Impairment of cellulose- and cellobiose-degrading soil Bacteria by two acidic herbicides


Correspondence: Steffen Kolb, Department of Ecological Microbiology, University of Bayreuth, Universitätsstr. 30, 95440 Bayreuth, Germany. Tel.: +49 0921 555620; fax: +49 0921 555799; e-mail:


Herbicides have the potential to impair the metabolism of soil microorganisms. The current study addressed the toxic effect of bentazon and 4-chloro-2-methylphenoxyacetic acid on aerobic and anaerobic Bacteria that are involved in cellulose and cellobiose degradation in an agricultural soil. Aerobic saccharide degradation was reduced at concentrations of herbicides above environmental values. Microbial processes (e.g. fermentations, ferric iron reduction) that were linked to anaerobic cellulose and cellobiose degradation were reduced in the presence of both herbicides at concentrations above and at those that occur in crop field soil. 16S rRNA gene transcript numbers of total Bacteria, and selected bacterial taxa (Clostridia [Group I], Planctomycetaceae, and two uncultivated taxa of Bacteroidetes) decreased more in anoxic than in oxic cellulose-supplemented soil microcosms in the presence of both herbicides. Collectively, the results suggested that the metabolism of anaerobic cellulose-degrading Bacteria was impaired by typical in situ herbicide concentrations, whereas in situ concentrations did not impair metabolism of aerobic cellulose- and cellobiose-degrading soil Bacteria.


Cellulose is metabolized by diverse aerobic and anaerobic, cellulolytic and saccharolytic microorganisms in soils (Falkowski et al., 2002; Lynd et al., 2002; Wei et al., 2009; Schellenberger et al., 2010). Increasing application of herbicides over the past decades in agriculture has resulted in accumulation of herbicide residues in soils that may affect microbial metabolism (Wainwright, 1978; Thorstensen et al., 2001; Chowdhury et al., 2008; Hiller et al., 2008). Herbicides can be degraded in soils (Müller et al., 2001; Gonzales et al., 2006), although, their degradation is slow compared with that of natural organic compounds (such as sugars or amino acids) and is primarily aerobic (Harrison et al., 1998; Knauber et al., 2000; Liu et al., 2010). Bentazon [3-isopropyl-1H2,1,3-benzothiadiazin-4(3H)-one-2,2-dioxide; pKa = 3.28] is a control agent of broadleaf weeds in agricultural crop plantations. It inhibits the photosynthetic electron flow in plants, and interacts with membranous proteins, which leads to an inhibition of ATPase (Hull & Cobb, 1998). Therefore, bentazon inhibits growth of pure cultures of various soil bacteria (e.g. Actinobacteria, rhizobia, cyanobacteria), and reduces dinitrogen fixation and nitrogen mineralization in soils (Cernakova et al., 1991; Galhano et al., 2009). MCPA (4-chloro-2-methylphenoxyacetic acid; pKa = 3.73) is also used as a control agent of broadleaf weeds, and acts as a plant growth hormone analog. MCPA enters the cytoplasm in the acidic form by diffusion, which causes a dissipation of the transmembrane proton-motive force (Cabral et al., 2003). The toxic effect of MCPA on cell growth has been observed with pure cultures of yeast, Pseudomonas putida and Vibrio fischeri (Ahtiainen et al., 2003; Cabral et al., 2003). Therefore, the toxic effects of these herbicides on cellulose-degrading soil Bacteria have been addressed in the current study.

Materials and methods

Soil from a wheat-planted agricultural cropland (Germany; 48°30.0′N, 11°20.7′E; sampled June 2009) was used (Table 1) to prepare soil microcosms. Cellulose- and cellobiose-supplemented soil microcosms were incubated at 15 °C in darkness. Two different experiments were set up. Microcosms with wet soil were supplemented with cellulose paper sheets. In the second experiment, soil was suspended in sterile water and supplemented with cellobiose. In each of the two experiments a set of replicates were incubated under oxic or anoxic conditions, and one set of experimental replicates was supplemented with bentazon and another set with MCPA. Microcosms without herbicides were used in both experiments as controls. Herbicide concentrations of 2.4 μmol gsoil DW−1 were used in cellulose-supplemented microcosms. Cellobiose-supplemented slurries received a ‘high’ (Bentazon, 8.5 μmol gsoil DW−1; MCPA, 3.01 μmol gsoil DW−1; Fig. 1) or a ‘low’ concentration (bentazon, 0.08 μmol gsoil DW−1; MCPA, 0.02 μmol gsoil DW−1; Supporting Information, Fig. S1). Low concentrations were assumed to be typical in herbicide-treated soils (Bentazon: 15.0 μg gsoil FW−1; MCPA: 2.8 μg gsoil FW−1; McGhee & Burns, 1995; Beulke et al., 2005; Baelum et al., 2006; Galhano et al., 2009). For cellulose-supplemented microcosms, 50 g of sieved wet soil (seven replicates) was mixed with crystalline herbicides and with cellulose sheets (Whatman, UK; > 98% cellulose; Munier-Lamy & Borde, 2000). Cellobiose-supplemented soil microcosms were prepared as duplicated slurries (250 μM cellobiose; Schellenberger et al., 2010). Microcosms were flushed with sterile air or N2 (Riessner Gase GmbH, Germany) to create oxic and anoxic conditions.

