FEMS Microbiology Ecology 28 (1999) 261^272
Monitoring ¢eld releases of genetically modi¢ed sugar beets for
persistence of transgenic plant DNA and horizontal gene transfer
Frank Gebhard, Kornelia Smalla *
Biologische Bundesanstalt fuër Land- und Forstwirtschaft, Institut fuër P£anzenvirologie, Mikrobiologie und biologische Sicherheit,
Messeweg 11^12, D-38104 Braunschweig, Germany
Received 24 July 1998; received in revised form 20 November 1998; accepted 21 November 1998
Abstract
Field releases of transgenic rizomania-resistant sugar beet (Beta vulgaris) plants were accompanied by a study of the
persistence of DNA from transgenic sugar beet litter in soil and of horizontal gene transfer of plant DNA to bacteria. The
transgenic sugar beets contained the marker genes nptII and bar under the control of the bidirectional TR1/2 promoter
conferring kanamycin (Km) and glufosinate ammonium resistance to the plant. Primer systems targeting the construct allowed
the specific and sensitive detection of the transgenic DNA in soil. Soil samples were analyzed by cultivation of bacteria on
nonselective and Km-selective media to determine the proportion of Km-resistant bacteria and to monitor the culturable
fraction for incorporation of transgenic plant DNA. To detect the presence of transgenic DNA independently from cultivation,
total soil DNA was extracted and amplified by PCR with three different primer sets specific for the transgenic DNA. Longterm persistence of transgenic DNA could be shown under field conditions (up to 2 years) and also in soil microcosms with
introduced transgenic plant DNA. No construct-specific sequences were detected by dot blot hybridizations of bacterial
isolates. The experimental limitations of detecting horizontal gene transfer from plants to bacteria under field conditions are
discussed. z 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
Keywords : Plant DNA persistence; Horizontal gene transfer
1. Introduction
Several thousand ¢eld releases of genetically modi¢ed plants (GMP) have been performed during the
last decade. At present, the majority of the GMP
tested in ¢eld releases or commercialized contain selectable markers such as bacterial antibiotic resistance genes. The hypothetical acquisition of plantharbored antibiotic resistance genes by bacterial
* Corresponding author. Tel.: +49 (531) 2993814;
Fax: +49 (531) 2993013; E-mail: [email protected]
communities from transgenic plants is often discussed as an undesired e¡ect of large-scale applications of GMP, due to well-known problems caused
by antibiotic-resistant bacteria [1^5]. However, experimental evidence that horizontal gene transfer of
genetic material from plants to bacteria can occur at
all has been lacking [6^9]. Recently, transformation
of Acinetobacter sp. BD413 with DNA from various
transgenic plants carrying the nptII gene was demonstrated [10,11], with detection of bacterial transformants based on recombinational repair of a deleted
nptII gene in the Acinetobacter sp. strains. Restora-
0168-6496 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 6 4 9 6 ( 9 8 ) 0 0 1 1 5 - 9
FEMSEC 999 8-3-99
262
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
tion of a 317-bp deletion of the nptII gene, resulting
in kanamycin (Km)-resistant transformants, was observed with plant DNA and plant homogenates at
frequencies of 5U1039 and 10310 , respectively [10].
However, it is unknown how relevant these ¢ndings
are with respect to horizontal gene transfer from
plants to bacteria under environmental conditions.
DNA can be released into soil from decaying plant
material [12]. If such DNA is rapidly degraded by
extracellular DNases it is unlikely to be taken up by
competent bacteria. Several groups have shown the
persistence of free DNA adsorbed to puri¢ed soil
components and to clay-amended soils [13^24].
DNA adsorbed to mineral surfaces remains to
some degree protected from the attack of nucleases
and thus maintains its transforming capacity [16,18^
21]. However, the soil systems used have often been
rather arti¢cial [14^19,25]. Long-term persistence of
even a small percentage of the released plant DNA is
assumed to enhance the likelihood of transformation
processes. Although competence for transformation
has been demonstrated for a wide range of bacteria
from marine or terrestrial habitats under laboratory
conditions [22], little is known of the frequency of
competent bacteria in environmental samples and the
environmental stimuli triggering the expression of
bacterial competence under natural conditions. Recently, natural transformation of Acinetobacter sp.
(formerly Acinetobacter calcoaceticus) and Pseudomonas stutzeri by bacterial DNA in nonsterile soils
has been shown [20,26]. However, in both studies
competent bacteria were introduced into the soil.
Other studies indicate that several environmental
factors which stimulate bacterial growth might enhance the probability of soil- or plant-associated
bacteria becoming competent [27,28]. Nielsen et al.
[27] were the ¢rst to show in situ transformation of
Acinetobacter sp. in soil by stimulating growth of
Acinetobacter sp. BD413.
The main foci of this study were to follow the
persistence of transgenic sugar beet DNA in soil
and assess, under ¢eld conditions, the appearance
of bacteria carrying parts of the transgenic construct,
in particular the nptII gene. First, tools to detect
speci¢cally the transgenic DNA had to be developed.
The fate of the transgenic plant DNA in soil samples
taken from the release site was then followed with
cultivation-dependent and cultivation-independent
methods by PCR and DNA probes. In addition, microcosm studies were performed to determine the
persistence of free plant DNA in soil from the ¢eld
release site.
