Biology and Fertility of Soils

, Volume 48, Issue 3, pp 315–324

Effects of conventionally bred and Bacillus thuringiensis (Bt) maize varieties on soil microbial biomass and activity

Authors

    • Research Institute of Organic Agriculture
  • Monika Messmer
    • Research Institute of Organic Agriculture
  • Bruno Nietlispach
    • Research Institute of Organic Agriculture
  • Valentina Infante
    • Research Institute of Organic Agriculture
    • Facultad de Agronomía e Ingeniería ForestalPontificia Universidad Católica de Chile
  • Paul Mäder
    • Research Institute of Organic Agriculture
Original Paper

DOI: 10.1007/s00374-011-0625-6

Cite this article as:
Fließbach, A., Messmer, M., Nietlispach, B. et al. Biol Fertil Soils (2012) 48: 315. doi:10.1007/s00374-011-0625-6

Abstract

Genetically modified (GM) maize containing genes from the soil bacterium Bacillus thuringiensis (Bt) was cultivated on 29% of the total maize production area worldwide in 2009. Most studies to date compare Bt-maize varieties with their near isogenic lines; however, there is little information on the variability of conventional maize breeding lines and how the effects of Bt varieties are ranked within. In our study on the potential risks of Bt-maize varieties, we analyzed tissue quality and compared the effects of ten conventional and GM maize varieties on soil microbiological properties in a replicated climate chamber experiment. All maize varieties were cultivated twice in the same soil microcosm. Shoot yields and soluble C in leaf tissue of Bt varieties were higher than the ones of non-Bt. Soil dehydrogenase activity was reduced by 5% under Bt varieties compared to non-Bt, while most of the other soil microbial properties (soil microbial biomass, basal respiration) showed no significant differences between Bt and non-Bt varieties. The leaves and roots of one Bt variety were decomposed to a greater extent than the ones of its near isogenic line; the conventional breeding lines also showed higher values. Changes in crop and soil parameters were found when comparing the first and the second crops, but the effects of repeated cropping were the same for all tested varieties. For the studied parameters, the variation among non-Bt-maize varieties was similar to the difference between Bt and non-Bt varieties.

Keywords

Bt-maizeSoil fertilityCry1AbGMOMicrobial biomassSoil respirationDehydrogenaseMycorrhiza

Introduction

Maize production worldwide was 817,000,000 t in 2009 on 160,000,000 ha (FAO 2011). Forty-two million hectares (29% of the total maize production area) was cultivated with genetically modified (GM) maize in 2009 after a rapid increase from 2,000,000 ha (1.4%) in 1998. In the US, the most abundant GM crop maize was cultivated on 85% of the total area under maize (GMO Compass 2011). The field area under GM crops in Europe peaked with 110,000 ha in 2007 and shrunk to 91,000 ha in 2010 with 85% of the area in Spain. The decrease in Europe over the last years was due to public skepticism and national cultivation bans for the maize variety MON810 in France and Germany and all GM crops in Switzerland.

The most commonly used genetic modifications of maize are based on two traits that are expressed in today’s maize varieties: herbicide tolerance that allows for using broad-spectrum herbicides like glyphosate or glufosinate for weed control and insect resistance that comes from the soil bacterium Bacillus thuringiensis by inserting the bacterial genes responsible for the production of larvicidal Cry proteins into the genome of the plant. The Cry protein from B. thuringiensis (protoxin) becomes effective when taken up and transformed to the active δ endotoxin in the gut of the target organism. Most of the Bt crops, however, express the active δ endotoxins directly and constitutively in all plant parts (Sanvido et al. 2006) and some of them also release Cry proteins into the rhizosphere of the living plant (Saxena et al. 1999, 2004). The Cry1A group targets lepidopteran larvae and Cry3 proteins target beetle larvae. The large-scale cultivation of Cry1-expressing crops has triggered the expectation that resistance to Bt crops would develop rapidly. For Cry1Ab, a case in South Africa is documented where a stem borer (Busseola fusca) has developed resistance towards Bt-maize (Tabashnik et al. 2009, 2003; Van Rensburg 2007) and a lepidopteran pest (Helicoverpa zea) in cotton systems has also developed resistance to Cry1Ac proteins (Tabashnik et al. 2008). Strategies to avoid the development of resistance are to use varieties with several stacked Bt genes, spraying insecticides, creating refuge areas with susceptible varieties for the pathogens to develop without the selection pressure of the Bt proteins, and post-harvest control measures (Bartsch et al. 2010).

