Holophaga
Plant organelles
30
21
7
14bb
10bc
Similarity to closest
relative (%)
Accession number
99.4
91.0
99.2
AJ308291
AJ308273
AJ308279
96.7
AJ344449
99.2
AJ308290
98.6
AJ308293
99.7
AJ308267
99.2
98.6
99.1
AJ308268
AJ308287
AJ308275
98.8
95.9
96.5
98.4
100
100
94.1
94.9
99.1
AJ308284
AJ308280
AJ308272
AJ308266
AJ308271
AJ308288
AJ308277
AJ308286
AJ308281
94.3
AJ308274
97.1
96.6
AJ308278
AJ308282
94.0
99.7
AJ308270
AJ308289
99.7
AJ308285
90.8
95.4
97.4
98.0
100
AJ308292
AJ308283
AJ308269
AJ308276
AJ308294
a
Corresponding to band positions indicated in Fig. 3.
Two di¡erent sequences obtained from the same band.
c
Isolation and reampli¢cation resulted in two distinct bands which were both sequenced.
b
the position which was expected and another one at a
di¡erent position. Sequences of both products could be
attributed to a di¡erent organism (Table 1). Two di¡erent
sequences from band number 14 were found when clones
were tested for inserted products during the cloning procedure in E. coli.
The nucleotide sequences of SSCP bands showed closest
relatives in the range of 90.8^100% (Table 1). The largest
number of sequenced products (50%) was related to organisms of the Proteobacteria with members of the K- and
L-subgroups more abundant than those of the Q-subgroup.
A total of 23% of the sequenced products matched to
closest relatives in the Cytophaga^Flavobacterium^Bacte-
roides (CFB) group. In addition, we found other phylogenetic groups: Chlamydiales^Verrucomicrobium, Holophaga, Planctomyces and Gram-positive bacteria with a high
G+C DNA content. Interestingly, band number 10 in the
pro¢le contained one sequence (10 b) of a maize plastid
rRNA gene, which must have been released from root
material during the extraction procedure.
4. Discussion
Our study aimed at detecting altered bacterial communities in response to a genetic modi¢cation of maize and to
FEMSEC 1336 13-5-02
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
Fig. 3. SSCP pro¢les of PCR-ampli¢ed partial 16S rRNA genes with
variable regions 4 and 5. Left lane, after isolation of bands for reampli¢cation and sequencing. Each number represents a targeted band position. The pro¢les were independent replicates, generated from DNA of
bacterial communities extracted from the rhizospheres of the transgenic
maize cultivar treated with conventional herbicides.
di¡erent herbicide applications using cultivation-independent PCR^SSCP to detect the bacterial community in rhizospheres. The SSCP method is, as are other pro¢ling
techniques such as DGGE or TGGE, currently restricted
to DNA sizes of a maximum length of approx. 500 bp for
most applications [35,39]. Thus, only partial sequences of
the 16S rRNA gene (total length approx. 1520 bp) were
analysed. To increase the information, each rhizosphere
community was analysed with three di¡erent primer pairs
amplifying di¡erent variable regions [34]. However, neither the ¢eld application of di¡erent herbicides nor the
genetic modi¢cation by insertion and expression of the
pat-gene a¡ected the composition of the bacterial community in rhizospheres of maize as detected by SSCP.
35
SSCP pro¢les were highly similar and digital image analysis could not identify groups with similar pro¢les according to the di¡erent treatments. Pro¢les of independent
replicates di¡ered in single bands (see Fig. 3) but this
variation did not correlate with experimental variables.
All plots were on the same ¢eld but due to the large
size, the replicate plots were located up to 150 m apart
from each other. Thus, ¢eld heterogeneity was not detected in the community composition in the rhizospheres
of maize or sugar beet. Since the SSCP-community patterns of maize and sugar beet were clearly di¡erent from
each other it can be concluded that each plant species
selected its own speci¢c bacterial community from the reservoir of bacteria present in the ¢eld soil or on seeds.
These results corroborate other studies based on rRNAgene pro¢ling techniques and community-level physiological pro¢les which demonstrated that plant species are
more important in the selection of bacterial communities
in rhizospheres than other factors, i.e. soil origin or agricultural treatments [40,41]. Maize plants collected 35 days
after sowing harboured a di¡erent bacterial community in
their rhizosphere than those collected after 70 days. Since
only ¢ne roots were compared, we assume that this modi¢cation was caused by di¡erent compositions or quantities of root exudates during di¡erent growth stages. Similar age-dependent e¡ects were also observed rhizospheres
of maize grown in tropical soil [42,43].
