Table Of ContentsNext Page

The development of endemic baculoviruses of Plutella xylostella (diamondback moth, DBM) for control of DBM in East Africa

David Grzywacz1, Mark Parnell 1, Gilbert Kibata2, George Oduor3, Walter Ogutu3, Douglas Miano2 & Doreen Winstanley4

1Sustainable Agriculture Group, Natural Resources Institute, University of Greenwich, Central Avenue, Chatham Maritime, Kent ME4 4TB, UK
Kenya Agricultural Research Institute, Waiyaki Way, PO Box 14733, Nairobi, Kenya
CAB International, Africa Regional Centre, PO Box 633, Village Market, Nairobi, Kenya
Horticultural Research International, Wellesbourne, Warwickshire, CV35 9EF, UK
Corresponding author:


A project to develop non-chemical methods of DBM control on Brassica crops in Kenya has been exploring the use of endemic pathogens as potential control agents. Initial surveys for endemic pathogens identified P. xylostella granulovirus (PlxyGV) on farms in Kenya. Subsequently 14 genetically distinguishable isolates were identified from field collected material. These were purified and ranging bioassays showed these isolates were pathogenic to Kenyan strains of DBM with LC50s varying from 2.36 x 106 to 3.95 x 107 occlusion bodies (OB) per ml for II instar DBM. One isolate (Nya-01) was selected and subsequently used for field trials in Kenya. The trials showed that unformulated PlxyGV applied at weekly intervals at a rate of 3.0 x 1013 OB/ha could control DBM on kale more effectively than available chemical insecticides. After application, infection rates in DBM can reach 90%. Further field trials are currently underway to determine the lowest effective dose rate for this virus when applied as a formulation. Initial virus production studies using in vivo propagation in II instar DBM reared on cabbage showed an initial productivity of 4.0 0.44 x 1010 OB/larva.


Brassica, granulovirus, biocontrol, Kenya


Larvae of the diamondback moth (DBM) Plutella xylostella, feed only on plants from the family Brassicaceae and are a major pest of Brassica vegetables (kale, cabbage, rapeseed, etc.) throughout Kenya (Michalik 1994). Presently, conventional chemical insecticides are heavily relied upon to control them (Kibata 1997). It is well known that DBM has become resistant to chemical insecticides in many countries throughout the world (Roush 1997) and current programs underway in Kenya have indicated that chemical resistance in DBM is also occurring there (Kibata 1997). The chemical insecticides currently recommended for control are expensive, damaging to the environment and in some areas simply not available to the small-scale farmers who account for a high percentage of the Brassica vegetable production of Kenya (Kibata 1997).

To address this issue, a collaborative project between the Natural Resources Institute (NRI), the Kenya Agricultural Research Institute (KARI) and CAB International, Africa Regional Centre (CABI-ARC) was set up to investigate alternative methods of DBM control. One component of the project investigated the possible use of endemic baculoviruses.

Before this study GVs of P. xylostella had been reported from Japan (Asayama & Osaki 1970), Taiwan (Wang & Rose 1978, Abdul Kadir 1986), China (Abdul Kadir et al. 1999) and India (Rabindra et al.1997), but there were no previous published records from Africa. A number of other NPVs, some uncharacterised (Padamvathamma & Veeresh 1989), have been reported as infecting DBM, but a review of the potential of DBM pathogens concluded that only the GV showed promising levels of pathogenicity (Wilding 1986). More recently an NPV has been identified from P. xylostella in China. This was characterised as being genetically similar to, though genetically distinct from, Autographa californica MNPV and A. falcifera MNPV (Kariuki & McIntosh 1999).

Materials and methods

Pathogen survey and identification

To collect baculoviruses, a survey of Brassica farms was conducted around Nairobi. In total, 27 farms were surveyed within a 170 km radius of Nairobi. In field sampling, larvae showing signs of baculovirus infection, puffy appearance and pale-yellow to white colouration (Asayama & Osaki 1970) were collected and individually stored for later examination. Standard, unstained wet mounts of infected larvae were examined using a microscope and dark-field contrast at x400 magnification to detect the presence of baculoviruses. Each candidate GV isolate was propagated in vivo in 15 II instar DBM following methods described by Parnell (1999). Restriction endonuclease analysis (REN) of the baculovirus isolates was performed on each of the GV isolates individually following the protocol of Smith and Summers (1978) as modified by Rabindra et al. (1997).