Most of the risk assessments with respect to GM crops were directed towards effects on nontarget invertebrates (Andow and Zwahlen 2006; Clark et al. 2005; Marvier et al. 2007). Nematodes as nontarget organisms have often been analyzed as risk indicators, but results showed small and contradictory effects (Griffiths et al. 2005; Lang et al. 2005; Manachini and Lozzia 2002; Saxena and Stotzky 2001a), as stated for most members of the soil fauna not directly targeted by the Bt toxins (Donegan et al. 1995; Escher et al. 2000; Saxena and Stotzky 2001a; Sims and Martin 1997; Stotzky 2004; Yu et al. 1997; Zwahlen et al. 2003).

Soil microorganisms as key players in decomposition and nutrient cycling processes in soils are of essential importance for natural and agricultural soil ecosystems. A recent review on the effects of the introduction of GM plants into agricultural ecosystems states little even though sometimes significant effects on microbial communities, while most of these effects were related to other factors than the genetic modification (Icoz and Stotzky 2008). Root exudates of Bt-maize (event 176) reduced presymbiotic hyphal growth of Glomus mossae compared to Bt 11 and its near isogenic line (Turrini et al. 2005). The rhizosphere microbial community of the same maize varieties was significantly changed and Bt 176 roots showed a lower level of mycorrhizal colonization (Castaldini et al. 2005) even though in Bt 176 only the green plant parts are expressing Cry proteins.

Differences in decomposition of the Bt crop residues compared to its near isogenic line were often ascribed to unintended changes in the plant tissue (e.g., different concentrations of lignin, carbohydrates, or soluble compounds), but results on concentrations of these compounds were contradictory (Escher et al. 2000; Flores et al. 2005; Lang et al. 2005; Lehman et al. 2008; Saxena and Stotzky 2001b). Undecomposed Bt-maize residues, however, may harbor up to 12% of their initial Cry1Ab concentration (Baumgarte and Tebbe 2005). When the toxin is bound to organic soil particles or clay, it may become protected against degradation and retains its insecticidal activity (Kravchenko et al. 2009; Tapp and Stotzky 1995, 1998). Poerschmann et al. (2005) claimed that higher lignin contents were mainly to be found in stems of Bt varieties and that little attention has been paid to investigating the roots of Bt crops.

Most of the studies on the environmental impact of GM crops compare the results with the so-called near isogenic breeding line (Griffiths et al. 2007). Broer and Schiemann (2009) claim that biological risk assessment needs to include the whole variability of a crop species or at least a broader spectrum of varieties in order to judge upon GM crop effects. Compared to the bulk of research on GM crops, little information is available on the variable impact of conventionally bred varieties to the environment. The research presented here shows maize shoot and root tissue analyses of ten varieties cultivated in a climate chamber, out of which two were GM. The varieties were grown twice in the same soil to evaluate repeated cropping effects of GM and conventionally bred maize varieties on soil microbial biomass and microbial activities.

Material and methods

Loessial soil was taken in June 2007 from the surroundings of the DOK long-term field trial (Therwil, CH; 7°33′ E, 47°30′ N) (Mäder et al. 2002). The soil had a pH(H2O) of 6.6, organic carbon (Corg) was 19.1 mg g−1 and total soil nitrogen (Ntotal) was 2.2 mg g−1, and particle size classes showed 23% clay, 66% silt, and 9% sand. Soil was taken from the 0- to 20-cm soil layer (plough layer) with a 3-cm diameter auger and combined to one bulk sample that was sieved twice for homogenization (8 mm mesh), and moisture was adjusted to 40–50% maximum water-holding capacity. The soil was filled into containers (0.352 × 0.252 × 0.320 m) by weight (24 kg dry matter per container) and was then compacted slightly to achieve a filling level of about 25 cm.

Two maize crops were grown one after another in the same soil containers. We used seven maize varieties used in Swiss agriculture that were tested in a series of field trials (Dierauer and Böhler 2008). Our study was restricted to two older Bt-maize varieties: Bt-NK4640 (Bt 11) together with the near isogenic line NK4640 and Bt-Max88 (Bt 176) without an appropriate isogenic line.

A climate chamber (2.80 × 4.70 m) equipped with ten sodium vapor lamps and ten mercury vapor lamps in 2.10 m height and a temperature control was used for cultivating the maize plants. Photosynthetic photon flux density at 50 cm height averaged at 654 ± 58 and 717 ± 85 μmol m−2 s−1 at 90 cm height. Air temperature averaged at 17.2°C with mean daily minimum air temperature of 12.8°C at night and mean daily maximum temperature of 19.2°C. Air humidity reached 100% at night (8 h) and varied between 41% and 72% during the day (16 h).