Figure 1.

Degradation of supplemental cellulose in soil microcosms in the presence of herbicides at 2.4 μmol gsoilDW−1 (i.e. Bentazon: 581 μg gsoilDW−1; MCPA: 485 μg gsoilDW−1). Tend values are the means of triplicates, error bars indicate standard deviations. Closed symbols, values from microcosms supplemented with Bentazon (black) or with MCPA (gray). Open symbols, microcosms without herbicide supplementation. Dotted lines, microcosms without sugar or herbicide supplementation. ● carbon dioxide; ▼ herbicide; ✚ cellulose; ▲ ferrous iron. Arrows, time points for calculation of inhibitory effect of supplemented herbicide (Tables S1 and S3).

Table 1. Physicochemical parameters of the agricultural soil used
  1. a

    According to FAO classification (Fuka et al., 2008).

  2. b

    Central Analytics, University of Bayreuth, Germany.

Soil typeaDystric cambisol
Textureb (clay : silt : sand) [%]27 : 33 : 40
Gravimetric water content [%]20.9
C contentb [%]1.23
N contentb [%]0.18
Total Feb [μmol gsoil DW−1]510 ± 14
Total Mgb [μmol gsoil DW−1]18.0 ± 0.6
inline imageb [μmol gsoil DW−1]0.2 ± 0.0
inline imageb [μmol gsoil DW−1]2.6 ± 0.1
inline imageb [μmol gsoil DW−1]0.4 ± 0.0

Molecular hydrogen, carbon dioxide, methane, pH, soluble sugars, organic acids, alcohols, herbicides, and ferrous iron were measured according to previously published protocols (Tamura et al., 1974; Daniel et al., 1990; Matthies et al., 1993; Küsel & Drake, 1995; Liu et al., 2010; Schellenberger et al., 2010). Cellulose-supplemented microcosms were incubated for 70 days and measured every 2 weeks. At each time point, one replicate was destroyed for measurement of cellulose weight loss (Munier-Lamy & Borde, 2000). Weight loss was converted into molar concentrations assuming that 1 mol of cellulose is equivalent to 1 mol of glucose. Cellobiose-supplemented microcosms were incubated for 1–2 days. Literature half-life times of herbicides (Bentazon: 42 days; MCPA: 24 days, Environmental Protection Agency, USA) were in same range or above. Thus, effective herbicide concentrations were probably stable and were not measured. Nucleic acids were purified from soil samples by a bead beating-based lysis procedure and phenol–chloroform extraction (Schellenberger et al., 2011). Pure RNA was obtained by DNase I (Fermentas GmbH, Germany) treatment of nucleic acid extracts (Schellenberger et al., 2011). RNA concentrations were quantified with the Quant-iT RiboGreen assay kit (Invitrogen, Germany). Quantification of 16S rRNA genes and transcripts was performed according to previously published qPCR protocols (Schellenberger et al., 2011). An assay-specific standard (100–108 transcripts per reaction) was included in every run. Transcript numbers were measured in three extracts of experimental replicates at the start and the end of the experiment in which wet soil was supplemented with cellulose sheets. Sample-specific inhibition and in vitro transcription efficiency were determined by quantification of spiked external RNA standard to each RNA extract and its quantification by a specific qPCR assay (Wieczorek et al., 2011).

Results and discussion

Cellulose and cellobiose were degraded under both oxic and anoxic conditions (Figs 1 and 2; Fig. S1). Products of cellulose hydrolysis (cellobiose or glucose) were not detected (≤ 0.5 μmol gsoil DW−1) suggesting an efficient assimilation of hydrolysis products. Small amounts of acetate, propionate, and butyrate accumulated in anoxic cellulose-supplemented microcosms (< 5 μmol gsoilDW−1), and ferrous iron formation was stimulated, i.e. ferric iron reducers were active (Fig. 1). Similar product profiles have been observed previously in other aerated soils (Küsel & Drake, 1995; Küsel et al., 2002). Hydrolysis of supplemental cellobiose led to a transient accumulation of glucose (Fig. 2; Fig. S2; Table S2) and could have been caused by β–glucosidases that were released by cellulolytic aerobes (Lynd et al., 2002) under the preceding oxic conditions. Traces of molecular hydrogen were detected in cellobiose-supplemented microcosms (Fig. 2; Fig. S1), and pH ranged from 4.7 to 6.2 (data not shown).