2. Materials and methods
2.1. Plant material, bacterial strains and plasmids
Transgenic and nontransgenic sugar beet plants
were grown by PLANTA Angewandte P£anzengenetik und Biotechnologie GmbH (Einbeck, Germany)
in Oberviehhausen (Bavaria, Germany). Plant material was obtained from ¢eld-grown sugar beet plants
provided by PLANTA. Escherichia coli strains carrying plasmids pGSFR160 (bar-TR2/TR1-nptII) or
pGSBNYC1 (35S/BNYVV cp) were provided by
Plant Genetic Systems (Ghent, Belgium) (for plasmid
maps see [9]). These plasmids contain the constructspeci¢c sequences of the released transgenic sugar
beets. Strains E. coli pSUP1021 and E. coli pESC1
containing plasmid-localized nptII were obtained
from J.D. van Elsas (IPO-DLO, Wageningen, The
Netherlands).
2.2. Soil
The soil at the ¢eld site in Oberviehhausen, which
was also used for the microcosm experiments, was a
parabrown earth. The soil type was determined by
H.-P. Malkomes (Biologische Bundesanstalt fuër
Land- und Forstwirtschaft, Braunschweig) to be silt
loam: 15.1% sand, 65.3% silt, and 19.7% clay; 1.6%
organic carbon, pH 6.3 (Oberviehhausen silt loam,
OSL).
2.3. Primers, probes, PCR and hybridization
conditions
Three primer systems targeting three di¡erent parts
of the construct (see Fig. 1) were designed with the
Oligo 4.0 program (National Biosciences, Plymouth,
UK) based on sequences kindly provided by PLANTA. Primer sequences, the size of the PCR products
and melting temperatures are given in Table 1. PCR
mixes contained 0.1 Wmol of both forward and reverse primer, 3.75 mM MgCl2 , 0.2 mM deoxynucleo-
FEMSEC 999 8-3-99
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
263
Fig. 1. Genetic construct in the transgenic sugar beet plants. I^III indicate the primer sets used for PCR ; structural genes: bar: phosphinotricin acetyltransferase gene (Streptomyces hygroscopicus); nptII: neomycin/kanamycin phosphotransferase (Tn5) ; BNYVV cp (coat protein gene of beet necrotic yellow vein virus); promoters: TR1/TR2 (Agrobacterium tumefaciens) ; 35S CaMV (cauli£ower mosaic virus); terminator : 3Pocs and 3Pnos (Agrobacterium tumefaciens).
side triphosphates, 1USto¡el bu¡er, 2.5 U Taq polymerase Sto¡el fragment (Perkin Elmer). To reduce
inhibitory e¡ects of co-extracted soil compounds,
such as humic acids, 1.25 Wl of 2 mg ml31 bovine
serum albumin (DNase-free, Pharmacia Biotech,
Sweden) was added. Ampli¢cation involved a 7-min
denaturation step at 92³C and 35 cycles consisting of
1 min denaturation at 92³C, 1 min primer annealing
(see Table 1), and 1.5 min at 72³C primer extension
followed by a ¢nal 10-min extension step at 72³C.
PCR products (10 Wl) were analyzed on 0.7% agarose
gels with Tris borate EDTA bu¡er (TBE) [29], Southern-blotted [29] and hybridized using digoxigenin-labelled probes. The probes were generated by PCR
using the conditions described above except that
the deoxynucleoside triphosphate mixture contained
0.13 mM dTTP and 0.07 mM digoxigenin-labelled
dUTP (Boehringer Mannheim, Germany).
2.4. Test of primer speci¢city
To test the speci¢city of primer system I (TR2/
nptII) for the construct, PCR was performed with
templates containing the natural nptII gene (E. coli
pSUP1021; E. coli pESC1) with and without transgenic sugar beet DNA added. E. coli pGSFR160 was
used as the positive control. In parallel, templates
were ampli¢ed with primers annealing inside the
nptII gene using published PCR conditions [30].
2.5. Determination of the limit of detection for primer
systems I^III
Samples of 5 g OSL soil were not inoculated (1),
or inoculated with 102 (2), 103 (3), 104 (3), 105 (2), or
106 (1) cells (number of replicates given in parentheses) of E. coli pGSFR160 (primer systems I, II) or
Table 1
PCR primer systems
Region
Primer sequence
Size (bp)
tannealing (³C)
(I) TR2/nptII
5P-GCATTCCGTTCTTGCTGTA-3P
5P-GATGTTTCGCTTGGTGGTC-3P
5P-CTTCAGCAGGTGGGTGTAGA-3P
5P-TCGTTTATTTCGGCGTGTAG-3P
5P-TGACGCACAATCCCACTATC-3P
5P-CAGCAAGACTAACGGTAAGC-3P
769
55
650
57
616
53
(II) bar/TR1
(III) 35S/BNYVV cp
FEMSEC 999 8-3-99
264
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
E. coli pGSBNYC1 (primer system III) per gram of
soil, incubated at room temperature for 30 min and
frozen at 370³C afterwards. Total community DNA
was extracted and puri¢ed as described in Section
2.11. DNA was used as a template for PCR ampli¢cation with all three primer systems. PCR products
were analyzed by agarose gel electrophoresis with
subsequent Southern blot hybridization.
2.6. Field design and soil sampling
In October 1993, transgenic sugar beet plants were
harvested from the ¢eld, shredded and introduced
into a disposal site (separate area of around 100
m2 ) coded DS93. In the spring and fall of 1994
and 1995, soil samples were taken from DS93 with
a Puërckhauer sampler (Baumann Saatzuchtbedarf,
Waldenburg, Germany). In addition, samples were
taken from another disposal site where sugar beet
litter had been incorporated into soil in the fall of
1994 (coded DS94, approx. 135 m2 ) and from plots
planted with transgenic sugar beet plants in the
spring of 1994. To achieve good representation of
the soil samples, these were taken randomly over
the whole site to a depth of 30 cm (diameter of
sampler: 1.8 cm) and composited. Each composite
sample thus consisted of 28 subsamples which were
combined and homogenized by sieving (pore size
4 mm) before further treatment. The soil was dried
if needed before sieving.
