Journal of Microbiological Methods 75 (2008) 12–18
Contents lists available at ScienceDirect
Journal of Microbiological Methods
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j m i c m e t h
A novel method for sampling bacteria on plant root and soil surfaces at the
microhabitat scale
Paul G. Dennis a,b,c,⁎, Anthony J. Miller a, Ian M. Clark a, Richard G. Taylor c,
Eugenia Valsami-Jones b, Penny R. Hirsch a
a
b
c
Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK
Department of Mineralogy, The Natural History Museum, Cromwell Road, London SW7 5BD, UK
Department of Geography, University College London, Gower Street, London WC1E 6BT, UK
A R T I C L E
I N F O
Article history:
Received 27 January 2008
Received in revised form 3 April 2008
Accepted 21 April 2008
Available online 18 May 2008
Keywords:
Micro-sampling
Rhizosphere bacteria
Soil bacteria
Microhabitat
A B S T R A C T
This study reports the first method for sampling bacteria at a spatial scale approximating a microhabitat. At the
core of this method is the use of tungsten rods with laser-cut tips of known surface area (0.013 mm2). Exposed
plant root or soil surfaces were viewed with a dissecting microscope and micro-sampling rods were guided to
sample sites using a micro-manipulator. Bacteria that adhered to the sampling tips were then recovered for
microbiological analyses. The efficiency of this method for removing bacteria from root surfaces was similar to
that with which bacteria are recovered from dissected root segments using the conventional technique of
washing. However, as the surface area of the micro-sampling tips was known, the new method has the advantage
of eliminating inaccuracy in estimates of bacterial densities due to inaccurate estimation of the root or soil surface
sampled. When used to investigate spatial distributions of rhizoplane bacteria, the new technique revealed trends
that were consistent with those reported with existing methods, while providing access to additional information
about community structure at a much smaller spatial scale. The spatial scale of this new method is ca. 1000-times
smaller than other sampling methods involving swabbing. This novel technique represents an important
methodological step facilitating microbial ecological investigations at a microhabitat scale.
Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
1. Introduction
Rhizosphere bacteria affect plant health and nutrition, but a detailed
understanding of their interactions in the rhizosphere is limited. In part,
this is due to a lack of methods that facilitate rhizosphere bacterial
community attributes, such as density, composition and diversity, to be
linked with habitat factors (e.g. pH and/or nutrient availability) at, or
close to, the microhabitat scale. In this context, a microhabitat is the area
surrounding a bacterial cell, population or community where the
interactions between the bacteria and their environment are particularly strong. Given the size of bacteria and the variability of soil and
rhizosphere environments the spatial scale of rhizosphere bacterial
microhabitats is likely to be very small, at the μm scale. Current
technologies, such as ion-selective microelectrodes, allow certain ion
activities to be measured at this scale (Plassard et al., 2002); however, to
link such measurements with rhizosphere bacterial community
attributes, novel microbiological methods are needed. This paper
reports the first method for sampling bacteria from plant root and soil
surfaces at a spatial scale that approximates a microhabitat.
⁎ Corresponding author. Present address: SCRI Living Technology, Invergowrie,
Dundee, DD2 5DA, UK. Tel.: +44 1382 562731; fax: +44 1382 562426.
E-mail addresses: [email protected], [email protected] (P.G. Dennis).
To investigate links between bacteria and their microhabitat, it is
necessary to preserve the spatial organisation of the environment to be
sampled. For rhizosphere research this is best achieved using specialized
plant growth containers (rhizotrons) designed to encourage root growth
along the surface of a soil monolith that can be exposed for visualization
and/or sampling of the root/soil surface. Rhizosphere bacterial density
can be determined microscopically (Newman and Bowen, 1974; Rovira
et al., 1974); however, community diversity and composition generally
require that a sample is collected. The microscopy-based technique,
fluorescent in-situ hybridization (FISH), facilitates the differentiation of
only limited numbers of bacterial types at the same time, restricted by
the number of fluorophores that are available (Amann et al., 1996;
Lenaerts et al., 2007). Furthermore, as FISH depends on rRNA
hybridization probes, a priori knowledge of the studied organisms is
required. Therefore, to characterize multiple attributes in poorly characterized bacterial communities, samples must be obtained.
Transfer of bacteria from a sample surface to a suitable medium for
microbiological analyses can be achieved directly by washing, or
indirectly, by washing bacteria from a surface previously used to blot
the sample. To investigate spatial distributions of rhizosphere bacteria
using direct washing techniques, it is necessary to section the root. For
investigations of bacterial communities at the micro-spatial scale,
however, excision of sub-millimetre root sections is challenging. Furthermore, the resulting sections are of variable size, and prone to
0167-7012/$ – see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.mimet.2008.04.013
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
desiccation and contamination. The destructive nature of direct washing
techniques also means that microbial community dynamics on the same
plant cannot be monitored over time. In our opinion, the most versatile
method for sampling rhizosphere bacteria reported to date involves
blotting root or soil surfaces with filter papers of known surface area
(20 mm2; Marschner and Crowley,1996). The principal advantages of this
technique are that it is minimally-destructive and facilitates expression
of bacterial community attributes per unit area. However, the technique
is limited to macro-spatial scale sampling as the preparation and
handling of filter papers smaller than 1 mm is not feasible.