All seeds were washed with 0.1% Tween 20 for 20 min, followed by a treatment with 0.1% NaClO for 10 min to remove seed treatments and pathogens from the seeds. Seeds were then rinsed with water for 1 h and spread on moist filter paper for germination. As soon as a suitable number of seeds were germinating, they were carefully planted in a container. Three replicate containers were randomly distributed in the climate chamber. Soils were irrigated automatically with demineralized water. On a weekly basis, each container was fertilized with 200 ml of a nutrient solution containing 8.14 g NH4NO3, 13.91 g KH2PO4, and 10.63 g MgSO4·7H2O L−1. The plants were grown until tassel emergence and then shoots, roots, and soil subsamples were collected.

The aboveground and belowground parts were harvested separately and dry matter yields were determined. For harvesting the roots, the whole soil was passed through a sieve of 8 mm mesh size and roots were handpicked and washed before they were allowed to dry superficially on filter paper. Subsamples were frozen at −80°C for analysis of the Cry1Ab protein and at −20°C for decomposition studies. Subsamples of the fresh fine roots were stored in ethanol (70%) for analysis of root colonization with arbuscular mycorrhizal fungi (AMF), and the remainder was dried at 60°C for 24 h for all other plant analyses. Subsamples (500 g dry matter) of soils were collected from each container for soil analysis. Subsamples were sieved (2 mm mesh) and the water content was readjusted to 40–50% of its maximum water-holding capacity. Soils were stored at 3°C in a cooling room until they were analyzed.

Plant analyses

Maize aboveground and belowground parts were analyzed for Cry1Ab proteins by enzyme-linked immunosorbent assay. Approximately 500 mg of frozen plant material was homogenized with 4.5 ml phosphate buffered saline Tween-20 (PBST) (BioReba, Switzerland) buffer in homogenization bags for 5 min. An aliquot of the same plant material was used for dry matter determination. After centrifugation for 2 min at 82×g, appropriate dilutions of the supernatant were filled into wells of an antibody-coated 96-well microtiter plate (Agdia, MPP 06200) together with a Cry1Ab standard kindly provided by Michael Sander (ETH, Zürich). The test was performed according to the guidelines provided by the manufacturer (https://orders.agdia.com/). In Bt-NK4640, Cry1Ab concentration averaged at 12.3 μg g−1 in the plant leaves and at 11.9 μg g−1 in the roots. In Max88, the leaves contained 11.9 μg g−1, but in the roots, Cry1Ab was not detectable. Neither in the near isogenic line NK4640 nor in the conventional breeding lines Cry1Ab was detectable.

Total C and N in leaf and root samples were analyzed with a CHN analyzer (Leco, CHN 1000). Hot water-soluble C (CHWS) and N (NHWS) was analyzed after autoclaving 500 mg dried and ground shoot and root material at 121°C for 15 min with 100 ml demineralized water (Leinweber et al. 1995). After cooling and readjusting the water volume, the filtered extract was analyzed for soluble organic C (TOC) by infrared spectrometry after combustion at 850°C (DIMA-TOC 100, Dimatec, Essen, Germany). Soluble N (TNb) was subsequently measured in the same sample by chemoluminescence (TNb, Dimatec, Essen, Germany).

Acid detergent fibers (Adf) and acid detergent lignin (Adl) in maize plant material were analyzed in a semiautomated fiber analyzer (Fibertec 2010, Foss, DK). The plant material was heated together with 100 ml of a 0.484 M H2SO4 solution with 20 g L−1 hexadecyl trimethyl ammonium bromide and three drops octanol per sample. After 60 min of simmering, the residue was washed with hot water and the sample was defattened with 25 ml acetone and then dried for 1 h at 130°C. The remainder was weighed as it represents the Adf fraction and was then mixed with 25 ml 11.82 M H2SO4 under repeated stirring for 3 h at room temperature. Then the remainder was washed repeatedly with demineralized water and dried at 130°C. This fraction represents the Adl fraction of the plant material.

For the assessment of root colonization with AMF, a modified method of Phillips and Hayman (1970) was used. Fine roots were cleared in 2.5% KOH, and fungal structures stained with 0.05% Trypan blue. The percentage of root segments colonized by AMF was assessed using a dissecting microscope (Leica M205 C, Leica Microsystems (Suisse) SA, 1020 Renens, Switzerland) at up to 100-fold magnification. One hundred and fifty intersections were counted per sample according to the gridline intersect method modified according to Giovannetti and Mosse (1980).