2.7. Cultivation of soil bacteria
OSL soil (1 g) was resuspended in 9 ml sterile
saline (0.85% w/v NaCl) by blending with a Stomacher blender (Seward Medical, London, UK) for
1 min at maximum speed. The suspension was plated
in a 10-fold dilution series onto plate count agar
(PCA; Merck, Darmstadt) containing 0, 10 Wg
ml31 , or 100 Wg ml31 Km and 100 Wg ml31 cycloheximide. From each composite sample two subsamples were processed and plated in duplicate.
Plates were incubated for 2^4 days at 28³C before
determining the number of colony forming units
(cfu). Colonies with di¡erent morphology were isolated from Km-containing plates of dilutions 1034
and 1035 (10 Wg ml31 or 100 Wg ml31 Km) after
plating and analyzed by dot blot hybridization with
all three probes for the presence of construct-speci¢c
DNA.
2.8. Dot blot hybridizations
Cells of single Kmr colonies from both replicates
were cultivated overnight in 1 ml Luria-Bertani
broth with 10 Wg ml31 Km. To a 150-Wl aliquot, 10
Wl lysis solution (6 mg ml31 lysozyme) and 0.2 Wl
RNase A (10 mg ml31 ) were added. After incubation
for 2 h at 37³C, 50 Wl of the lysate was applied to a
nylon Hybond-N membrane (Amersham, UK) using
a dot blot apparatus. This was done only for colonies isolated from ¢eld samples. The membranes
were treated as described for colony blots [29]. The
blots were hybridized with a mixture of probes speci¢c for all three parts of the construct under high
stringency conditions. Hybridization-positive cells
were further analyzed by PCR with primer systems
I^III.
2.9. Extraction of the bacterial fraction
The bacterial fraction was extracted from soil
based on established protocols [31,32]. Soil portions
of 5 g were treated sequentially in a Stomacher blender (1 min at maximum speed) with 45 ml sterile
saline, 1 g Chelex (Bio-Rad) in 10 ml NDP solution
(NDP: 0.1% sodium deoxycholate (Sigma, St. Louis,
MO), 2.5% polyethylene glycol 6000 (Merck, Darmstadt) in bidistilled water), 10 ml NDP solution, 10
ml Tris-HCl solution (50 mM Tris-HCl, pH 7.4), and
30 ml bidistilled water. Chelex addition was followed
by incubation on ice for 15 min with a 1-min Stomacher treatment after each 5 min. Between additions,
the suspended cells were separated from the soil particles by low-speed centrifugation (500Ug) and
stored on ice. The remaining soil pellet was re-extracted. The individual supernatants of each step
were pooled and collected by centrifugation
(5000Ug for 10 min). The resulting pellet was resuspended in 5 ml saline and underlaid with 10 ml Percoll (Pharmacia Biotech, Sweden) containing 0.25 M
sucrose [33]. Centrifugation for 15 min at 25 000Ug
produced a density gradient from 1 to 1.3 g ml31 .
After centrifugation, the interphase and the liquid
phase above it were removed, mixed with 30 ml saline and centrifuged to obtain the bacterial fraction.
FEMSEC 999 8-3-99
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
To remove residual Percoll the pellets were washed
twice with sterile saline. The pellets were resuspended
in 100 Wl water and treated with 100 Wl DNase I (1
mg ml31 ; overnight). DNase I activity was stopped
by the addition of 1 vol 0.5 M EDTA pH 8.0.
2.10. Community DNA extraction
DNA extractions from soil samples (5 g), from the
bacterial pellet obtained from soil (Section 2.9), or
from the bacterial pellet of culturable bacteria (Section 2.7) were carried out according to Smalla and
van Elsas [34]. To ensure lysis of dwarf cells, endospores or bacteria inside cavities, samples were
treated with a bead beater (1 g glass beads (diameter
0.11 mm), 60 s, 4000 rpm; Braun Melsungen, Germany). DNA from soil was further puri¢ed as described by van Elsas and Smalla [35] with the Wizard
DNA puri¢cation kit (Promega Corp., Madison, WI,
USA) and resulting DNA (size s 15 kb, data not
shown) analyzed by PCR and Southern blot hybridization.
2.11. Microcosm studies
DNA extracted from 25 g transgenic sugar beet
leaves according to [36] was introduced into 130 g
nonsterile OSL. After intensive mixing with a sterile
spatula, portions of 6 g soil were partitioned into 50ml falcon tubes (Becton Dickinson, Franklin Lakes,
NJ, USA) and incubated at 20³C in the dark. After
1 and 5 h, 1, 2, 4, 7, 10, 15, 18, 23, and 30 days and
3 and 6 months, two soil microcosm replicates were
analyzed by plating on PCA (1 g) and total DNA
extraction (5 g) with subsequent PCR analysis. For
plating on PCA, 1 g soil was resuspended in 9 ml
saline by blending with a Stomacher blender for
1 min at maximum speed. Tenfold dilutions of this
suspension were spread on PCA containing 100 Wg
ml31 cycloheximide. After 4 days incubation, the
bacterial lawns of dilutions 1032 and 1033 were resuspended in 4 ml saline and pelleted by centrifugation. The resulting bacterial pellets were treated with
DNase I (1 mg ml31 ; overnight at 37³C) and the
reaction was stopped by addition of EDTA. Cells
of 100 Wl resuspension were boiled for 10 min, and
1 Wl was used as template for PCR ampli¢cation with
primers I^III.
