C. Mulder and M. Raubuch conceived and designed the experiments. C. Mulder, M. Wouterse, M. Raubuch, and M. Rutgers performed the experiments. M. Raubuch, W. Roelofs, and M. Rutgers contributed materials/analysis tools. C. Mulder and W. Roelofs designed and analyzed the models. C. Mulder and M. Wouterse performed the calculations. C. Mulder wrote the paper.
The authors have declared that no competing interests exist.
The aim of the experiment was to determine if temporal variations of belowground activity reflect the influence of the Cry1Ab protein from transgenic maize on soil bacteria and, hence, on a regulatory change of the microbial community (ability to metabolize sources belonging to different chemical guilds) and/or a change in numerical abundance of their cells. Litter placement is known for its strong influence on the soil decomposer communities. The effects of the addition of crop residues on respiration and catabolic activities of the bacterial community were examined in microcosm experiments. Four cultivars of
There is a huge number of studies regarding the effects of exposure to transgenic maize pollen on nontarget insects. These concerns are due to the key role of pollinating insects in ecosystem functioning. Many results are highly debated and controversial. However, soil microbes as nontarget organisms are neglected. This is surprising, given the wide variation in agro-ecosystems that arises from several bacterial-driven processes (soil respiration, priming effect of organic matter, litter quality and quantity, and many others). Should the inclusion of bacteria in these nontarget organisms be regarded as artifactual? This seems unambiguous, as soil bacteria themselves are known to be the most crucial element of ecosystem functioning. Effects on the metabolism of soil bacteria grown in microcosms treated with either transgenic or conventional maize straw were studied in the laboratory. In contrast to the authors' expectations, empirical evidence is available to show different utilization of carbon compounds, especially of carbohydrates, according to the concentration of the
(A) Photo of
(B) Transgenic crop residues and soil sample after the addition of reaped maize straw shortly before starting the experiment (complete straw analysis available as
Differences occurred in the microbial respiration activity of soil samples with added transgenic straw in comparison with soil samples with conventional maize residues (
Daily Respiration of Soils Amended with Straw from Different Cultivars of Maize (μg CO2/g Dry Soil ± SD) Whose Cultivars Show Strong Differences (ANOVA:
Temporal variance in gross soil respiration of a stagnic luvisol after the addition of maize straw. Dotted lines for crop residues from conventional maize, continuous lines for transgenic maize. Untreated soils without any addition of straw showed initial respiration values of 12 μg CO2-C at d 0 (
Parallel with measurements of soil respiration, both the counting of colony-forming units (CFUs) and the measurement of metabolic fingerprints by BIOLOG were performed along a time gradient. At the beginning of the experiment, no significant difference was detectable between the 5 × 2 soil microcosms (average CFU equals 5 × 106 with
CFU growing curves in amended soils. Dotted lines for crop residues from conventional maize, continuous lines for transgenic maize. Transgenic maize straws show a much higher number of CFUs than conventional straws during the first week (ANOVA:
The utilization of 31 sole-carbon sources by bacterial communities of soils in the presence of increasing concentrations of the Cry1Ab protein was measured by a color development assay performed on BIOLOG microtitre plates (see
Correlations between WCD Values of BIOLOG Carbon Sources and the Concentration of the Cry1Ab Protein as Measured at the Beginning of the Experiment
Regardless of the elapsed days, the counts of a given color development (a WCD value that is specific for the respective carbon source) showed differences within each isoline in the catabolic activity of the soil bacteria during the entire experiment. Such differences in the potential carbon utilization are clearly related to the addition of transgenic material and describe a lower utilization of some substrates. Among the 31 × (31 – 1) = 930 possible matrix correlations between the EcoPlate substrates, 9.3% correlations were (slightly) responsive at α = 0.10 and 30.7% were highly significant at α = 0.05. There was no prevalence of either direct or indirect correlations between these 31 independent carbon sources, as 52% of the correlations were positive and 48% were negative.
Time was also modeled as canonical covariable (see [
PRCs of the summarized metabolic fingerprints of the two
Using the untreated control as an appropriate nesting specification (see
Catabolic Activity of Soil Bacteria (as Measured by the Change of the Dye from Colorless to Purple) on the BIOLOG Carbon Sources (Time as Fixed Effect, and Log-Transformed WCD Values as 31 Dependent Variables)
We may conclude that the possibly adaptive radiation of bulk soil bacteria in our microcosms shortly after the addition of
These possibilities need to be addressed in future analyses. Overall, however, the trend was a clearly different catabolic activity of the bacteria in all the microcosms treated with transgenic plant material in comparison with those of the control and of the soils amended with conventional crop residues.
In microcosm experiments we examined the effects of the addition of maize straw on properties such as respiration and catabolic activities of the bacterial community. We hypothesized that variations in the belowground activity would reflect the influence of transgenic plants, due to the effect of the Cry1Ab protein on the bulk soil bacteria, and hence on the ability of the microbial community to metabolize carbon sources. We combined microbiological techniques and mixed generalized linear models to facilitate a more comprehensive understanding of possible stress.