Soil analyses

Soil samples from each replicate container obtained at plant harvest were analyzed for pH, dry matter, soil microbial biomass C (Cmic) and N (Nmic), dehydrogenase activity (DHA), basal respiration, and hot water-extractable C and N. A study on the decomposition of shoot and root material was also performed for selected samples.

Dry matter of soil samples was analyzed gravimetrically after drying of subsamples at 105°C for 24 h. Soil pH was measured in a suspension with demineralized water (1:5 w/v) by the use of a pH electrode after standing overnight. Total C and N in dried and finely ground soil samples were analyzed with a CHN analyzer (Leco, CHN 1000). Hot water-extractable C and N were measured after autoclaving a soil suspension of 10 g soil in 100 ml demineralized H2O for 15 min at 121°C. The filtered extracts were analyzed for soluble C and N in a TOC-TNb analyzer as described above.

Soil microbial biomass C (Cmic) and N (Nmic) were estimated by chloroform fumigation–extraction according to Vance et al. (1987) and Brookes et al. (1985). Chloroform fumigation–extraction was done on 20 g (dry matter) subsamples that were extracted with 80 ml of a 0.5-M K2SO4 solution. The filtered extracts were analyzed for soluble C and N in a TOC-TNb analyzer as described above. Soil Cmic and Nmic were determined according to the following formulas:
$$ {{\text{C}}_{\text{mic}}} = {\text{EC}}/{k_{\text{EC}}} $$
where EC = (TOC in fumigated samples − TOC in unfumigated control samples) and kEC = 0.45 (Joergensen 1996).
$$ {{\text{N}}_{\text{mic}}} = {\text{EN}}/{k_{\text{EN}}} $$
where EN = (Nt in fumigated samples − Nt in unfumigated control samples) and kEN = 0.54 (Joergensen and Mueller 1996).

Soil basal respiration was measured in preincubated (7 days at 22°C) samples as carbon dioxide (CO2) evolved over a period of 72 h. Soil samples (20 g dry matter) were weighed into perforated centrifuge tubes and placed into a screw bottle (Schott, 250 ml). After another 24-h preincubation period in the bottle, the actual measurement started by adding exactly 20 ml of 0.025 M NaOH. After 72 h, the soil was taken out of the bottle and the alkali was titrated with 0.025 M HCl after precipitation of absorbed CO2 by excess BaCl2. The measurement was done according to the reference methods of the Swiss agricultural research stations (FAL et al. 1996). The metabolic quotient for CO2 (qCO2) was calculated from basal respiration rates divided by the amount of Cmic in the respective sample (Anderson and Domsch 1993).

DHA was measured according to Tabatabai (1982) in 5 g soil samples that were incubated at 30°C for 24 h in the presence of an alternative electron acceptor (triphenyl tetrazolium chloride). The red-colored product (triphenylformazan) was extracted with acetone and quantified in a spectrophotometer at 546 nm.

For the decomposition of maize plant material, frozen leaves and roots were mixed with the same soil, where the plant was previously grown, and the CO2 evolution from this soil was used as an indicator of decomposition. Before adding maize plant material, soil CO2 evolution of 20 g subsamples was assessed at 20°C for 7 days in order to break down any easily available organic C sources. Maize leaf and root material corresponding to 160 mg dry matter both stored at −20°C were carefully mixed with the soil in a glass beaker. The mixture was filled into perforated centrifuge tubes and placed into a screw bottle (Schott, 250 ml) in the presence of NaOH at appropriate concentration to capture the evolved CO2. The NaOH was changed at fixed intervals and was titrated with appropriate HCl solutions after precipitation of carbonates with excess BaCl2. Water loss in the incubated soil samples was regularly readjusted with demineralized water. Taking into account the C content of the plant material, the percentage of the decomposition was calculated, which, however, may also include some soil-derived C that was mineralized due to priming effects.

Statistics

The experiment was run under controlled conditions in a climate chamber and three replicate containers were prepared for each treatment. Single-container values for each of the parameters were subject to statistical evaluation (JMP 2008) with the variety and the experiment as factors and the replicate blocks as random effects of the full factorial model. Decomposition rates were determined only with plants and soils of the second crop. In case of significant model effects, a Tukey post hoc test was applied to check for differences among varieties. The overall effect of Bt-maize vs. non-Bt maize varieties was evaluated by contrasts added to the factorial effect.