265
3. Results
3.1. Speci¢city and sensitivity of primer systems I^III
Three sets of primers were used for speci¢c ampli¢cation of the transgenic DNA by PCR (Fig. 1),
each consisting of one primer annealing to the foreign promoter sequence (TR1/2; 35S) and one targeting a structural gene (nptII, bar, BNYVVcp). To
determine the limit of detection for the construct in
soil, di¡erent numbers of E. coli cells carrying plasmid pGSFR160 or pGSBNYC1 were introduced
into nonsterile OSL soil. The limit of detection as
determined by PCR ampli¢cation followed by
Southern blot hybridization was around 5U102 cells
g31 dry soil corresponding to approx. 103 target sequences for primer pairs I and II. The ampli¢cation
with primer set III was slightly less e¤cient (around
104 templates g31 soil). An additional hybridizing
band was consistently observed for all primers. Since
this product gave a strong hybridization signal it can
be due either to self-priming structures of the speci¢c
PCR product, to single-stranded PCR products or to
an additional primer annealing site. PCR products
obtained with primer systems I^III were always con¢rmed by hybridization under high stringency conditions.
As expected, the nptII gene located on plasmids
pSUP1021 and pESC1 was not PCR-ampli¢able
with primer set I (data not shown) due to the absence of the annealing site for the primer complementary to the TR2 promoter. PCR products of
the expected size were observed only in the presence
of transgenic sugar beet DNA. However, both templates were ampli¢able with primers designed to amplify an internal part of the nptII gene [30]. Thus
primer set I should allow the speci¢c detection of
the nptII gene of transgenic plant origin even in the
presence of naturally occurring nptII sequences.
3.2. Monitoring the fate of transgenic sugar beet
DNA under ¢eld conditions
The persistence of transgenic sugar beet DNA and
putative horizontal gene transfer of parts of the
transgenic DNA to soil bacteria was tracked by cultivation-based and cultivation-independent analysis
of soil samples. Before starting the ¢eld release
FEMSEC 999 8-3-99
266
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
study, tests were performed to determine the variance between composite samples and subsamples
for cfu numbers of Kmr bacteria [37]. This information was required to estimate the number of samples
needed to detect di¡erences in the number of Kmr
bacteria greater than twice the standard deviation
between ¢eld sites planted with transgenic and nontransgenic sugar beet plants using the computer program CADEMO [38]. The variance coe¤cient was
17.5% between subsamples and 46.1% between composite samples.
3.2.1. Cultivation-based analysis of soil samples
Soil samples taken from the ¢eld release site were
plated onto PCA with and without Km to determine
the proportion of Kmr bacteria and to determine, by
DNA hybridization, whether the Km resistance phenotype was nptII encoded. Low (10 Wg ml31 ) and
high levels (100 Wg ml31 ) of Km were chosen because
the level of expression of the nptII gene driven by the
TR1 promoter in di¡erent soil bacteria was not
known. Recently, Gebhard and Smalla [10] showed
that the nptII gene under the control of the TR1
promoter did not confer high level Km resistance
to E. coli and Acinetobacter sp. The proportion of
Kmr bacteria in the total bacterial population ranged
from 0.6% to 5.2% for high level and 4% to 30% for
low level Kmr bacteria. Introduction of sugar beet
litter into the soil led to an increase in both the Kmr
and the total bacterial population (Fig. 2). However,
resistance quotients did not increase signi¢cantly.
A total of 4000 isolates picked from Km-selective
agar (low and high level) was screened by dot blot
hybridizations for the presence of construct-speci¢c
sequences. A few isolates giving weak hybridization
signals with a probe mix (I^III) were checked by
PCR using primer systems I^III. None of the isolates
gave PCR products indicating that the hybridization
signal might be due to a reaction of the probes with
exopolysaccharides.
Fig. 2. Plate counts (cfu g31 soil) of soil samples taken from DS93 and DS94 after di¡erent times following the incorporation into soil
determined on PCA, PCA (10 Wg ml31 Km) and PCA (100 Wg ml31 Km) after 2^4 days incubation at 28³C.
FEMSEC 999 8-3-99
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
Fig. 3. Southern blot hybridization of PCR products ampli¢ed
with primer system bar/TR1 from DNA extracted directly from
soil samples taken from the disposal site 93 in spring 1994. A:
Lanes 1^8: PCR products ampli¢ed from DNA extracted directly
from composite soil samples 1^8 taken from the disposal site approx. 6 months after sugar beet litter was plugged into soil; 9:
negative control; 10: positive control; 11: dig ladder. B: Lanes
1^8: PCR products ampli¢ed from DNA extracted directly from
composite soil taken from the ¢eld plots with young sugar beet
plants ; 9: positive control; 10: negative control; 11: dig ladder.
3.2.2. Detection of the transgenic DNA in soil DNA
3.2.2.1. Analysis of directly extracted community
DNA. PCR analysis of total extracted DNA allows
the detection of construct-speci¢c DNA in soil. An
average of 5^20 Wg DNA ranging in size from 3 to 30
kb was obtained from soil samples and two DNA
puri¢cation steps using the Wizard DNA puri¢cation
kit were su¤cient to obtain PCR-ampli¢able DNA.
To exclude false positives due to contamination during the extraction and puri¢cation procedure, a control soil sample was included. All DNA samples were
267
subjected to PCR using the three di¡erent primer
sets. PCR products were analyzed on agarose gels
and by Southern blot hybridizations. Empty lanes
only with primers visible were obtained if the PCR
ampli¢cation of the soil DNA was inhibited. Weak
randomly ampli¢ed PCR products were observed if
the PCR was not inhibited. Fig. 3 shows a Southern
blot hybridization of PCR products ampli¢ed with
primer set II from soil DNA obtained from composite soil samples taken from the disposal site 1993
(Fig. 3A) or from the plots with young transgenic
sugar beets (Fig. 3B). After 6 months, construct-speci¢c DNA was detectable in the majority of the composite soil samples with all three primer sets. On
later sampling occasions (after 18 and 24 months),
PCR ampli¢cation products were still observed but
only with primer set I. Table 2 summarizes the results of PCR analyses of soil samples taken from
DS93 at di¡erent times after incorporation of the
transgenic sugar beet litter. Persistence of transgenic
DNA was also found for DS94. However, it was not
possible to obtain information on whether the transgenic DNA persisted as free DNA bound to soil
particles, in decaying or rotting plant material, or
in transformed bacteria.