Generally, pro¢ling techniques can be useful to detect
changes in community structures after herbicide applications. Altered TGGE patterns were found in response to
applications of the herbicides Dinoterb and Metamitron,
but the application rates were above those recommended
for agricultural practices [44]. Phenylurea herbicides affected DGGE patterns of bacteria in soil that was treated
with these herbicides for more than 10 years [45]. The fact
that we could not detect any di¡erence between rhizosphere communities from glufosinate-treated plants and
those treated with other herbicides indicates that concentrations chosen in agricultural practices are tolerable to
most soil bacteria and will not result in dramatic changes.
In addition, the high rate of glufosinate degradation in soil
[12] may have resulted in only a short exposure of that
compound to soil bacteria.
Two cultivars, one transgenic and one non-transgenic
were compared. The insertion of the pat-gene confers the
capability of the transgenic plant to detoxify glufosinate
by N-acetylation. The detoxi¢cation takes place in the
green parts of the plant and no direct e¡ect on root exudation can be expected with this genetic modi¢cation.
Nevertheless, in the context of risk-assessment studies
with transgenic plants, it is important to understand how
much a microbial community composition in rhizospheres
di¡ers between cultivars of the same species. In fact, in a
thorough cultivation-based approach such di¡erences were
found between a rhizosphere of a herbicide-tolerant, transgenic rape and a non-transgenic cultivar [23], the former
FEMSEC 1336 13-5-02
36
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
showing a decrease of Arthrobacter and Bacillus isolates.
Assuming that such an alteration would not be restricted
to the culturable part of the bacterial community, a similar
change would have been detected with PCR^SSCP. Therefore, our results indicate that no comparable change in
species diversity existed between both maize cultivars.
However, our results cannot be extrapolated beyond the
species level, i.e. the resolution of rRNA genes. Di¡erences
of this kind can exist between cultivars of the same plant
species, as shown for Burkholderia cepacia strains in maize
rhizospheres [46].
Sequencing of SSCP bands indicated that the dominant
bacterial organisms in the rhizosphere were members of
the Proteobacteria, in accordance with other studies on
the bacterial diversity in maize rhizospheres [27,34]. It
cannot be ruled out that the freeze^thaw lysis procedure
selected in our study failed to release high amounts of
DNA from bacteria with lysis-resistant cell walls and,
thus, Gram-positive bacteria may have been underrepresented [47]. In fact, we could not detect the common maize
rhizosphere coloniser Paenibacillus polymyxa [48]. In a
study where bead-beating was used for lysis, several sequences of Gram-positive bacteria with a low G+C
DNA content were recovered from maize rhizospheres
after PCR and cloning of partial 16S rRNA genes [27].
In addition to the lysis procedure, the selection of variable
regions which are PCR-ampli¢ed is an important factor.
In a previous study we found that in SSCP pro¢les, 14^
40% of the sequenced bands could be attributed to bacteria with a low G+C DNA content depending on the
selection of primers and ampli¢ed variable regions [34].
Even though most nucleotide sequences were not 100%
identical to those found in the database, the identi¢cation
of closest relatives indicated a high correlation to typical
soil and rhizosphere inhabiting bacteria. The importance
of bacteria from the CFB group, which include typical
gliders and many cellulose-degrading organisms, from
the group of Planctomyces and Chlamydiales^Verrucomicrobium and of the Holophaga group was con¢rmed. In
contrast to a previous study [34] and recent work [27] we
could not detect any sequence of an Archaea.
In summary, this study shows that cultivation-independent PCR^SSCP can be a useful tool in risk-assessment
studies to analyse the immediate impact of plants on the
diversity of soil bacteria in rhizospheres. The fact that no
e¡ect between transgenic and non-transgenic plants was
detected is not a weakness of the method but more likely
an indication of the minor physiological changes caused
by the insertion of the pat-gene into the maize genome. It
remains to be determined if community changes are more
common beyond the level of 16S rRNA gene resolution.