Results

Maize yield and quality

Total biomass (shoot and root) yield in the first maize crop was 43% higher (p = 0.05) than in the second, but no significant differences between single varieties were found. Nonetheless, the contrast between Bt and non-Bt varieties indicated 13% higher total biomass yield for Bt (p = 0.020). The contrast between aboveground biomasses of Bt and non-Bt showed 14% higher shoot yield for Bt varieties (p = 0.017). Differences in either shoot or root yield of single varieties were never significant (Table 1). Likewise, the shoot-to-root ratio was not significantly affected. The average C content of the shoot was 405 mg g−1 dry matter and the one of the root averaged at 400 mg g−1 dry matter without significant differences between single varieties, but Bt varieties showed 1% higher C content than non-Bt varieties (p = 0.013) as indicated by the contrast of these two groups(data not shown).
Table 1

Mean values of shoot, root, and biomass yield of the tested varieties and the two cropping cycles

Level

Shoot yielda [g pot−1]

Root yield [g pot−1]

Biomass yielda [g pot−1]

Bt-Max88

69.0

12.3

81.3

Bt-NK4640

59.3

12.3

71.6

NK4640

59.6

13.0

72.6

5133 Eco

56.7

10.7

67.4

Amadeo

51.5

10.0

61.5

Ciclixx

61.4

12.8

74.1

Coxximo

59.1

11.1

70.2

Gavott

57.1

13.1

70.2

PR39G12

51.2

10.4

61.7

Ronaldinio

53.3

11.1

64.4

1st crop

68.0 A

10.1 B

78.1 A

2nd crop

47.0 B

13.3 A

60.7 B

Effect test from 2-way ANOVA [probability > F]

 Variety

n.s.

n.s.

n.s.

 Cropping cycle

<0.0001

<0.0001

<0.0001

 Variety × cropping cycle

n.s.

n.s.

n.s.

In case of significant model effects, different letters in a column indicate significant differences according to a post hoc Tukey HSD test (n = 3)

n.s. not significant

aSignificant contrast between Bt and non-Bt varieties

Soluble C in the shoot material averaged at 29% of the total C content. In the two Bt varieties, it was significantly higher than in five other varieties, but Bt-NK4640 was not different from its near isogenic line NK4640. Bt varieties had 17% higher hot water-soluble C content, while they had 18% lower soluble N content than the non-Bt varieties (p < 0.001). Soluble N showed no difference between Bt-NK4640 and its near isogenic line and was negatively correlated to the soluble C content (r = −0.7095, p < 0.0001; Table 2).
Table 2

Mean values of soluble C and N in shoot and root material of tested maize varieties and the two cropping cycles

Level

Soluble C [mg g−1 d.m.]

Soluble N [mg g−1 d.m.]

Shoota

Root

Shoota

Roota

Bt-Max88

132 a

62 ab

10.7 bc

10.3 c

Bt-NK4640

128 ab

66 ab

9.3 cd

10.6 c

NK4640

121 abc

70 a

8.4 d

10.5 c

5133 Eco

106 cd

58 ab

12.8 a

10.5 c

Amadeo

121 abc

67 ab

12.3 ab

13.8 a

Ciclixx

99 d

56 b

13.7 a

10.9 c

Coxximo

104 cd

66 ab

12.2 ab

11.9 bc

Gavott

128 ab

62 ab

12.3 ab

11.3 c

PR39G12

105 cd

65 ab

13.7 a

13.4 ab

Ronaldinio

110 bcd

60 ab

12.2 ab

11.6 c

1st crop

115

63

11.7

11.4

2nd crop

115

63

11.8

11.5

Effect test from 2-way ANOVA [probability > F]

 Variety

<0.0001

0.0173

<0.0001

<0.0001

 Cropping cycle

n.s

n.s.

n.s.

n.s.

 Variety × cropping cycle

n.s.

n.s.

n.s.

n.s.

Values are given in milligrams per gram oven d.m. of the plant material. In case of significant model effects, different letters in a column indicate significant differences according to a post hoc Tukey HSD test (p = 0.05; n = 3)

d.m. dry matter, n.s. not significant

aSignificant contrast between Bt and non-Bt

While the soluble C in shoots indicated markedly higher values for the Bt varieties, there was no significant contrast in the soluble C content of the roots. Amadeo and PR39G12 contained relatively high amounts of soluble N compared to most of the other varieties and the contrast showed 11% lower soluble N in the roots of Bt vs. non-Bt varieties (Table 2).