3.2.2.2. Analysis of DNA extracted from the bacterial
fraction. To track the presence of construct-speci¢c
sequences in bacterial cells independently from their
culturability a protocol to extract the bacterial fraction from soil particles was developed, based on established cell extraction procedures [31,32]. Several
Stomacher blending steps using di¡erent bu¡ers
and the ion exchange resin Chelex-100 (Bio-Rad,
Hercules, CA, USA) were applied to homogenize
e¤ciently the soil particles and to dislodge surfaceattached bacterial cells. Low speed centrifugation
was performed to separate the bacterial cells from
the soil matrix. Due to the high clay content of the
Table 2
PCR-based detection in total community DNA extracted from soil samples taken from DS93 at di¡erent time points after introduction of
sugar beet litter into soil
Primer system
6 months
number of PCR-positives
(number of samples)
18 months
number of PCR-positives
(number of samples)
24 months
number of PCR-positives
(number of samples)
(I) TR2/nptII
(II) bar/TR1
(III) 35S/BNYVV cp
6 (8)
7 (8)
7 (8)
7 (7)
not determined
0 (7)
1 (4)
0 (4)
0 (4)
FEMSEC 999 8-3-99
268
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
OSL soil samples and a similar sedimentation behavior of small clay particles and bacteria, the bacterial
fraction recovered still contained a considerable
amount of clay, necessitating additional puri¢cation
of the bacteria using a Percoll-sucrose gradient. The
amount of DNA recovered was low compared to
directly extracted DNA (less than 20%). To determine whether the DNA recovered was mainly of
bacterial origin, DNA of plasmid pGSFR160 (0.2
or 2 Wg g31 of soil) or 50 mg of homogenized leaf
material of transgenic sugar beets were introduced
into soil prior to the extraction of the bacterial fraction. DNase I treatment of the bacterial pellet drastically reduced the intensity of the hybridization signal of PCR products obtained after ampli¢cation of
DNA from soils with 0.2 Wg pGSFR160 DNA g31
soil, whereas for soils to which 2 Wg g31 was added
strong hybridization signals were observed even after
DNase treatment (data not shown). A faint PCR
signal was also observed after DNase treatment for
one of the soil samples with leaf homogenates added.
The protocol was used to extract the bacterial fraction from soil samples taken from DS93. Two out of
seven samples taken after 18 months yielded PCR
products with primer system I (data not shown),
while all other samples taken after 18 and 24 months
from DS93 were negative with all primer systems.
3.3. Microcosm experiments
Microcosm experiments with sugar beet DNA
added to OSL soil were performed to provide data
on the persistence of free plant DNA. The fate of the
transgenic plant DNA was monitored by PCR detection of the construct in total DNA extracted directly
from soil after di¡erent incubation times. PCR products were analyzed by agarose gel electrophoresis
and Southern blot analysis. After introduction of
Fig. 4. Soil microcosm studies. Persistence of free transgenic sugar beet DNA in soil. Southern blot hybridizations of PCR products ampli¢ed with primer systems TR2/nptII (A), bar/TR1 (B), and 35S/BNYVV cp (C) from soil at di¡erent time points before (lane 1) and
after introducing the sugar beet DNA into soil : lane 2: 1 h; 3: 5 h; 4: 1 day ; 5: 2 days; 6: 4 days; 7: 7 days; 8: 10 days; 9: 15 days;
10: 18 days; 11: 23 days; 12: 30 days; 13: negative control; 14: positive control.
FEMSEC 999 8-3-99
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
269
Fig. 5. Soil microcosms : detection of construct-speci¢c sequences in DNA extracted from resuspended bacterial lawn grown on dilutions
1032 and 1033 on PCA plates by PCR using primer systems bar/TR1 (A), TR2/nptII (B), and 35S/BNYVV cp (C). Lane 1: before introducing the sugar beet DNA into soil ; lane 2: after 1 day; 3: 2 days; 4: 4 days; 5: 7 days; 6: 10 days; 7: 15 days; 8: 18 days; 9: 23
days; 10 : 30 days; 11 : negative control; 12: positive control.
sugar beet DNA the intensity of the hybridization
signal decreased during the ¢rst days (Fig. 4) and
subsequently increased, in particular for ampli¢cations with primer sets I and II. PCR ampli¢cation
products from directly extracted soil DNA were
obtained with primer systems I, II and III up to
3 months, until the end of the experiment (6 months),
and up to 4 weeks, respectively. In parallel, the soil
samples were plated onto PCA. The bacterial lawn
grown after 4 days incubation (1032 , 1033 dilution)
was resuspended, treated with DNase I and the
DNA released from the bacterial pellet by boiling
and freezing. PCR ampli¢cation with all three primer
sets resulted in several positive signals (Fig. 5), which
might indicate uptake of transgenic DNA by competent bacteria. Because the isolates carrying the
construct-speci¢c DNA sequences were not accessible, an interpretation of the signals remains inconclusive.