The objectives of our study were to: i) develop the first method for
sampling bacteria at a spatial scale approximating a microhabitat; ii)
compare how efficiently the new method removes bacteria from root
surfaces with the conventional technique of root dissection/washing;
and iii) demonstrate that when used to investigate spatial distributions of rhizoplane bacteria the new technique reveals trends
consistent with those reported with existing methods while providing
access to additional information concerning community structure at a
much smaller spatial scale. At the core of this method is the use of
micro-engineered tungsten rods which can be directed towards
sampling targets on root and soil surfaces by micromanipulation.
13
2. Materials and methods
2.1. Preparation of micro-sampling rods
Metal wire is used routinely to transfer bacteria from one environment to another, e.g. microbiological loops and inoculators.
Tungsten is a metal with low toxicity, high tensile strength, melting
point and corrosion resistance compared to most other metals.
Tungsten rods are cheap and readily available in sub-millimetre
diameter sizes. The ends of tungsten rods (diameter 130 µm; Science
Products GmbH, Hofheim, Germany) were laser-cut at the Laser
Micromachining Centre IBMM (Bangor, Wales, UK; Fig. 1). The rods
were mounted in single-barreled borosilicate glass capillary tubes
(outer and inner diameters 1.0 and 0.58 mm, respectively; Hilgenberg
Glass, Malsfeld, Germany) with a pointed tip produced by heating and
pulling the glass using a PE-2 vertical micropipette puller (Narishige,
Tokyo, Japan). The pointed tips of the glass tubes were blunted to allow
the cut rod to be inserted and fixed in place with Loctite® Easy Brush
Super Glue (Henkel Consumer Adhesives, Cheshire, UK; Fig. 1). Rods
were sterilized by washing in ethanol for 30 s then sterile water for 30 s
followed by 30 min ultra-violet irradiation with the tips positioned
Fig. 1. Preparation and use of micro-sampling tips. Scanning electron micrographs of un-washed sampling rods a) before and b) after laser-cutting; (c) micro-sampling rod (i)
embedded in a borosilicate glass capillary (ii) using super glue (iii). Equipment used for micro-sampling (d). The image shows B. napus plants in a growth pouch. A hole was cut in the
plastic to enable a micro-sampling rod to be directed towards a sampling site on the rhizoplane (e) or a soil crumb (f) using a micro-manipulator.
14
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
upright in a laminar flow cabinet. Autoclaving the assembled rods is
not practical as the glue degrades during this treatment.
2.2. Plants and growth system
Growth pouches (145 × 155 mm; Mega International, Minneapolis,
MN, USA) were filled with 100 g of general purpose horticultural soil
(Petersfield Products, Leicester, UK) and 50 ml water. One Brassica
napus seedling, pre-germinated for 16 h on tissue paper dampened
with 0.2 mM CaSO4, was then planted in each pouch. Pouches were
placed in envelopes to protect the roots from light and maintained in a
controlled environment chamber (day length 16 h; photon flux density
300 µmol m− 2 s− 1; temperature 20 °C; humidity 70%; night length 8 h;
temperature 18 °C, humidity 80%). The pouches were positioned at a 45°
angle to ensure root growth occurred along the surface of the soil. In this
system, the root growth rate was approximately 1.8 cm per day.
2.3. Comparison of sampling efficiencies between methods
2.3.1. Root dissection method
Four days after planting, a sterile scalpel blade was used to cut 1 cm
segments from the root base and apex of 24 plants. To remove adhered
rhizosphere soil, each root segment was gently agitated using sterile
forceps in sterile phosphate buffered saline (PBS; 137 mM NaCl,
1.5 mM KH2PO4, 6.5 mM Na2HPO4·H2O, 2.7 mM KCl, pH 7.2)
for 5–10 s. After this initial washing process root segments were
transferred to vials containing 15 ml of sterile PBS buffer. Rhizoplane
bacteria were extracted at 4 °C for 45 min by vigorous shaking at
2200 rpm using a IKA-Vibrax-VXR shaker (Janke and Kunkel, Staufen,
Germany). Extracts were diluted and then plated on 1/10 TSA (3 g L− 1
Tryptic Soy Broth solidified with 1.5% Technical Agar No. 3, Oxoid,
Basingstoke, UK) containing 100 µg ml− 1 cycloheximide (Sigma, UK).
