Effect of Bacillus thuringiensis and a pyrethroid insecticide on the leaf microflora of Brassica oleracea


Dr J.M. Chard, Scottish Agricultural Science Agency, East Craigs, Edinburgh EH12 8NJ, UK.


In a field experiment comparing the effect of treatment with Bacillus thuringiensis (Bt) and the pyrethroid insecticide, cypermethrin, on leaf microflora across the growing season, no significant differences were found, apart from more bacteria on cypermethrin-treated leaves on 24 September (P < 0·05). No effect was seen on fungal diversity or numbers. In contrast, the pyrethroid insecticide inhibited growth of 44% of bacterial isolates tested in vitro; Bt was not inhibitory. Bacillus thuringiensis numbers on treated leaves increased throughout the season, following repeated applications.

Strains of Bacillus thuringiensis (Bt) have been used to control insect pests ( Entwistle et al. 1993 ) and new strains have been isolated and characterized from a range of sources ( Chilcott & Wigley 1994). Bacillus thuringiensis has been shown to be ubiquitous in soil ( Dulmage & Aizawa 1982; Martin & Travers 1989), but it has also been isolated from leaves ( Smith & Couche 1991). The latter authors consider that it could be a common phylloplane organism on many plants. The environmental impact of Bt as a biopesticide has been addressed and Bt is considered environmentally benign. However, gaps in our knowledge of the ecology, physiology and population genetics of Bt have been identified ( Meadows 1993). Studies on the effects of pesticides on non-target organisms are a necessary part of the pesticide registration process in many countries. For most pesticides, effects on micro-organisms are targeted to the soil environment and published data on the effects of insecticides on phylloplane organisms are more limited ( Andrews 1981). When used for insect control, Bt is usually applied to leaves and little is known of its interaction with other phylloplane organisms.

This paper describes a preliminary study to determine the effects on the microflora of Brassica oleracea (Brussels sprout) leaves over a growing season following treatments with Bt and a pyrethroid insecticide.

Materials and methods

Experimental design and field collections

A field experiment was set up in 1985 according to the European and Mediterranean Plant Protection Organization guidelines for the biological evaluation of insecticides on caterpillars of leaf brassicas ( Anonymous 1984). Twelve plots, each with 48 plants of the dwarf variety Lunet, were used with four replicates of three treatments in a randomized design. The three treatments were the pyrethroid insecticide, cypermethrin (‘Ambush C’, formerly ICI, now Zeneca Crop Protection, an emulsifiable concentrate formulation containing 100 g l−1 active ingredient applied at 250 ml ha−1), Bt (‘Biobit’ formerly Microbial Resources, now Novo Nordisk, a wettable powder formulation containing 16000 IU mg−1 var kurstaki active ingredient applied at 1 kg ha−1) and control plots which were left unsprayed. The treatments were applied as foliar sprays every 2 weeks. Plots were sampled 1 week after the application of sprays. The middle west-facing leaf from five randomly selected plants per plot was removed for analysis.

Meteorological data were derived from Meteorological Office records from the Royal Botanic Garden, Edinburgh.

Laboratory procedures

Four media were used for assessment of leaf microbial populations, potato dextrose agar (PDA, Oxoid), Kings medium B (KB, King et al. 1954 ), nutrient agar (NA, Oxoid) and, for recovery of Bt, Bacillus cereus medium (BCM, Oxoid).

Leaves were weighed and measured for surface area calculation (length × breadth × correction factor = surface area (cm2)). The correction factor was worked out for 20 leaves and was calculated to be 0·0081 cm2. Each sample of five leaves was agitated with 100 ml quarter strength ringers solution (Oxoid) containing 0·1% w/v bacteriological peptone (Oxoid) using a stomacher for approximately 2 min. Serial dilutions of the washings (to 10–5) were made and aliquots spread over three replicate plates of each medium per dilution. NA, KB and BCM plates were incubated for 2 d at 27 ± 2 °C, and PDA plates were incubated at the same temperature for 7 d. Numbers of micro-organisms were calculated by leaf area (sq cm−1) and fresh weight (g−1). Fungi were identified on the basis of spore morphology and colony characteristics; Bt forms characteristic colonies on the BCM medium.

Eighteen bacterial isolates were obtained from control plates and tested for resistance to the pyrethroid insecticide and Bt. The bacterial isolates were identified to genus using API strips (API 20B, API System S.A.) and the following laboratory tests: motility, Gram reaction, spore staining, Hugh and Leifsons test ( Hugh & Leifson 1953). For insecticide resistance, 5 ml molten NA was cooled to 45 °C and spread-inoculated with a bacterial suspension from bacteria isolated from control plots. Three sterile filter paper discs (10 mm diameter) were stacked on top of each other on the solidified agar and a 150 μl drop of filter-sterilized cyper-methrin (100 ppm in acetone) placed on top; control discs were set up with acetone alone. Four filter paper stacks were placed on each plate. Plates were incubated for 7 d at 27 ± 2 °C and the inhibition zone measured from the edge of the filter paper to the start of bacterial growth. Three replicate plates were used per isolate and mean inhibition zones calculated. Resistance to Bt was assessed by streaking Bt down the centre of an NA plate and allowing it to grow for 24 h at 27 ± 2 °C; test isolates were streaked at 90° to the Bt and grown for 7 d, after which time inhibition zones were measured. Three replicate plates were set up for each bacterial isolate.