The fiber content of the plant material varied between 289 and 364 mg g−1 dry matter in the aboveground part and between 389 and 434 mg g−1 in the belowground part of the maize varieties. Bt-Max88 had a lower fiber content compared to the two other varieties in the study. No differences were found with the fiber content of the roots. Highest lignin values were found for Gavott and the lowest ones for NK4640. Lignin values between 44 and 54 mg g−1 in the roots of the maize varieties were not significantly different (Table 3). The colonization of roots with arbuscular mycorrhiza varied between 19% and 39%, but no differences between varieties were obtained (Table 4).
Table 3

Mean values of Adf and Adl content of shoot and root material

Level

Adf [mg g−1 d.m.]

Adl [mg g−1 d.m.]

Shoota

Root

Shoot

Root

Bt-Max88

289 b

412 a

32.3 abc

52.3 a

Bt-NK4640

307 ab

418 a

31.1 abc

46.7 a

NK4640

302 ab

411 a

26.1 c

43.8 a

5133 Eco

330 ab

406 a

28.6 abc

50.2 a

Amadeo

305 ab

390 a

27.0 bc

46.9 a

Ciclixx

345 a

434 a

32.0 abc

48.1 a

Coxximo

326 ab

402 a

31.0 abc

53.5 a

Gavott

311 ab

423 a

33.2 a

49.4 a

PR39G12

339 a

401 a

32.5 ab

47.2 a

Ronaldinio

320 ab

423 a

27.9 abc

50.3 a

1st crop

315

415

30.5

48.4

2nd crop

320

407

29.8

49.4

Effect test from 2-way ANOVA [probability > F]

 Variety

0.0111

n.s.

0.0034

n.s.

 Cropping cycle

n.s.

n.s.

n.s.

n.s.

 Variety × cropping cycle

n.s.

n.s.

n.s.

n.s.

Values are given in milligrams per gram oven d.m. of the plant material. Different letters in a column indicate significant differences according to a post hoc Tukey HSD test (p = 0.05; n = 3)

Adf acid detergent fiber, Adl acid detergent lignin, d.m. dry matter, n.s. not significant

aSignificant contrast between Bt and non-Bt varieties

Table 4

Root colonization with AMF after the second maize crop

Level

Colonization of roots with AMF [%]

Bt-Max88

22 a

Bt-NK4640

33 a

NK4640

29 a

5133 Eco

31 a

Amadeo

20 a

Ciclixx

31 a

Coxximo

20 a

Gavott

29 a

PR39G12

21 a

Ronaldinio

39 a

Different letters indicate significant differences according to a post hoc Tukey HSD test (p = 0.05; n = 3)

AMF arbuscular mycorrhizal fungi

Soil chemical parameters

The soils were analyzed in the beginning of an experimental series. Soil Corg in soils was 15.65 mg g−1 in the beginning and 18.21 mg g−1 at the end the first cropping period, and then no further change was observed. Ntotal with a value of 2.2 mg g−1 in the beginning decreased to 2.0 mg g−1 after the first maize crop and increased again to 2.1 mg g−1 at the end of the second. Hot water-extractable soil C analyzed after the first maize crop accounted for 905 μg g−1 soil (5.0% of Corg). This value increased significantly to 1,002 μg g−1 soil after the second maize crop, which corresponds to 5.5% of the soil organic C. Hot water-extractable N from soils after the first maize crop was 161 μg g−1 soil dry matter, corresponding to 8.2% of the Ntotal. The increase in hot water-extractable N at the end of the second crop was not significant. The extractable part of total N tended to decrease with increasing Ntotal (r2 = 0.21, p < 0.01). None of these soil parameters analyzed showed a significant effect of the varieties (data not shown).

Soil microbiological parameters

Microbial biomass C (Cmic) in the soil used was 397 μg g−1 soil dry matter at the start of the experiment. During the first maize cultivation, Cmic decreased by 9% to 363 μg g−1 soil dry matter and decreased further by 21% to reach an average of 328 μg g−1 soil dry matter at the end of the second cropping period (p < 0.0001). No significant effects of the maize varieties were found (Fig. 1a). The decrease in soil microbial N (Nmic) was 31% during the first cultivation and did not decrease further during the second cultivation. Soil microbial biomass N was also not affected by the different maize varieties (Fig. 1b).
https://static-content.springer.com/image/art%3A10.1007%2Fs00374-011-0625-6/MediaObjects/374_2011_625_Fig1_HTML.gif
Fig. 1

Soil microbiological properties before and after repeated cultivation of ten maize varieties. Mean values and standard error bars of three replicates are shown for a soil microbial biomass C (Cmic), b soil microbial biomass N (Nmic), and c DHA. The horizontal bar in c indicates differences of the predefined contrast between Bt and non-Bt varieties (p < 0.05)

The ratio of Cmic and Nmic showed no significant effects of the varieties, but the overall mean increased from a starting value of 7.1 to 9.5 after the first maize crop and decreased again to 8.5 after the second (p = 0.0139). Soil microbial biomass C as percentage of Corg, which was 2.1 in the beginning, decreased to 2.0 (−4.33%) after the first maize crop and to 1.8 (−9.55%) after the second (p < 0.0001).