Plates were incubated for ten days at 28 °C and CFU were recorded
after three, seven and ten days. The maximum count for each plate
was used as the final result. The 1/10th TSA was chosen as it is used
routinely in soil and rhizosphere microbiology and it supports a wide
range of commonly isolated bacterial groups (Davis et al., 2005).
2.3.2. Micro-sampling method
Four days after planting, two micro-samples per plant were taken from
the root apex (0–10 mm from the root cap) and one from the root base (0–
20 mm from the root–shoot junction). Twenty eight plants were sampled
in total. During sampling, the roots were viewed with a dissecting
microscope and the micro-sampling rods were guided to sample sites
using a micro-manipulator (Prior Scientific, Cambridge, UK; Fig. 1). As the
roots were exposed in rhizotrons it was possible to ensure that only the tip
of the sampling rods touched the rhizoplane (Fig.1). Light contact between
the sampling tip and the root was maintained for 10 s; this process did not
damage the root surface. The tip was then withdrawn from the sample
specimen and removed from the manipulator. Bacteria were recovered
from the tip in a microtube containing 10 µl PBS by sonicating for 60 s at
low power in a sonicating water bath (Ultrasonik 300; JM Ney, Bloomfield,
CT, USA) containing ice and water. This level of sonication did not appear to
be detrimental to bacteria as viable counts were similar to those obtained
when tips were agitated in buffer by hand (data not shown). Following
their primary wash, rods were placed in new microtubes each containing
10 µl sterile PBS and the washing process was repeated. This process
enabled the efficiency with which bacteria were recovered by washing to
Fig. 2. ERIC profiles from colonies in micro-samples ran on a 1.5% agarose gel with 100 bp ladders (a). Gel images are analysed using specialist computer software (b) that organises
ERIC fingerprints into a dendrogram (c). The final branches of the dendrogram represent the cluster or bacterial type (d).
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
15
be assessed. Sterile controls were included to monitor for contamination.
The samples were made up to a final volume of 200 µl sterile PBS, spread
on agar plates and the culturable bacteria were counted as above.
2.3.3. Statistical analysis
For both sampling methods, bacterial densities (i.e. CFU tip− 1 for
micro-samples and CFU cm− 1 of root for the root dissection method) four
days after planting, were expressed as CFU mm− 2 of root. The surface
area of each dissected root segment was assumed to be that of a 10 mm
long cylinder with a diameter of 1 mm minus its ends. Differences in
bacterial densities (CFU mm− 2 root) between the two sampling methods
within the apical and basal root zones were then investigated using the
method Residual Maximum Likelihood (REML; Patterson and Thompson, 1971). This method took account of the unbalanced design structure
(unequal numbers of plants between sampling methods) and a data
transformation (square root (CFU mm− 2 root + 1)) ensured that the
assumption of Normality was fulfilled. A mixed model was fitted with
random effects due to nested design terms (plants, and zones within
plants) and fixed effects due to the treatment terms (zone, method and
their interaction). The effects of treatment terms were tested for
significance using the Wald test (Welham and Thompson, 1997). The
REML method is similar to ANOVA in that it allows the mean of treatment
groups to be compared. However, in contrast to ANOVA, it is well suited
to analysis of unbalanced data (Patterson and Thompson, 1971). Other
groups have applied REML to analyse data from microbial sampling (e.g.
Hamilton et al., 2006; Bremer et al., 2001). All REML analyses were
implemented using the GenStat statistical system (GenStat 8th edition,
Lawes Agricultural Trust; VSN International, Hemel Hempstead, UK).
2.4. Spatiotemporal trends in culturable bacterial density and diversity
2.4.1. Micro-sampling and enumeration of culturable bacteria
Four, six and eight days after planting, two samples per plant were
taken from the root apex and one from the root base. In addition, two
samples per plant were taken from randomly selected points in the
bulk soil by touching the tip on a soil crumb. Bacteria adhered to the
tip were recovered and enumerated as described above.
2.4.2. Colony picking and DNA extraction
Five plants were selected with basal region bacterial counts closest
to the overall basal region mean density, both four and six days after
planting. The selection provided 262 colonies from which DNA was
obtained by the alkali extraction method of Klimyuk et al. (1993).
Fig. 3. REML predicted mean bacterial density mm− 2 root (SQRT CFU + 1) within each zone
by sampling method combination: root dissection method (●), micro-sampling method
(○). Numbers shown are the back-transformed REML predicted means. Error bars
represent standard errors of means. Converting these values to express them as CFU per
sampling tip for comparison with Figs. 5 and 6: 305 = 4, 675= 9, 1520 = 20, 44520 = 579.
Fig. 4. Basal region including lower stem (a) and apical region (b) of a 4 day-old B. napus root.
High densities of root hairs and attached soil particles were associated with the root base but
were absent at the root apex. As bacterial density for the root dissection/washing method was
calculated by relating bacterial numbers to the surface area of a 10 mm long cylinder with a
diameter of 1 mm minus its ends, the presence of root hairs and attached soil particles in basal
root segments lead to an overestimation of bacterial density. Scale bar equals 1 mm.