Figure 2.

Degradation of glucose and supplemental cellobiose in soil slurries under ‘high’ (Bentazon: 8.5 μmol gsoilDW−1 or 2.0 μg gsoilDW−1; MCPA, 3.01 μmol gsoilDW−1 or 0.6 μg gsoilDW−1) herbicide concentrations. Values are the means of duplicates, error bars indicate standard deviations. Closed symbols, microcosms supplemented with Bentazon (black) or with MCPA (gray). Open symbols, microcosms without herbicide supplementation. Dotted lines, microcosms without substrate or herbicide addition. Symbols: ● carbon dioxide; ■ molecular hydrogen; ✖ cellobiose; ♦ glucose. Arrows, time point for calculations of inhibitory effect of supplemented herbicide as presented in Tables S2 and S3.

Cellulose degradation was analysed only at high herbicide concentrations (Fig. 1) and revealed that both pesticides have the potential to impair cellulose degradation at oxic and anoxic conditions. The toxic effect of Bentazon and MCPA on cellobiose degradation under oxic conditions was only apparent at concentrations above values that are typical of crop field soils. At typical in situ herbicide concentrations, inhibition of aerobic cellobiose degradation was not apparent (Fig. 2; Table S3). Under anoxic conditions, Bentazon and MCPA impaired consumption of glucose in cellobiose-supplemented soil microcosms (Figs 1 and 2). Cellobiose consumption rates were not affected (Table S3). This toxic effect was observed at high and low herbicide concentrations (Figs 1 and 2; Fig. S1). Concentrations of formed organic acids (i.e. acetate, propionate, butyrate) were below the quantification limit (i.e. < 1.5 μmol gsoil DW−1 in total) (data not shown). The production of carbon dioxide and molecular hydrogen was decreased up to 85% and 100%, respectively, and ferrous iron production was negligible (Table S3). Thus, anaerobic cellulose-degradation was highly sensitive to the toxicity of both herbicides.

The findings on the toxic effects of the tested two herbicides agree with observations (1) that MCPA that was applied at the recommended dose did not affect either carbon dioxide production, or oxygen uptake or N-mineralization in an cropland soil and (2) that aerobic cellulose degradation was only slightly decreased even when MCPA was spread directly on cellulose sheets (Grossbard, 1971; Schröder, 1979). Nonetheless, reduction of nitrogen mineralization and soil respiration (i.e. carbon dioxide production) in the presence of Bentazon has been reported previously (Marsh et al., 1978; Cernakova et al., 1991; Piutti et al., 2003). Heterogeneous distribution of herbicides in field crops may lead to local maxima of herbicide concentration that exceed reported mean values (Marsh et al., 1978) but current and previous data suggest that the impact of MCPA and Bentazon on oxygen-dependent cellulose and cellobiose degradation is minimal under environmental concentrations.

16S rRNA gene transcript numbers of total soil Bacteria and five family-level taxa of Bacteria that have previously been identified as active members of the cellulolytic and saccharolytic community of the same soil (Schellenberger et al., 2010) were determined in soil samples of cellulose-supplemented microcosms using reverse transcriptase quantitative PCR (RTqPCR). In the oxic, cellulose-supplemented microcosms, fungal hyphae grow on the cellulose sheets, whereas in anoxic treatments, fungal hyphae were not observed (data not shown). Thus, it is very likely that fungi contributed to aerobic cellulose degradation. The metabolic response to Bentazon and MCPA of well known and novel, i.e. as yet uncultivated, taxa that have all been proven to contribute to cellulose and cellobiose degradation in the investigated soil (Schellenberger et al., 2010) was evaluated to reveal the taxa that may cause the reduced degradation rates under anoxic conditions. The specificity of the utilized RT qPCR assays has been demonstrated previously in the same soil (Schellenberger et al., 2011). In the presence of 2.4 μmol gsoil DW−1, Bentazon and MCPA, transcript numbers of total soil Bacteria and all analysed family-level taxa were lower in both oxic and anoxic microcosms at the end of the experiment compared with herbicide-free microcosms (Fig. 3; Table 2). Reports about a reduction of microbial growth in pure culture by both herbicides support these findings (Cernakova et al., 1991; Ahtiainen et al., 2003; Cabral et al., 2003; Galhano et al., 2009). Transcript numbers of Planctomycetaceae and uncultured ‘Sphingo’ (Bacteroidetes) were significantly lower under oxic conditions, whereas those of uncultured ‘Cellu’ (Bacteroidetes) and Clostridia of group I (Clostridiaceae; according to Collins et al., 1994) were significantly lower under anoxic conditions (Table 2). Most known anaerobic cellulolytic bacteria that have been isolated belong to Clostridia group III (Collins et al., 1994). Clostridiaceae assimilated carbon from supplemented 13C-enriched cellulose and were metabolically stimulated under anoxic conditions in the same soil (Schellenberger et al., 2010, 2011). Development of primers that exclusively target these organisms failed. Thus, it cannot be excluded that the metabolism of not only Clostridia of group I but also group III was inhibited by herbicides.