4. Discussion
This study, accompanying the ¢eld release of
transgenic sugar beets, has shown that the DNA of
transgenic sugar beet plants was detectable for several months in soil under ¢eld conditions. Since it
was supposed that long-term persistence of transgenic plant DNA is partly attributable to a protection of
the DNA inside cells of decaying plant residues, microcosm studies with free transgenic sugar beet DNA
added to OSL soil have been performed. Constructspeci¢c sequences were detected in microcosm soil
samples for up to 6 months. However, the construct
was detectable with primer I under ¢eld conditions in
one sample for up to 2 years. Therefore, it is assumed that the construct-speci¢c sequences detected
in soil samples from the disposal site originated
partly from DNA localized in sugar beet litter and
from free DNA. Similar observations were made by
FEMSEC 999 8-3-99
270
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
Widmer et al. [24] who followed the persistence of
transgenic DNA in nonsterile soil microcosms after
introducing plasmid DNA or leaf tissue of transgenic
tobacco. Persistence of transgenic plant DNA in ¢eld
soil was also reported by Paget et al. [23] as unpublished results. We have no satisfactory explanation
for the observation that in soil microcosms the intensity of hybridization signals of PCR products ¢rst
decreased and then increased. In several cases PCR
ampli¢cations done on the same DNA extracts from
soils (Table 2; Fig. 4) or the cultured fraction (Fig.
5) were observed only with one primer set. This
might be due to uptake of a respective fragment by
competent bacteria, or di¡erent parts of the transgenic construct may be di¡erently accessible due to
secondary structures or interactions with the plant
genome.
Di¡erent abiotic and biotic factors seem to a¡ect
the persistence of free DNA in soil making general
predictions on the persistence of plant DNA in soils
di¤cult. The content and type of clay minerals in£uence the extent to which free DNA is adsorbed to
mineral surfaces and thus is protected from degradation by nucleases [20,39,40]. In addition, high microbial activity accompanied by an enhanced presence
of bacterial DNase a¡ects the persistence of free
DNA in soil [39]. Ogram et al. [40] reported that
the largest amounts of adsorbed DNA were present
in soils with the highest clay content.
A prerequisite for transformation processes to occur in soil is the presence of free DNA and of bacteria developing competence in the close vicinity [41].
In aquatic systems, transformation processes depending on cell-to-cell contact and involving dead
bacterial cells as donors have been described [42].
Based on the observations by Gebhard and Smalla
[10] and Bertolla et al. [28] it can be hypothesized
that bacteria growing on rotting plant materials
which allow an intimate contact between decaying
plant cells releasing transgenic DNA and bacterial
cells could be another scenario for horizontal gene
transfer of transgenic plant DNA by transformation
to plant-associated bacteria. In this study, di¡erent
strategies were followed to detect horizontal gene
transfer from transgenic sugar beets to soil bacteria
under ¢eld and microcosm conditions. Plating of soil
bacteria on Km-selective and nonselective media
showed that incorporation of plant residues into
soil increased the total cfu numbers as well as the
number of Kmr bacteria. Donegan et al. [43] also
reported that the addition of plant material nearly
always caused a signi¢cant increase in bacterial and
fungal numbers. Screening of bacteria isolated from
Km-selective media by hybridization with a probe
speci¢c to all three parts of the construct and PCR
analysis of a few weakly hybridizing cells showed the
absence of construct-speci¢c sequences in the most
dominant aerobically grown bacteria. To improve
the limit of detection, the DNA released from the
bacterial lawn (104 ^105 cfu per plate) obtained after
plating the lowest dilution of resuspended soil samples from microcosms was analyzed by PCR using
the construct-speci¢c primers. To track the presence
of construct-speci¢c sequences in bacterial cells independently from their culturability, the DNA extracted from the bacterial fraction extracted directly
from soil samples from the disposal site was analyzed by PCR. In a few samples, PCR detection of
construct-speci¢c sequences in DNA obtained from
the bacterial fraction directly extracted from soil
samples or from resuspended bacterial cell lawns obtained after a cultivation step might indicate horizontal gene transfer to members of the bacterial
community. However, the results remain di¤cult to
interpret because transient uptake of DNA, free
DNA adhering to cells or to co-extracted soil particles cannot be excluded. Therefore, this study could
not provide a valid proof that horizontal gene transfer of plant DNA to bacteria occurs under ¢eld or
microcosm conditions.
While PCR analysis of DNA extracted directly
from soil enabled speci¢c and sensitive detection of
transgenic DNA in soil, detection of horizontal gene
transfer events is di¤cult due to the limitations of
the techniques currently available. Unequivocal
proof of horizontal gene transfer from plants to bacteria requires the isolation of the putative transformants for thorough genetic characterization. However,
the strategy to monitor the transfer of complete
genes or larger DNA fragments might fail because
transformation often involves the stable integration
of short DNA fragments resulting in gene mosaics
[44]. Based on sequence comparisons, it has been
assumed that horizontal gene transfer from plants
to bacteria might have occurred during evolution
[45^47]. The presence of bacterial genes, promoters,
FEMSEC 999 8-3-99
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
terminators, or origins of vegetative replication in
transgenic plants will enhance the probability of stable integration of DNA stretches based on recombination events [10,11,28]. The lack of information on
the abundance of naturally competent bacteria in the
environment, frequencies of transformation processes and environmental factors triggering these
processes currently still impairs predictions of the
extent of horizontal gene transfer from plants to
bacteria.
Acknowledgments
This work was supported by Grant 0310642 of the
German Ministry of Education, Science, Research
and Technology. We thank E.-M. Brambilla, Julia
Boëtel, and Ellen Kroëgerrecklenfort for excellent technical assistance. Furthermore, we are grateful to
K.M. Nielsen, Trondheim (Norway) and J.D. van
Elsas, Wageningen (The Netherlands) for critically
reading and discussing the manuscript.