DHA, as a microbial activity parameter of the soil, indicated no significant changes over the whole time range (Fig. 1c). While soils under single varieties were not significantly different from each other with respect to DHA, the contrast between Bt and non-Bt varieties showed 5% lower values for Bt (p = 0.046).

Basal respiration increased by 71% (p < 0.0001) from the end of the first crop to the end of the second. No significant effect of maize varieties on basal respiration was found. The metabolic quotient for CO2 (qCO2) increased by 87% between the end of the first and the end of the second crop (p < 0.0001). The maize varieties produced no significant effect in qCO2 (Table 5).
Table 5

Mean values of soil basal respiration and the metabolic quotient (qCO2) after cultivation of the varieties

Level

Basal respiration [μg CO2–C g−1 soil h−1]

qCO2 [μg CO2–C mg−1 Cmic h−1]

Bt-Max88

0.38

1.21

Bt-NK4640

0.32

1.05

NK4640

0.41

0.98

5133 Eco

0.35

1.16

Amadeo

0.41

0.90

Ciclixx

0.37

1.22

Coxximo

0.36

1.01

Gavott

0.33

1.03

PR39G12

0.31

0.89

Ronaldinio

0.36

1.09

1st crop

0.26 B

0.73 B

2nd crop

0.45 A

1.37 A

Effect test from 2-way ANOVA [probability > F]

 Variety

n.s.

n.s.

 Cropping cycle

<0.0001

<0.0001

 Variety × cropping cycle

n.s.

n.s.

In case of significant model effects, different letters in a column indicate significant differences according to a post hoc Tukey HSD test (p = 0.05; n = 3)

n.s. not significant

Decomposition of plant material can also be followed by trapping CO2 from the decomposition process. Carbon dioxide evolution from soils without added plant material was higher for soils cultivated with Gavott, Max88, and Bt-NK4640 than for those with Ronaldinio and NK4640 (p < 0.0001). After adding shoot and root material from maize varieties to the same soil they grew in before, the mineralization rates in the first 2 days were between 11.4% and 14.5% of the applied leaf material, while roots were mineralized only by 3.4% to 5.2% in the same time span (Fig. 2). The initial flush of CO2 from Gavott and Bt-NK4640 leaves were higher than the ones of NK4640 and Max88 (p < 0.0001), while only roots of Bt-NK4640 produced a higher initial flush than NK4640 and Max88 (p < 0.0001). Endpoints of mineralization after 24 days were 21% higher for Bt-NK4640 leaves compared to its near isogenic line (Table 6). Root mineralization of Bt-NK4640 was 24% higher than the one of its near isogenic line. Shoot and root material of Gavott were also decomposed to a larger extent. Differences in C mineralization rates between varieties decreased with time, but even at low rates, Bt-NK4640 leaves showed higher mineralization rates than NK4640, while this was only found for the first three set points in root mineralization (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00374-011-0625-6/MediaObjects/374_2011_625_Fig2_HTML.gif
Fig. 2

Carbon dioxide evolution rates before and after adding leaves and roots of five maize varieties to the same soil the plants were cultivated in. Mean values of the three replicate containers are given (n = 3). Mean coefficient of variation was 12% for leaves and 13% for the roots

Table 6

Endpoints of leaf and root decomposition over 24 days as percentage of the applied maize C for selected varieties in the second cropping cycle

 

Leaf

Root

[% of applied C]

Bt-NK4640

62.4 a

26.2 a

NK4640

51.6 c

21.1 c

Max88

53.8 c

22.0 bc

Gavott

59.9 ab

26.0 ab

Ronaldinio

54.9 bc

23.1 abc

Different letters indicate significant differences according to a Tukey HSD test (p = 0.05; n = 3)

Discussion

Maize varieties compared in our study showed variable shoot and root yield, which contrasted in higher yield for Bt varieties. Shoot yield of Bt varieties was 14% (contrast: p = 0.039) and root biomass 6% (contrast: n.s.) higher than the average of the other varieties even under conditions where environmental stresses are excluded in the climate chamber and the Bt trait is not at any advantage due to the lack of pests. No difference in shoot yield was found when comparing Bt-NK4640 to its near isogenic line.