2.4.3. DNA fingerprinting of culturable bacteria
PCR fingerprinting, using Enterobacterial Repetitive Intergenic Consensus (ERIC) primers (Versalovic et al., 1991), was routinely performed
using: 1 µl DNA sample, 12.5 µl REDTaq™ 2× ReadyMix™ PCR Reaction
Mix (Sigma, UK), 50 pmole R1CIRE (5′- CAC TTA GGG GTC CTC GAA
TGT A-3′), 50 pmole ERIC2 (5′- AAG TAA GTG ACT GGG GTG
AGC G-3′), 1 µl 2.5 mM MgCl2 and 8.5 µl water per reaction. The
following thermocycler (T Gradient; Biometra, Göttingen, Germany)
conditions were used: 30 cycles of 94 °C for 1 min, 46 °C for 1 min,
72 °C for 1 min, final step 72 °C for 5 min, then maintained at 4 °C. PCR
products and size markers (100 bp DNA ladder; Fermentas Life Sciences,
York, UK) were run on 1.5% agarose gels and stained with ethidium
bromide prior to image capture using a transilluminator. ERIC primers
were originally designed to target Intergenic Repeat Units, know as ERIC
sequences, that are short (127 nucleotides long) palindromic repetitive
DNA elements found in enterobacteria (Versalovic et al., 1991). However,
their specificity is so poor that ERIC-PCR patterns can be obtained from
bacteriophages, invertebrates, fungi, plants and vertebrates (Hulton et al.,
1991; Gillings and Holley, 1997). The method is now used routinely to
produce highly reproducible inter and intra species specific profiles. Here
ERIC-PCR fingerprinting was used to demonstrate that the tips picked a
Fig. 5. Numbers of bacteria (CFU) in micro-samples within treatment combinations: BR
(basal region), AR (apical region), BS (bulk soil), 4, 6, and 8 days after planting.
16
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
ward comparisons between different curves. It was not possible to detect
correspondence between bacterial types within the treatment combinations as too little of the variation was accounted for in the first three
inertias (23%; implemented using the GenStat statistical system). Therefore, the number of bacterial types that appeared in more than one
sample within each treatment combination was assessed manually.
3. Results
3.1. Preparation of micro-sampling rods
Fig. 6. REML predicted mean bacterial densities per sampling tip (Log10 CFU + 0.5)
within each zone by plant age combination: apical region (●), basal region (○), bulk soil
(▼). Error bars represent standard errors of means. Numbers shown are the backtransformed REML predicted means for comparison with Figs. 3 and 5.
wide range of bacterial types and that the method could reproduce known
trends reliably. The ERIC primer binding sites are dependent on the
genomic sequence and vary in number and distribution, thus bacterial
clusters can be discriminated by the number and size of PCR products on a
gel. ERIC profiles were analysed using Phoretix 1D analysis software v.
2003.02 (Phoretix International, Newcastle upon Tyne, UK). Lanes were
identified; then bands were assigned and matched by comparison with
molecular size ladders (100 bp). Results were summarized in a
dendrogram which clustered the profiles obtained from each colony
depending on their phenetic similarity. Each cluster was designated an
operational taxonomic unit (OTU) for diversity analyses (Fig. 2).
2.4.4. Statistical analyses
Differences in mean bacterial density, i.e. CFU tip− 1 between
treatment combinations (plant ages within zones) were compared
using the REML method which accounted for the unbalanced design
structure (unequal numbers of replicates between zones) and any
correlation between plant ages, given the repeated measures on plants.
Data were transformed (log10 (CFU tip− 1 + 0.5)); this ensured that the
assumption of Normality was fulfilled. A mixed model was fitted with
random effects due to nested design terms (plants, zones within plants,
and plant ages within zones within plants) and fixed effects due to the
treatment terms (zones, plant ages and their interaction).
Individual-based rarefaction curves were used to investigate the effect
of plant ages and zones on taxonomic diversity. These were computed by
EstimateS (Colwell, 2005) using the method of Colwell et al. (2004). This
method estimates rigorous confidence intervals that enable straightfor-
Laser-cutting produced rods with flat tips perpendicular to their
length and standardized surface area (0.013 mm2; Fig. 1). By mounting
the laser-cut rods in glass capillaries they were compatible with the
micro-manipulator. This in conjunction with the microscope made it
easy to achieve finely controlled movement towards targeted sampling
sites on the rhizoplane or soil surfaces (Fig. 1). It was possible to observe
when the tip had made contact with the root and control the pressure
with which the tip pressed against it thus avoiding any damage.