Figure 3.

16S rRNA gene transcript numbers of Bacteria in cellulose-supplemented microcosms in the presence and absence of herbicides: (a) oxic conditions, (b) anoxic conditions. White bar, value at the start of the experiment, i.e. the same value for the oxic and anoxic experiment. Black bars, no herbicide addition. Light gray bar, Bentazon-treated microcosms; dark gray, MCPA-treated microcosms. The experiment lasted 70 days. Small letters indicates statistical differences of values (P ≤ 0.05; t-test).

Table 2. 16S rRNA gene transcript numbers of family-level taxa that are linked to cellulose degradation (based on results of Schellenberger et al., 2010)
TaxonaTranscript number [ngRNA−1]b
0 days70 days
  1. a

    Name and affiliation of family-level taxa according to Schellenberger et al. (2010). A family was based on a minimal 16S rRNA gene similarity of 87% (Yarza et al., 2008).

  2. b

    Values are the mean of three replicates (with standard deviation).

  3. c

    Herbicide concentration was 2.4 μmol gsoil DW−1.

  4. d

    Numbers in herbicide-supplemented microcosms, marked with an asterisk are significantly different from herbicide-free controls (t-test; S. Schellenberger, H.L. Drake & S. Kolb 2011; submitted).

Micrococcaceae and Cellulomonadaceae(4.6 ± 1.5) × 105(3.4 ± 0.8) × 105(1.7 ± 0.7) × 105(1.9 ± 0.9) × 105
Group I Clostridiaceae(1.3 ± 0.2) × 104(8.3 ± 1.8) × 103(6.5 ± 3.9) × 103(2.0 ± 1.4) × 104
Planctomycetaceae(3.8 ± 0.3) × 105(2.3 ± 0.6) × 105(9.1 ± 2.6) × 104d(2.8 ± 0.9) × 104d
‘Cellu’(5.0 ± 0.7) × 102(4.3 ± 1.3) × 102(4.3 ± 1.2) × 102(5.2 ± 3.2) × 102
‘Sphingo’(6.5 ± 0.4) × 105(5.1 ± 1.4) × 105(1.4 ± 0.7) × 104d(2.0 ± 0.8) × 104d
Micrococcaceae and Cellulomonadaceae(4.6 ± 1.5) × 105(2.5 ± 0.9) × 105(2.7 ± 1.3) × 105(1.4 ± 0.6) × 105
Group I Clostridiaceae(1.3 ± 0.2) × 104(4.3 ± 1.2) × 104(1.1 ± 0.4) × 104d(7.0 ± 2.0) × 103d
Planctomycetaceae(3.8 ± 0.3) × 105(7.8 ± 5.9) × 104(9.8 ± 3.9) × 104(4.2 ± 0.1) × 104
‘Cellu’(5.0 ± 0.7) × 102(9.7 ± 5.3) × 103(3.0 ± 0.6) × 101d(6.5 ± 1.6) × 102d
‘Sphingo’(6.5 ± 0.4) × 105(2.4 ± 2.0) × 104(1.1 ± 0.5) × 105(2.7 ± 1.1) × 104

Collectively, the results of the current study suggest that Bentazon and MCPA at environmental relevant concentrations likely do not impair aerobic cellobiose degradation, whereas anaerobic cellobiose degradation and the Bacteria involved may be negatively impacted. However, both aerobic cellulose and cellobiose degradation were impaired by higher herbicide concentrations. The analysed bacterial taxa were also metabolically impaired under oxic conditions, which could suggest that they consumed less cellulose; however, the visibly present fungi compensated for this loss of activity in the presence of pesticides. Agricultural soil is normally well aerated and has only small anoxic microzones. Impairment of anaerobic processes in such soils is probably of minor importance for the overall degradation of cellulose, as this process is mainly aerobic. However, when such soils become water-saturated due to rain, the observed toxic effect of Bentazon and MCPA on anaerobes may be of importance for cellulose degradation.


The authors thank Christina Hirsch for technical assistance, M. Schloter, and S. Schulz (Technical University Munich) for providing soil, and the Deutsche Forschungsgemeinschaft (Priority Program 1315), and the University of Bayreuth for funding the study.