The often-stated differences of Bt and non-Bt varieties in tissue composition, especially lignin compounds (Flores et al. 2005; Poerschmann et al. 2005), was not supported by our results and thus merely supports the notion that differences in lignin are genetically determined (Tarkalson et al. 2008) or due to different levels of maturity (Griffiths et al. 2008). As our plants were grown in a climate chamber under controlled conditions and were harvested before maturity, these results have to be judged cautiously, as they may be different in the field, when maturity stage has been reached (Raubuch et al. 2007). Differences in lignin and soluble plant compounds were hypothesized to influence decomposition, but it was only the work of Flores et al. (2005) that confirmed the hypothesis based on a series of experiments with transgenic Bt plants. Hopkins and Gregorich (2003) found no differences in decomposition of Bt and non-Bt plant materials. In our study, decomposition rates of leaf and root material from Gavott and Bt-NK4640 were higher compared to Max88, Ronaldinio, and NK4640. This difference was not only most pronounced shortly after mixing the plant material into the soil, but also the sum of evolved CO2 at the end of the study showed this effect. In plant material of Bt varieties, hot water-soluble C and N was 17% (contrast: p < 0.001) higher than in the average of non-Bt varieties. Higher concentration of easily available C in residue may stimulate a faster decomposition of the residue, but neither the initial shoot nor the root decomposition rates were correlated to the hot water-extractable C. Our results confirm earlier findings of Raubuch et al. (2007) who stated a strong initial flush of CO2 evolving from leaves of Bt varieties grown in the field until maturity compared to their near isogenic counterparts.

While most of the soil parameters showed only minor effects or remained unaffected by the maize varieties, it was only the DHA that contrasted with 5% (p = 0.045) lower values in soils under the Bt varieties compared to non-Bt. Devare et al. (2004, 2007) found no significant effects of Cry3Bb transgenic maize on soil microbial biomass, N mineralization, nitrification, and soil respiration in a field study, and in a recent study of Icoz et al. (2008), no or only minor effects of transgenic maize varieties on soil enzyme activities and on cultivable bacteria and fungi as well as cultivation-independent 16S DNA profiles were stated. Microbial community structure in rhizosphere and bulk soil samples under Bt-maize showed minor differences as revealed by phospholipid fatty acid (PFLA) profiles and community level physiological profiles (Blackwood and Buyer 2004). Baumgarte and Tebbe (2005) characterized the rhizosphere bacterial community of transgenic maize (MON810) by single-strand conformation polymorphism with small or negligible effects. Wei Xiang et al. (2004a, b) found some transient effects on some microbial activities after the incorporation of Bt-rice to water-flooded soils and also persistent effects on DHA, methanogenesis, and anaerobic respiration. In a decomposition experiment with Bt-maize residues, Raubuch et al. (2010) stated a transient but significantly increased CO2 production rate with Bt-maize compared to the soil amended with non-Bt- maize. Contrasting to our study, the authors state that the microbial biomass uses less of the energy provided by the Bt varieties for growth, followed by an increased death rate of the microbial biomass.

The cultivation of GM maize expressing larvicidal proteins from Bt has not shown major reductions in soil microbial biomass and activity. No interactions were found between varieties and the two cropping cycles. Therefore, the repeated cultivation of Bt-maize under our growth conditions has not increased the risk to soil microbiota. Despite the differences in decomposition and the significant contrast in DHA between Bt and non-Bt plants, effects of Bt-maize varieties were found within the variation of the effects of the ten varieties used in this study. Our results are in line with most of the previous studies on Bt-maize influence on soil microbial biomass and activity that were merely stated as transient or negligible.

Acknowledgements

The work of Kathi Hothum and Antje Stotz as part of their practical stage at FiBL is gratefully acknowledged. We are especially grateful to Prof. Dr. Geneviève Défago (ETH, Zürich) who kindly provided the seeds of genetically modified Bt-maize that we were otherwise not able to achieve. The project was funded by the Swiss National Science Foundation in the frame of the NFP59 “Benefits and Risks of the Deliberate Release of Genetically Modified Plants.” Within the framework of NFP59, we thank Michael Sander and Michael Madliger (both ETH, Zürich) for their support in the quantification of the Cry1Ab proteins and Claudia Zwahlen (Uni Neuchâtel) for the helpful comments on the outcome of this project. Two unknown reviewers provided valuable input to the submitted manuscript and are gratefully acknowledged here.

Copyright information

© Springer-Verlag 2011