References
[1] Levy, S.B. (1987) Environmental dissemination of microbes
and their plasmids. Swiss Biotechnol. 5, 32^37.
[2] Tschaëpe, H. (1994) The spread of plasmids as a function of
bacterial adaptability. FEMS Microbiol. Ecol. 15, 23^32.
[3] Davies, J. (1994) Inactivation of antibiotics and the dissemination of resistance genes. Science 264, 375^382.
[4] Van Elsas, J.D and Smalla, K. (1996) Antibiotic (kanamycin
and streptomycin) resistance traits in the environment.
In : Key Biosafety Aspects of Genetically Modi¢ed Organisms
(Landsmann, J. and Casper, R., Eds.), pp. 61^69. Mitt.
Biol. Bundesanst. Land-Forstwirtsch. Berlin-Dahlem, 309,
1995.
[5] Smalla, K., Wellington, E.M.H. and van Elsas, J.D. (1997)
Natural background of bacterial antibiotic resistance genes
in the environment. In : Nordic Seminar on Antibiotic Resistance Marker Genes and Transgenic Plants, pp. 43^57. Norwegian Biotechnology Advisory Board, Oslo.
[6] Becker, J., Siegert, H., Logemann, J. and Schell, J. (1994)
Begleitende Sicherheitsforschung zur Freisetzung genetisch
veraënderter Petunien. In: Biologische Sicherheit (Buchholz,
C., Ed.), pp. 563^578. Bundesministerium fuër Forschung
und Technologie, Bonn.
[7] Schluëter, K., Fuëtterer, J. and Potrykus, I. (1995) Horizontal
gene transfer from a transgenic potato line to a bacterial
pathogen (Erwinia chrysanthemi) occurs ^ if at all ^ at an
extremely low frequency. Biotechnology 13, 94^98.
271
[8] Broer, I., Droëge-Laser, W. and Gerke, M. (1996) Examination
of the putative horizontal gene transfer from transgenic plants
to agrobacteria. In: Transgenic Organisms and Biosafety
(Schmidt, E.R. and Hankeln, T., Eds.), pp. 67^70. Springer
Verlag, Berlin.
[9] Nielsen, K.M., Gebhard, F., Smalla, K., Bones, A.M. and
Van Elsas, J.D. (1997) Evaluation of possible horizontal
gene transfer from transgenic plants to the soil bacterium
Acinetobacter calcoaceticus BD413. Theor. Appl. Genet. 95,
815^821.
[10] Gebhard, F. and Smalla, K. (1998) Transformation of Acinetobacter sp. BD413 by transgenic sugar beet DNA. Appl. Environ. Microbiol. 64, 1550^1554.
[11] De Vries, J. and Wackernagel, W. (1998) Detection of nptII
(kanamycin resistance) genes in genomes of transgenic plants
by marker-rescue transformation. Mol. Gen. Genet. 257, 606^
613.
[12] Widmer, F., Seidler, R.J., Donegan, K.K. and Reed, G.L.
(1997) Quanti¢cation of transgenic plant marker gene persistence in the ¢eld. Mol. Ecol. 6, 1^7.
[13] Aardema, B.W., Lorenz, M.G. and Krumbein, W.E. (1983)
Protection of sediment-adsorbed transforming DNA against
enzymatic inactivation. Appl. Environ. Microbiol. 46, 417^
420.
[14] Lorenz, M. and Wackernagel, W. (1987) Adsorption of DNA
to sand and variable degradation rates of adsorbed DNA.
Appl. Environ. Microbiol. 53, 2948^2952.
[15] Romanowski, G., Lorenz, M.G. and Wackernagel, W. (1991)
Adsorption of plasmid DNA to mineral surfaces and protection against DNase I. Appl. Environ. Microbiol. 57, 1057^
1061.
[16] Khanna, M. and Stotzky, G. (1992) Transformation of Bacillus subtilis by DNA bound on montmorillonite and e¡ect of
DNase on the transforming ability of bound DNA. Appl.
Environ. Microbiol. 58, 1930^1939.
[17] Paget, E., Jocteur Monrozier, L. and Simonet, P. (1992) Adsorption of DNA on clay minerals: protection against DNase
I and in£uence on gene transfer. FEMS Microbiol. Lett. 97,
31^40.
[18] Romanowski, G., Lorenz, M.G., Sayler, G. and Wackernagel,
W. (1992) Persistence of free plasmid DNA in soil monitored
by various methods, including a transformation assay. Appl.
Environ. Microbiol. 58, 3012^3019.
[19] Chamier, B., Lorenz, M.G. and Wackernagel, W. (1993) Natural transformation of Acinetobacter calcoaceticus by plasmid
DNA adsorbed on sand and groundwater aquifer material.
Appl. Environ. Microbiol. 59, 1662^1667.
[20] Nielsen, K.M., van Weerelt, D.M., Berg, T.N., Bones, A.M.,
Hageler, A.N. and Van Elsas, J.D. (1997) Natural transformation and availability of transforming chromosomal DNA
to Acinetobacter calcoaceticus in soil microcosms. Appl. Environ. Microbiol. 63, 1945^1952.
[21] Gallori, E., Bazzicalupo, M., Dal Canto, L., Fani, R., Nannipieri, P., Vettori, C. and Stotzky, G. (1994) Transformation of
Bacillus subtilis by DNA bound on clay in non-sterile soil.
FEMS Microbiol. Ecol. 15, 119^126.
[22] Lorenz, M.G. and Wackernagel, W. (1994) Bacterial gene
FEMSEC 999 8-3-99
272
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
F. Gebhard, K. Smalla / FEMS Microbiology Ecology 28 (1999) 261^272
transfer by natural genetic transformation in the environment.