For ecological risk identi¢cation, however, the observation
of shifts at the strain level is not su⁄cient. A positive
identi¢cation of a risk would also need to demonstrate
changes at the functional level.
Acknowledgements
Karin Trescher supported this work with her excellent
technical assistance which we gratefully acknowledge. The
work was funded by the Germany Ministry for Education
and Research (bmb+f), Projekttra«ger Biologie, Energie,
Umwelt (BEO), Grant number 0311740.
References
[1] Hoerlein, G. (1994) Glufosinate (phosphinothricin), a natural amino
acid with unexpected herbicidal properties. Rev. Environ. Contam.
Toxicol. 138, 73^145.
[2] Abell, L.M. and Villafranca, J.J. (1991) Investigation of the mechanism of phosphinothricin inactivation of Escherichia coli glutamine
synthetase using rapid quench kinetic technique. Biochemistry 30,
6135^6141.
[3] Sauer, H., Wild, A. and Ru«hle, W. (1987) The e¡ect of phosphinothricin (glufosinate) on photosynthesis. II: The cause of inhibition of
photosynthesis. Z. Nat.forsch. 42c, 270^287.
[4] Wild, A., Sauer, H. and Ru«hle, W. (1987) The e¡ect of phosphinothricin (glufosinate) on photosynthesis. I. Inhibition of photosynthesis and accumulation of ammonia. Z. Nat.forsch. 42c, 263^269.
[5] Ebert, E., Leist, K.H. and Mayer, D. (1990) Summary of safety
evaluation toxicity studies of glufosinate ammonium. Food Chem.
Toxicol. 28, 339^349.
[6] Hack, R., Ebert, E., Ehling, G. and Leist, K.H. (1994) Glufosinate
ammonium ^ some aspects of its mode of action in mammals. Food
Chem. Toxicol. 32, 461^470.
[7] Quinn, J.P., Heron, J.K. and McMullan, G. (1993) Glufosinate tolerance and utilisation by soil and aquatic bacteria. Biol. Environ.
Proc. R. Ir. Acad. 93B, 181^186.
[8] Bartsch, K. and Tebbe, C.C. (1991) Initial steps in the degradation of
phosphinothricin (glufosinate) by soil bacteria. Appl. Environ. Microbiol. 55, 711^716.
[9] Kriete, G. and Broer, I. (1996) In£uence of the herbicide phosphinothricin on growth and nodulation capacity of Rhizobium meliloti.
Appl. Microbiol. Biotechnol. 46, 580^586.
[10] Tebbe, C.C. and Reber, H.H. (1988) Utilization of the herbicide
phosphinothricin as a nitrogen source by soil bacteria. Appl. Microbiol. Biotechnol. 29, 103^105.
[11] Allen-King, R.M., Butler, B.J. and Reichert, B. (1995) Fate of the
herbicide glufosinate-ammonium in the sandy, low-organic-carbon
aquifer at CFB Borden, Ontario, Canada. Can. J. Contam. Hydrol.
18, 161^179.
[12] Tebbe, C.C. and Reber, H.H. (1991) Degradation of [14 C]phosphinothricin (glufosinate) in soil under laboratory conditions ^ e¡ects of
concentration and soil amendments on 14 CO2 production. Biol. Fertil. Soils 11, 62^67.
[13] Gallina, M.A. and Stephenson, G.R. (1992) Dissipation of
[14 C]glufosinate ammonium in two Ontario soils. J. Agric. Food
Chem. 40, 165^168.
[14] Ismail, B.S. and Ahmad, A.R. (1994) Attenuation of the herbicidal
activities of glufosinate-ammonium and Imazapyr in two soils. Agric.
Ecosyst. Environ. 47, 279^285.
[15] Faber, M.J., Stephenson, G.R. and Thompson, D.G. (1997) Persistence and leachability of glufosinate-ammonium in a northern Ontario terrestrial environment. J. Agric. Food Chem. 45, 3672^3676.
[16] Ahmad, I. and Malloch, D. (1995) Interaction of soil micro£ora with
the bioherbicide phosphinothricin. Agric. Ecosyst. Environ. 54, 165^
174.