Data were analysed by analysis of variance using Genstat 5, Release 2 (Numerical Algorithms Group Ltd). Counts were transformed to log10 + 1 before analysis.

Results and discussion

No significant differences (P > 0·05) were found in microbial counts on NA or PDA at any of the sampling dates ( Fig. 1a-c) apart from samples taken on 24 September, when more bacteria were recorded on NA from cypermethrin-treated plants than from the other two treatments. The microbial counts were highest on leaves on 26 July and counts may have been affected by exceptional rainfall that day (82 mm) ( Fig. 1d). Lower numbers were recorded on subsequent sampling dates with a trend of increased numbers at the end of the season. Similar fluctuations in phylloplane populations have been observed in many studies, with numbers being related to age of leaf and weather conditions ( Kinkel 1997).

Figure 1.

Incidence of (a) bacteria, (b) fungi and (c) yeasts on Brassica oleracea leaf surfaces treated with Bacillus thuringiensis (Bt) and the pyrethroid insecticide, cypermethrin. (d) Rainfall (mm) during the experimental period, July to October 1985. Counts are means of three replicates (log 10 cfu+1 cm−2 leaf) at each sampling date. (bsl00008) Cypermethrin, (░) Bt, and (▪) untreated. *Significantly different (P < 0·05)

The relative frequency of four fungal species throughout the sampling period is shown in Fig. 2; no major differences in species composition nor any significant differences (P > 0·05) in numbers were found with treatment. A succession of fungi was recorded, with Aureobasidium pullulans, Fusarium sp., Cladosporium sp. and Monilia sp. being found throughout the season. Genera such as Mucor, Sporobolomyces and Candida were recorded only at the first sampling date, Botrytis, Phoma and Epicoccum were found in September and October, and Aspergillus, Penicillium and Stemphyllium were found only at the end of the season.

Figure 2.

Incidence of four fungi on Brassica oleracea leaves at five sampling times. Counts are means of three replicates (log 10 cfu+1 cm−2 leaf). (bsl00008) Cypermethrin, (░) Bt, and (▪) untreated. (a) Aureobasidium pullulans; (b) Fusarium spp.; (c) Cladosporium spp.; (d) Monilia spp.

The counts of fluorescent pseudomonads isolated on KB medium were higher in samples from cypermethrin-treated plots than from the other treatment plots, but not significantly (P > 0·05) (data not shown). Numbers of pseudomonads increased throughout the season in all treatments.

The number of Bt per leaf were relatively low at the start of the experiment but increased significantly throughout the season in the Bt-treated plots. The numbers of cfu cm−2 (log 10+1) ± s. e. were: 12 August, 0·036 ± 0·043; 3 September, 0·061 ± 0·042; 24 September 0·493 ± 0·192; 17 October, 1·476 ± 0·135. At the final sampling time, counts were within one order of magnitude of the total bacterial count. No Bt was isolated on the BCM medium in samples from the other treatments. Many studies have been done on the survival and efficacy of Bt following field applications ( Meadows 1993; Bryant 1994) and these have shown its rapid decline following exposure to u.v. after application to leaves.

Thirteen of the 18 bacterial isolates from the control plots were identified to genus or family level (Bacillus, Corynebacterium, Enterobacteriaceae, Flavobacterium, Micrococcus and Pseudomonas/Alcaligenes). Eight of the isolates failed to grow in the presence of cypermethrin and five other isolates had inhibition zones in excess of 10 mm ( Table 1). The most sensitive isolates were members of the Enterobacteriaceae; Bt did not inhibit the growth of any of the isolates.

Table 1.  Number of bacteria inhibited by the pyrethroid insecticide, cypermethrin, in an in vitro assay
Zone of inhibition (mm)
Bacteria1–56–1011–1516–20No growth
  1. Results are means of inhibition of 3 replicate plates containing 4 filter paper discs (1.5×10−6 mg cypermethrin).

Bacillus spp.  1 1
Corynebacterium spp.  1  
Enterobacteriaceae    3
Flavobacterium spp. 1   
Micrococcus spp.    1
Pseudomonas/Alcaligenes 2111
Unidentified111 2

The Bt treatments in this study did not appear to affect the phylloplane microflora on Brassica oleracea. Cypermethrin treatment similarly had no effect apart from an increase in bacterial numbers on one occasion which was unexpected. The potential of cypermethrin to affect bacterial population diversity was demonstrated by inhibition of growth in vitro which is contrary to the observed effect. On all the other sampling dates, insecticide-treated leaves had lower bacterial numbers, which, although not significant, would be more expected from the in vitro work. This experiment did not address whether ‘resistant’ bacterial species had developed in the field plots.