3.2. Comparison of the sampling efficiency between sampling methods
Bacterial density was significantly different (Wald tests, P b 0.05)
between root zones irrespective of the sampling method (Fig. 3). Using
the micro-sampling approach, bacterial densities were approximately
1.5 × 103 103 CFU mm− 2 root at the base and 3 × 102 102 CFU mm− 2 root at
the apex. The corresponding values using the root dissection method
were 4 × 104 104 CFU mm− 2 root at the base and 7 ×102 102 CFU mm− 2 root
at the apex. Two-tailed t-tests revealed that between methods, there was
no difference in bacterial densities at the root apex, but at the root base
the bacterial density was significantly greater using the root dissection
method (P b 0.05). The surface area of dissected root segments used for
calculating bacterial density was similar to the actual root surface area
(RSA) of apical root segments. However, the actual RSA of basal root
segments was much greater due to the abundance of root hairs and
adhered soil particles (Fig. 4). Both methods revealed greater bacterial
density at the root base than at the apex (2-tailed student t-test, P b 0.05).
Micro-sampling rods were clearly able to pick-up bacteria and the
efficiency with which they were washed off the tip was excellent —
carry-over of bacteria on the tips between the first and second washes
was never detected. In addition, the absence of colonies in control plates
indicates that sterilization was effective (data not shown).
3.3. Spatiotemporal trends in culturable bacterial density and diversity
Bacterial density was highly variable between samples indicating a
high degree of patchiness at the micro-scale (Fig. 5). The treatment terms
(zones, plant ages and their interaction) all had a significant effect (Wald
tests, Pb 0.05) on bacterial densities. Although there were repeated
Fig. 7. Individual-based rarefaction curves (solid lines) with their associated 95% confidence limits (dotted lines), representing bacterial diversity: 4 (a) and 6 (b) days after planting,
on the rhizoplane (black) or in bulk soil (grey).
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
measures within plants, there was no significant autocorrelation between
plant ages within zones within plants (PN 0.05). At the centimetre scale
(pooled micro-samples), the density of culturable bacteria followed the
order: basal regionN bulk soilN apical region (Fig. 6). Using standard errors
with the exception of the basal region and bulk soil, four and eight days
after planting, all zones were significantly different from one another. A
significant temporal effect on bacterial density was observed only in the
basal region between four and six days.
ERIC profiling revealed that the 262 colonies analysed represented a
total of 135 OTU. Only 4.6% of these OTU were observed on more than
one plant within each zone within plant age category. Plant age had a
significant effect on bacterial diversity; however, differences between
zones were not significant. Nonetheless, bacterial diversity within zones
followed the order: bulk soil N apical region N basal region. Differences in
bacterial diversity between root zones within plants of the same age, or
between plants of different ages within root zones were not significant;
however, the order bulk soil N apical region N basal region was observed
in both plant age categories, and increasing dominance was more
apparent in root samples than in soil samples. In pooled root samples,
bacterial diversity decreased significantly between four and six days,
whereas in the bulk soil there was no change. The difference between
root and bulk soil samples was significant six days after planting (Fig. 7).
4. Discussion
4.1. Preparation of micro-sampling rods
The diameter of the tungsten rods varied by less than 1 µm, so the
laser-cutting resulted in tips of adequately standardized surface area
(Fig. 1). Mounting them in glass capillaries enabled them to be used with
the micro-manipulator, which, in conjunction with a microscope made it
simple to achieve finely controlled movement towards targeted sampling
sites on soil surfaces or the rhizoplane. For rhizoplane samples, this level of
control made it possible to see when the tip had made contact with the
root and to control the pressure with which the tip pressed against it, thus
avoiding damage to the epidermal cells. The results clearly demonstrate
that the micro-sampling tips picked up bacteria and indicate that the tipwashing, sterilizing processes were effective. It is possible that the microsampling tips pick-up some bacterial species more effectively than others
due to the properties of the tungsten surface. There are many different
complex processes involved in the bacterial attachment to metals, the
most important of which are thought to be electrostatic and hydrophobic
interactions (van Loosdrecht et al., 1990). This bacterial sampling method
will be particularly useful in comparative studies in the rhizosphere.
4.2. Comparison of the sampling efficiency between sampling methods
At the root base the root dissection/washing method yielded ca. 30
times greater bacterial densities (CFU mm− 2 root) than the microsampling technique, but no significant difference was observed at the
root apex (Fig. 3). This suggests that the efficiency of the sampling
methods is similar at the root apex and that in basal root zones the root
dissection/washing method is more efficient. However, at the root
base, bacterial density was overestimated by the root dissection/
washing method. This is because dissected basal root segments were
more or less covered with root hairs and adhered soil particles, which
greatly increased the sampled surface area, but were not accounted for
when calculating bacterial density. The estimate generated by the root
dissection/washing method at the root apex was more accurate due to
the absence of root hairs and adhered soil particles (Fig. 4). Under some
environmental conditions there may be fewer root hairs in the basal
region and this variable needs to be assessed in any experiment for the
adequate interpretation of data. It appears, therefore, that the new
micro-sampling technique yields a similar number of bacterial cells as
the root dissection/washing method, but is faster and eliminates
inaccuracy due to sampled surface area.