The performed growth models established a striking dichotomy between the soils amended with either conventional or transgenic straw. The concentrations of the Cry1Ab protein clearly influenced the composition of the microbial community. Evidence for increasing soil respiration during the second and third days of the treatment can be interpreted in relation to the presence of transgenic plant material.
Microbial uptake of sugars seemed to be affected by the presence of transgenic material (
Our data suggest that, in contrast to previous studies, the introduction of transgenic maize influences abundance, diversity, and ecosystem functioning of the bulk soil bacteria. Overall, the data provide support for the notion that a
There were short but robust differences in the soil microbial community after the addition of transgenic plant material.
These results may have important implications for the ecological risk-assessment of genetically modified organisms at the community level [
The technique of metabolic fingerprinting enabled the detection of rapid shifts in the functional diversity of the microbial community after the introduction of the Cry1Ab toxin in the environment. A resulting question for further research is therefore to determine genetically which components of the microbial community contribute to the increase of activity shortly after the treatment with transgenic material.
The soil was a stagnic luvisol containing 75% silt, 21% clay, and 4% sand from the Hebenshausen 37249 site (Meierbreite, 51°21′ N, 9°52′ E) of the Department of Soil Ecology and Plant Nutrition at the University of Kassel (Germany). Total soil carbon and total soil nitrogen were analyzed by dry combustion, and the percentage of organic carbon was calculated by subtracting the carbonate content from the total soil carbon. Soil pH was measured in deionized water with a glass electrode. The soil contained 1.45% organic carbon and 0.13% total nitrogen and had a pH of 6.4. Plant material was collected near Halle (Sachsen-Anhalt) at the field site of a research network of the German Federal Ministry for Education and Research. Residues of shoots (mixture of stems and leaves) of four field-grown cultivars of maize were collected at harvest at the end of the vegetation period in October 2002. Novelis (MON810, Monsanto,
Ecosystem properties such as land use and soil structure are known to control the ratio of CO2 evolution per O2 uptake in nonsterile soils [
The incubation experiments were done in climate chambers adjusted to constant 15 °C. Two grams of ground plant residues (dry weight) were added to 50 g of sieved (2 mm) fresh soil adjusted to 75% of the water-holding capacity in glass pots/glass jars, and mixed (some soil samples after treatment are shown in
Some soil samples before entering the 1.0-L pots after the addition of crop residues (reaped maize straw, see
Bacteria were extracted from soil immediately before each titration: 15 g of soil sample, based on dry weight, was blended with 150 ml sterile buffer (10 mM bis[2-Hydroxyethyl]imino-tris[hydroxymethyl]methane (BisTris, Sigma,
The colonies of viable cells were grown on 0.1TSA (Trypton Soya nutrient broth by Oxoid,
For the analysis of the CLPP, the inoculum-density independent approach was applied [
Analysis of variance (PROC ANOVA) and nested (and non-nested) generalized linear models (PROC MIXED, GLM) were used to assess the effect of the concentration of the Cry1Ab protein on the microbial respiration. The strength of a relationship between each WCD and Cry1Ab value was measured by partial correlation (PROC CORR) using the sums of squares and cross-products. The fitting of multilevel models (here, the glass pots according to their treatment) and growth models (a dilution series in 12 sections of four BIOLOG EcoPlates (three sections each) comprehending 31 individual carbon sources repeated 2×) required a flexible statistical approach. The PROC MIXED routine by SAS 9.1.3 has been chosen, as PROC MIXED was written by agricultural scientists seeking a mixed generalization of linear models that allows for both fixed and random effects [
Two structural specifications in the PROC MIXED statement are the covtest option (hypothesis tests for the variance and covariance components; see [
The statistics of the goodness-of-fit of multiple models with the same fixed effects (here, time) but different random effects (here, crop residues, if any) is assessed by the Akaike's information criterion (AIC) and by the much more stringent Sawa's Bayesian information criterion (BIC). The AIC test penalizes for adding parameters to the model, wherefore the model with the smallest AIC value is chosen [
The isolines of the transgenic cultivars Novelis (Event MON810) and Valmont (Event 176) and conventional cultivars Nobilis and Prelude are in roman numbers on the upper row. Valmont showed the highest contents of proteins, fats, and sugars. Using the control as appropriate nesting specification to fit the data on soil respiration to a multilevel generalized linear model (untreated or treated soils; if treated: conventional or transgenic straw; if transgenic: straw with either a low or a high Cry1Ab content), sugars became the most significant predictor for the rate of CO2 production (
(42 KB DOC)
(41 KB PDF)
We thank the anonymous referees for their many helpful suggestions. Leo Posthuma, Frans Kuenen, Anton M. Breure, and Marie-Elène Boivin are acknowledged for their insights. Appreciation is expressed to Kerstin Behr and Katja Roose for providing the crop residues and to Bjarne Munk Hansen for a photo of
Akaike's information criterion
Bayesian information criterion
colony-forming unit
community level physiological profiles
well-color development