Microbiol. Rev. 58, 563^602.
Paget, E. and Simonet, P. (1994) On the track of natural
transformation in soil. FEMS Microbiol. Ecol. 15, 109^118.
Widmer, F., Seidler, R.J. and Watrud, L.S. (1996) Sensitive
detection of transgenic plant marker gene persistence in soil
microcosms. Mol. Ecol. 5, 603^613.
Paget, E. and Simonet, P. (1997) Development of engineered
genomic DNA to monitor the natural transformation of Pseudomonas stutzeri in soil-like microcosms. Can. J. Microbiol.
43, 78^84.
Sikorski, J., Graupner, S., Lorenz, M.G. and Wackernagel,
W. (1998) Natural transformation of Pseudomonas stutzeri
in a non-sterile soil. Microbiology 144, 569^576.
Nielsen, K.M., Bones, A.M. and Van Elsas, J.D. (1997) Induced natural transformation of Acinetobacter calcoaceticus in
soil microcosms. Appl. Environ. Microbiol. 63, 3972^3977.
Bertolla, F., van Gijsegem, F., Nesme, X. and Simonet, P.
(1997) Conditions for natural transformation of Ralstonia solanacearum. Appl. Environ. Microbiol. 63, 4965^4968.
Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular
Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
Smalla, K., Van Overbeek, L.S., Pukall, R. and Van Elsas,
J.D. (1993) Prevalence of nptII and Tn5 in kanamycin-resistant bacteria from di¡erent environments. FEMS Microbiol.
Ecol. 13, 47^58.
Herron, P.R. and Wellington, E.M.H. (1990) New method for
the extraction of streptomycete spores from soil and application to the study of lysogeny in sterile amended and non
sterile soil. Appl. Environ. Microbiol. 56, 1406^1412.
Hopkins, D.W., Macnaughton, S.J. and O'Donnell, A.G.
(1991) A dispersion and di¡erential centrifugation technique
for representatively sampling microorganisms from soil. Soil
Biol. Biochem. 23, 217^225.
Bakken, L.R. (1985) Separation and puri¢cation of bacteria
from soil. Appl. Environ. Microbiol. 49, 1482^1487.
Smalla, K. and Van Elsas, J.D. (1995) Application of the PCR
for detection of antibiotic resistance genes in environmental
samples. In: Nucleic Acids in the Environment (Trevors, J.T.
and Van Elsas, J.D., Eds.), pp. 241^256. Springer Verlag,
Berlin.
Van Elsas, J.D. and Smalla, K. (1995) Extraction of microbial
community DNA from soils. In: Molecular Microbial Ecol-
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
ogy Manual (Akkermans, A.D.L. and Van Elsas, J.D., Eds.),
pp. 1.3.3/1^11. Kluwer Academic, Dordrecht.
Trinker, N.A., Fortin, M.G. and Mather, D.E. (1993) Random ampli¢ed polymorphic DNA and pedigree relationship in
spring barley. Theor. Appl. Genet. 85, 976^984.
Van Elsas, J.D. and Smalla, K. (1996) Methods for sampling
soil microbes. In: Manual of Environmental Microbiology
(Hurst, C.J., Knudsen, G.R., McInerney, M.J., Stetzenbach,
L.D. and Walter, M.V., Eds.), pp. 383^390. ASM Press,
Washington, DC.
Rasch, D., Guiard, V. and Nuërnberg, G. (1992) Statistische
Versuchsplanung : Einfuëhrung in die Methoden und das Dialogsystem CADEMO. Gustav Fischer Verlag, Stuttgart.
Blum, S.A.E., Lorenz, M.G. and Wackernagel, W. (1997)
Mechanisms of retarded DNA degradation and prokaryotic
origin of DNases in nonsterile soils. Syst. Appl. Microbiol. 20,
513^521.
Ogram, A., Sayler, G.S., Gustin, D. and Lewis, R.J. (1988)
DNA adsorption to soils and sediments. Environ. Sci. Technol. 22, 982^984.
Nielsen, K.M., Bones, A.M., Smalla, K. and Van Elsas, J.D.
(1998) Horizontal gene transfer from transgenic plants to terrestrial bacteria ^ a rare event ? FEMS Microbiol. Rev. 22, 79^
103.
Paul, J.H., Thurmond, J.M., Frischer, M.E. and Cannon, J.P.
(1992) Intergeneric natural plasmid transformation between E.
coli and a marine Vibrio species. Mol. Ecol. 1, 37^46.
Donegan, K.K., Palm, C.J., Fieland, V.J., Porteous, L.A.,
Ganio, L.M., Schaller, D.L., Bucao, L.Q. and Seidler, R.J.
(1995) Changes in levels, species and DNA ¢ngerprints of
soil microorganisms associated with cotton expressing the Bacillus thuringiensis var. kurstaki endotoxin. Appl. Soil Ecol. 2,
111^124.
Smith, J.M., Dowson, C.G. and Spratt, B.G. (1991) Localized
sex in bacteria. Nature 349, 29^31.
Wakabayashi, S., Matsubara, H. and Webster, D.A. (1986)
Primary sequence of a dimeric bacterial haemoglobin from
Vitreoscilla. Nature 322, 481^483.
Doolittle, R.F., Feng, D.F., Anderson, K.L. and Alberro,
M.R. (1990) A naturally occurring horizontal gene transfer
from a eukaryote to a prokaryote. J. Mol. Evol. 31, 383^388.
Smith, M.W., Feng, D.-F. and Doolittle, R.F. (1992) Evolution by acquisition : the case for horizontal gene transfers.
Trends Biochem. Sci. 17, 489^493.
FEMSEC 999 8-3-99