17
4.3. Spatiotemporal trends in culturable bacterial density and diversity
In all plant age categories bacterial density followed the order:
basal region N bulk soil N apical region (Fig. 6) which is in agreement
with others (Parke et al., 1986; Olsson et al., 1987; Liljeroth et al., 1991;
Chin-A-Woeng et al., 1997; Duineveld and van Veen, 1999), highlighting that the micro-sampling technique generates data sets that
reliably reproduce known trends at the macro-scale (i.e. centimetre
scale) while providing additional micro-scale information. For
example, between individual micro-samples, bacterial density was
highly heterogeneous and varied between zero and approximately
one thousand bacteria per tip (Fig. 5). Others have reported similar
observations using scanning electron and optical microscopy (Rovira
et al., 1974; Newman and Bowen, 1974, Chin-A-Woeng et al., 1997;
Nunan et al., 2001; Nunan et al., 2003; Gamalero et al., 2004; Watt
et al., 2006); however, the advantage of the micro-sampling approach
over microscopy-based techniques is that it also facilitates assessment
of bacterial community attributes such as composition and diversity.
The ERIC-PCR fingerprinting revealed that less than 5% of bacterial
OTU were present in more than one plant/soil replicate within the
treatment combinations (zones within plant age categories). In addition,
individual-based rarefaction curves showed no sign of reaching an
asymptote and were highly linear. These findings show that the microsampling tips picked a wide range of bacterial types and are in agreement with others (Grundmann and Debouzie, 2000; Grundmann and
Normand, 2000) demonstrating that bacterial community composition
is heterogeneous at the micro-spatial scale. The individual-based
rarefaction curves also showed that in contrast to soil samples, bacterial
diversity within pooled root samples decreased over time. These
observations support other studies in which bacterial diversity was
found to decrease with increasing proximity to a root (Marilley et al.,
1998; Marilley and Aragno, 1999).
4.4. Special considerations for micro-sample analysis
A micro-sample typically contains less than 50 bacterial cells; therefore, to ensure that the appropriate community attributes are measured,
special consideration must be given to their analysis. For example, any
division of such a small sample is likely to result in sub-samples that
poorly represent the sampled community. In addition, the very low cell
numbers are problematic for culture-independent analyses as any minor
contamination from bacterial DNA, for example in reagents (Corless
et al., 2000; Nadkarni et al., 2002; Shen et al., 2006) or attached to microsampling tips, will have a major effect. It is possible to eliminate DNA
contamination in PCR reagents (Rochelle et al., 1992; Maiwald et al.,
1994; Steinman et al., 1997; Carroll et al., 1999; Sleigh et al., 2001);
however, elimination of contaminating bacterial DNA on the microsampling tip surfaces is likely to be more of a challenge. This may be
achieved using the techniques for decontaminating PCR reagents and is a
future avenue for further development of the micro-sampling technique.
Standard cultivation methods are widely reported to underestimate
both bacterial density and diversity (Weisburg et al., 1991); however,
cultivation facilitates enumeration and fingerprinting of individual
bacterial colonies, thereby allowing diversity comparisons using
individual-based rarefaction. Culture-independent methods are likely
to be more useful where questions concern community composition
alone. For example, the method could be used to probe for bacterial
species that commonly dominate features such as sites of pathogen
infection.
4.5. Concluding remarks
The micro-sampling method may be applied not only to soil surfaces
and the rhizoplane but potentially to any surface. Any limitation on the
analysis of microbial communities due to size of the rod (surface area)
can be overcome by changing the dimensions of the metal wire used. A
18
P.G. Dennis et al. / Journal of Microbiological Methods 75 (2008) 12–18
future challenge is to combine the micro-sampling technique with
culture-independent analytical techniques. This would allow for a more
comprehensive analysis of the microbial communities sampled. The
data collected with the micro-sampling technique has demonstrated
that the density and distribution of bacterial types in soil and on the
rhizoplane are highly heterogeneous at the micro-spatial scale. It should
be emphasized that most macro-ecological studies consider the
influence of organisms on their environment and vice versa at the
habitat scale; however, due to a lack of suitable methods, similar studies
concerning microorganisms have been impossible. Our new microsampling technique represents an important methodological advance
and should make a valuable contribution to the field of surface-based
microbial ecology.
Acknowledgements
We gratefully acknowledge Prof. David Hopkins for his help with
the manuscript, Dr Stephen Powers (Rothamsted Research, Harpenden, UK) for statistical advice and Dr Alex Ball (The Natural History
Museum, London, UK) for the scanning electron microscopy.