[17] van Leeuwen, W., Ruttink, T., Borst Vrenssen, A.W.M., van der
Plas, L.H.W. and van der Krol, A.R. (2001) Characterization of
FEMSEC 1336 13-5-02
A. Schmalenberger, C.C. Tebbe / FEMS Microbiology Ecology 40 (2002) 29^37
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
position-induced spatial and temporal regulation of transgene promoter activity in plants. J. Exp. Bot. 52, 949^959.
Lynch, J.M. (1990) The Rhizosphere. Wiley, New York.
Maloney, P.E., Vanbruggen, A.H.C. and Hu, S. (1997) Bacterial
community structure in relation to the carbon environments in lettuce
and tomato rhizospheres and in bulk soil. Microb. Ecol. 34, 109^
117.
Miller, H.J., Henken, G. and van Veen, J.A. (1989) Variation in the
composition of bacterial populations in the rhizosphere of maize,
wheat, and grass cultivars. Can. J. Microbiol. 35, 656^660.
Rengel, Z., Ross, G. and Hirsch, P. (1998) Plant genotype micronutrient status in£uence colonization of wheat roots by soil bacteria.
J. Plant Nutr. 21, 99^113.
Siciliano, S.D., Theoret, C.M., de Freitas, J.R., Hucl, P.J. and Germida, J.J. (1998) Di¡erences in the microbial communities associated
with roots of di¡erent cultivars of canola and wheat. Can. J. Microbiol. 44, 844^851.
Siciliano, S.D. and Germida, J.J. (1999) Taxonomic diversity of bacteria associated with the roots of ¢eld-grown transgenic Brassica napus cv. Quest, compared to the non-transgenic B. napus cv. Excel and
B. rapa cv. Parkland. FEMS Microbiol. Ecol. 29, 263^272.
Lukow, T., Dun¢eld, P.F. and Liesack, W. (2000) Use of the
T-RFLP technique to assess spatial and temporal changes in the
bacterial community structure within an agricultural soil planted
with transgenic and non-transgenic potato plants. FEMS Microbiol.
Ecol. 32, 241^247.
Lottmann, J. and Berg, G. (2001) Phenotypic and genotypic characterization of antagonistic bacteria associated with roots of transgenic
and non-transgenic potato plants. Microbiol. Res. 156, 75^82.
Torsvik, V., Salte, K., Sorheim, R. and Goksoyr, J. (1990) Comparison of phenotypic diversity and DNA heterogeneity in a population
of soil bacteria. Appl. Environ. Microbiol. 56, 776^781.
Chelius, M.K. and Triplett, E.W. (2001) The diversity of archaea and
bacteria in association with the roots of Zea mays. Microb. Ecol. 41,
252^263.
Stahl, D.A. (1997) Molecular approaches for the measurement of
density, diversity, and phylogeny. In: Manual of Environmental Microbiology (Hurst, C.J., Knudsen, G.R., McInerney, M.J., Stetzenbach, L.D. and Walter, M.V., Eds.), pp. 102^114. ASM Press, Washington, DC.
Schwieger, F. and Tebbe, C.C. (1998) A new approach to utilize
PCR-single-strand-conformation polymorphism for 16S rRNA
gene-based microbial community analysis. Appl. Environ. Microbiol.
64, 4870^4876.
Wehrmann, A., Van Vliet, A., Opsomer, C., Botterman, J. and
Schulz, A. (1996) The similarities of bar and pat gene products
make them equally applicable for plant engineers. Nat. Biotechnol.
14, 1274^1278.
Sambrook, J.E., Fritsch, E.F. and Maniatis, T. (1989) Molecular
Cloning : A Laboratory Manual. Cold Spring Harbor Press, Cold
Spring Harbor, NY.
Neefs, J.-M., Van der Peer, Y., De Rejk, P., Chapelle, S. and De
Wachter, R. (1993) Compilation of small ribosomal subunit RNA
structures. Nucleic Acids Res. 21, 3025^3049.
Nu«bel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A.,
Amann, R.I., Ludwig, W. and Backhaus, H. (1996) Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus polymyxa
detected by temperature gradient gel electrophoresis. J. Bacteriol.
178, 5636^5643.
Schmalenberger, A., Schwieger, F. and Tebbe, C.C. (2001) E¡ect of
primers hybridizing to di¡erent evolutionarily conserved regions of
the small-subunit rRNA gene in PCR-based microbial community
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
37
analyses and genetic pro¢ling. Appl. Environ. Microbiol. 67, 3557^
3563.