Rothamsted Research receives grand-aided support from the Biotechnology and Biological Sciences Research Council (BBSRC) of the United
Kingdom. PGD thanks the BBSRC and the Natural History Museum for
a CASE studentship.
References
Amann, R., Snaidr, J., Wagner, M., Ludwig, W., Schleifer, K.H., 1996. In situ visualization of
high genetic diversity in a natural microbial community. J. Bacteriol. 178,
3496–3500.
Bremer, P.J., Monk, I., Osborne, C.M., 2001. Survival of Listeria monocytogenes attached to
stainless steel surfaces in the presence or absence of Flavobacterium spp. J. Food
Prot. 64, 1369–1376.
Carroll, N.M., Adamson, P.O., Khravi, N., 1999. Elimination of bacterial DNA from Taq
DNA polymerases by restriction endonuclease digestion. J. Clin. Microbiol. 37,
3402–3404.
Chin-A-Woeng, T.F.C., de Priester, W., van der Bij, A.J., Lugtenberg, B.J.J., 1997. Description
of the colonization of a gnotobiotic tomato rhizosphere by Pseudomonas fluorescens
biocontrol strain WCS365, using scanning electron microscopy. Mol. Plant-Microb.
Interact. 10, 79–86.
Colwell, R.K., 2005. EstimateS: Statistical Estimation of Species Richness and Shared
Species from Samples. Version 7.5. User's guide and application published at:
http://purl.oclc.org/estimates.in.
Colwell, R.K., Mao, C.X., Chang, J., 2004. Interpolating, extrapolating, and comparing
incidence-based species accumulation curves. Ecology 85, 2717–2727.
Corless, C.E., Guiver, M., Borrow, R., Edwards-Jones, V., Kaczmarski, E.B., Fox, A.J., 2000.
Contamination and sensitivity issues with a real-time universal 16S rRNA PCR.
J. Clin. Microbiol. 38, 1747–1752.
Davis, K.E.R., Joseph, S.J., Janssen, P.H., 2005. Effects of growth medium, inoculum size,
and incubation time on culturability and isolation of soil bacteria. Appl. Environ.
Microb. 71, 826–834.
Duineveld, B.M., van Veen, J.A., 1999. The number of bacteria in the rhizosphere during
plant development: relating colony-forming units to different reference units. Biol.
Fertil. Soils 28, 285–291.
Gamalero, E., Lingua, G., Giusy Capri, F., Fusconi, A., Berta, G., Lemanceau, P., 2004.
Colonization pattern of primary tomato roots by Pseudomonas fluorescens A6ri
characterized by dilution plating, flow cytometry, fluorescence, confocal and
scanning electron microscopy. FEMS Microbiol. Ecol. 48, 79–87.
Gillings, M., Holley, M., 1997. Repetitive element PCR fingerprinting (rep-PCR) using
enterobacterial repetitive intergenic consensus (ERIC) primers is not necessarily
directed at ERIC elements. Lett. Appl. Microbiol. 25, 17–21.
Grundmann, G.L., Debouzie, D., 2000. Geostatistical analysis of the distribution of NH+4
and NO−2oxidizing bacteria and serotypes at the millimeter scale along a soil
transect. FEMS Microbiol. Ecol. 34, 57–62.
Grundmann, G.L., Normand, P., 2000. Microscale diversity of the genus nitrobacter in
soil on the basis of analysis of genes encoding rRNA. Appl. Environ. Microb. 66,
4543–4546.
Hamilton, A.J., Stagnitti, F., Premier, R., Boland, A.M., Hale, G., 2006. Quantitative
microbial risk assessment models for consumption of raw vegetables irrigated with
reclaimed water. Appl. Environ. Microb. 72, 3284–3290.
Hulton, C.S.J., Higgins, C.F., Sharp, P.M., 1991. ERIC sequences — a novel family of
repetitive elements in the genomes of Escherichia coli, Salmonella typhimurium and
other enterobacteria. Mol. Microbiol. 5, 825–834.
Klimyuk, V.I., Carroll, B.J., Thomas, C.M., Jones, J.D.G., 1993. Alkali treatment for rapid
preparation of plant-material for reliable PCR analysis. Plant J. 3, 493–494.
Lenaerts, J., Lappin-Scott, H.M., Porter, J., 2007. Improved fluorescent in situ
hybridization method for detection of bacteria from activated sludge and river
water by using DNA molecular beacons and flow cytometry. Appl. Environ. Microb.
73, 2020–2023.
Liljeroth, E., Burgers, S., van Veen, J.A., 1991. Changes in bacterial populations along roots
of wheat (Triticum aestivum L) seedlings. Biol. Fertil. Soils 10, 276–280.
Maiwald, M., Ditton, H.J., Sonntag, H.G., Vonknebeldoeberitz, M., 1994. Characterization
of contaminating DNA in Taq polymerase which occurs during amplification with a
primer set for legionella 5S ribosomal-RNA. Mol. Cell. Probes 8, 11–14.
Marilley, L., Aragno, M., 1999. Phylogenetic diversity of bacterial communities differing
in degree of proximity of Lolium perenne and Trifolium repens roots. Appl. Soil Ecol.
13, 127–136.
Marilley, L., Vogt, G., Blanc, M., Aragno, M., 1998. Bacterial diversity in the bulk soil and
rhizosphere fractions of Lolium perenne and Trifolium repens as revealed by PCR
restriction analysis of 16S rDNA. Plant Soil 198, 219–224.
Marschner, P., Crowley, D.E., 1996. Physiological activity of a bioluminescent Pseudomonas fluorescens (strain 2-79) in the rhizosphere of mycorrhizal and nonmycorrhizal pepper (Capsicum annuum L). Soil Biol. Biochem. 28, 869–876.
Nadkarni, M.A., Martin, F.E., Jacques, N.A., Hunter, N., 2002. Determination of bacterial
load by real-time PCR using a broad-range (universal) probe and primers set.
Microbiology-SGM 148, 257–266.
Newman, E.I., Bowen, H.J., 1974. Patterns of distribution of bacteria on root surfaces. Soil
Biol. Biochem. 6, 205–209.
Nunan, N., Ritz, K., Crabb, D., Harris, K., Wu, K.J., Crawford, J.W., et al., 2001.
Quantification of the in situ distribution of soil bacteria by large-scale imaging of
thin sections of undisturbed soil. FEMS Microbiol. Ecol. 37, 67–77.
Nunan, N., Wu, K.J., Young, I.M., Crawford, J.W., Ritz, K., 2003. Spatial distribution of
bacterial communities and their relationships with the micro-architecture of soil.
FEMS Microbiol. Ecol. 44, 203–215.
Olsson, S., Baath, E., Soderstrom, B., 1987. Growth of Verticillium dahliae Kleb. hyphae
and of bacteria along the roots of rape (Brassica napus L) seedlings. Can. J. Microbiol.
33, 916–919.
Parke, J.L., Moen, R., Rovira, A.D., Bowen, G.D., 1986. Soil water flow affects the
rhizosphere distribution of a seed borne biological control agent, Pseudomonas
fluorescens. Soil Biol. Biochem. 18, 583–588.
Patterson, H.D., Thompson, R., 1971. Recovery of inter-block information when block
sizes are unequal. Biometrika 58, 545–554.
Plassard, C., Guerin-Laguette, A., Very, A.A., Casarin, V., Thibaud, J.B., 2002. Local
measurements of nitrate and potassium fluxes along roots of maritime pine. Effects
of ectomycorrhizal symbiosis. Plant Cell Environ. 25, 75–84.
Rochelle, P.A., Weightman, A.J., Fry, J.C., 1992. DNAse-I treatment of Taq DNApolymerase for complete PCR decontamination. Biotechniques 13, 520.
Rovira, A.D., Newman, E.I., Bowen, H.J., Campbell, R., 1974. Quantitative assessment of
the rhizoplane microflora by direct microscopy. Soil Biol. Biochem. 6, 211–216.
Shen, H., Rogej, S., Kieft, T.L., 2006. Sensitive, real-time PCR detects low-levels of
contamination by Legionella pneumophila in commercial reagents. Mol. Cell. Probes
20, 147–153.
Sleigh, J., Cursons, R., La Pine, M., 2001. Detection of bacteraemia in critically ill patients
using 16S rDNA polymerase chain reaction and DNA sequencing. Intensive Care
Med. 27, 1269–1273.
Steinman, C.R., Muralidhar, B., Nuovo, G.J., Rumore, P.M., Yu, D., Mukai, M., 1997.
Domain-directed polymerase chain reaction capable of distinguishing bacterial
from host DNA at the single-cell level: characterization of a systematic method to
investigate putative bacterial infection in idiopathic disease. Anal. Biochem. 244,
328–339.
van Loosdrecht, M.C.M., Lyklema, J., Norde, W., Zehnder, A.J.B., 1990. Influence of
interfaces on microbial activity. Microb. Rev. 54, 75–87.
Versalovic, J., Koeuth, T., Lupski, J.R., 1991. Distribution of repetitive DNA-sequences in
Eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids
Res. 19, 6823–6831.
Watt, M., Hugenholtz, P., White, R., Vinall, K., 2006. Numbers and locations of native
bacteria on field-grown wheat roots quantified by fluorescence in situ hybridization
(FISH). Environ. Microbiol. 8, 871–884.
Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J., 1991. 16S ribosomal DNA
amplification for phylogenetic study. J. Bacteriol. 173, 697–703.
Welham, S.J., Thompson, R., 1997. Likelihood ratio tests for fixed model terms using
residual maximum likelihood. J. R. Stat. Soc., B Meth. 59, 701–714.