Tebbe, C.C., Schmalenberger, A., Peters, S. and Schwieger, F. (2001)
Single-strand conformation polymorphism (SSCP) for microbial
community analysis. In: Environmental Molecular Microbiology :
Protocols and Applications (Rochelle, P.A., Ed.), pp. 161^175. Horizon Scienti¢c Press, Wymondham.
Bassam, B.J., Caetano-Anolles, G. and Gressho¡, P.M. (1991) Fast
and sensitive silver staining of DNA in polyacrylamide gels. Anal.
Biochem. 196, 80^83.
Peters, S., Koschinsky, S., Schwieger, F. and Tebbe, C.C. (2000)
Succession of microbial communities during hot composting as detected by PCR-single-strand-conformation polymorphism-based genetic pro¢les of small-subunit rRNA genes. Appl. Environ. Microbiol. 66, 930^936.
Ludwig, W., Strunk, O., Klugbauer, S., Klugbauer, N., Weizenegger,
M., Neumeier, J., Bachleitner, M. and Schleifer, K.H. (1998) Bacterial phylogeny based on comparative sequence analysis. Electrophoresis 19, 554^568.
Heuer, H., Wieland, G., Scho«nfeld, J., Scho«nwa«lder, A., Gomes,
N.C.M. and Smalla, K. (2001) Bacterial community pro¢ling using
DGGE or TGGE analysis. In: Environmental Molecular Microbiology: Protocols and Applications (Rochelle, P.A., Ed.), pp. 177^190.
Horizon Scienti¢c Press, Wymondham.
Germida, J.J., Siciliano, S.D., de Freitas, J.R. and Seib, A.M. (1998)
Diversity of root-associated bacteria associated with ¢eld grown canola (Brassica napus L.) and wheat (Triticum aestivum L.). FEMS
Microbiol. Ecol. 26, 43^50.
Miethling, R., Wieland, G., Backhaus, H. and Tebbe, C.C. (2000)
Variation of microbial rhizosphere communities in response to crop
species, soil origin, and inoculation with Sinorhizobium meliloti L33.
Microb. Ecol. 40, 43^56.
Gomes, N.C.M., Heuer, H., Scho«nfeld, J., Costa, R., Mendonca
Hagler, L. and 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, 167^180.
von der Weid, I., Paiva, E., Nobrega, A., van Elsas, J.D. and Seldin,
L. (2000) Diversity of Paenibacillus polymyxa strains isolated from
the rhizosphere of maize planted in Cerrado soil. Res. Microbiol. 151,
369^381.
Engelen, B., Meinken, K., von Wintzintgerode, F., Heuer, H., Malkomes, H.P. and Backhaus, H. (1998) Monitoring impact of a pesticide treatment on bacterial soil communities by metabolic and genetic
¢ngerprinting in addition to conventional testing procedures. Appl.
Environ. Microbiol. 64, 2814^2821.
El Fantroussi, S., Verschuere, L., Verstraete, W. and Top, E.M.
(1999) E¡ect of phenylurea herbicides on soil microbial communities
estimated by analysis of 16S rRNA gene ¢ngerprints and communitylevel physiological pro¢les. Appl. Environ. Microbiol. 65, 982^988.
Dalmastri, C., Chiarini, L., Cantale, C., Bevivino, A. and Tabacchioni, S. (1999) Soil type and maize cultivar a¡ect the genetic diversity of
maize root-associated Burkholderia cepacia populations. Microb.
Ecol. 38, 273^284.
Kuske, C.R., Banton, K.L., Adorada, D.L., Stark, P.C., Hill, K.K.
and Jackson, P.J. (1998) Small-scale DNA sample preparation method for ¢eld PCR detection of microbial cells and spores in soil. Appl.
Environ. Microbiol. 64, 2463^2472.
Seldin, L., Rosado, A.S., da Cruz, D.W., Nobrega, A., van Elsas,
J.D. and Paiva, E. (1998) Comparison of Paenibacillus azoto¢xans
strains isolated from rhizoplane, rhizosphere, and non-root-associated soil from maize planted in two di¡erent Brazilian soils. Appl.
Environ. Microbiol. 64, 3860^3868.
FEMSEC 1336 13-5-02