Bioassay of pathogen strains

The pathogenicity of the different isolates was determined by means of two bioassay methods. Comparative bioassays using single discriminate doses were performed on nine GV isolates displaying different DNA profiles. Subsequently, in order to obtain LC50 values, dose series bioassays were carried out on three of those eight isolates and the PlxyGV-Tw isolate. The concentration of GV was determined by counting using a 0.02 mm depth bacterial spore-counting chamber viewed under dark phase illumination at x200 magnification. Discriminate dose bioassays and dose response bioassays were carried out as per Parnell (1999). Bioassay data were corrected using Abbot's correction for control mortality and dose series data analysed using a probit analysis with the SPSS data analysis package.

Field trials

To evaluate the potential of the Kenyan PlxyGV to control crop loss caused by DBM, isolate Nya-01 was selected for mass production and use in small-plot field trials. This isolate was selected because it had been indicated as the most pathogenic strain in the laboratory bioassays. The virus was applied as a simple unformulated suspension using standard farmer equipment. Volume application rate for all treatments was 800 litres/ha. The first field trial was carried out on the research farm at Jomo Kenyatta University of Agricultural Technology (JKUAT) 25 km outside Nairobi and ran for 12 weeks in late1998. This was a randomised-block design trial carried out on small plots of 5 m x 5 m with a one metre gap between plots and a plant/row spacing of 60 cm. The test crop was kale (var. Thousand headed). This trial compared two virus treatments, a weekly application of high application rate of 3.0 x 1014 (occlusion bodies {OB}) and a medium rate of 3.0 x 1013 OB/ha. There was a no treatment control and a standard farmer insecticide treatment schedule based upon weekly application of the local standard pyrethroid insecticide (Karate- lambda-cyhalothrin).

A second field trial was carried out at the National Agricultural Research Laboratory (NARL) farm on the outskirts of Nairobi in 2000. In this trial there were five treatments arranged in randomised replicated plot design. The treatments were three virus application rates (3 x 1014, 3 x 1013 and 3 x 1012 OB/ha) a no treatment control and a standard insecticide treatment with Karate as before. The plots were 5 x 5 m with a one metre gap between plots and a plant spacing of 60 x 60 cm.

In both trials, 10 random plants in the central area of the plot were sampled weekly for numbers of DBM larvae, numbers showing symptoms of GV infection and damage caused by DBM. In addition, in the second trial yield data were also collected. To assess more precisely the disease incidence in the plots, after three weeks of the trial 45 larvae of each instar were collected from each treatment and reared individually in the laboratory and the disease occurrence recorded. The yield data were analysed using 2 way ANOVA on the SigmaStat statistical package (SPSS Inc., USA).

PlxyGV productivity

In order to estimate the productivity of the PlxyGV when produced in vivo, two hundred II and III instar larvae were inoculated with a range of concentrations of the strain Nya-01 and reared under standard conditions until death. Progeny virus was collected, counted and its identity confirmed using REN.


During the field survey, 127 larvae with disease symptoms were collected from eight of the 27 farms included in the survey. Microscopic examination confirmed that 95 larvae collected from four of the eight farms were suffering from GV infection. The areas in which GV-infected larvae were found were Nyathuna (84 larvae-two farms), South Kinangop (9 larvae) and Naivasha (2 larvae).

The REN analysis of the 95 PlxyGV isolates showed that 14 had fragment profiles that could be distinguished from any other with both EcoR1 and Pst1 cuts (Figure 1). Comparison of these 14 Kenyan PlxyGV isolates to an isolate of PlxyGV from Taiwan (PlxyGV-Tw) revealed that, although the profiles had many similarities, there were major band differences between all isolates. Both the Pst1 and EcoR1 digests revealed between 2 and 6 major band differences between isolates, even in those collected from the same location (Figure 1).

Figure 1. Comparison of PlxyGV isolates. DNA of each isolate was digested with Pst1 restriction endonuclease, fragments were separated on 0.6% agarose gel. Track 1 (far left of page), 1kb molecular size standard; tracks 2-16, Kenyan PlxyGV isolates from Nyathuna (Nya-01, Nya-02, Nya-03, Nya-06, Nya-07, Nya-14, Nya-15, Nya-25, Nya-27, Ny-29, Nya-35, Nya-37, Nya-40, Nya-42, Nya-52 respectively), track 17, PlxyGV isolate from South Kinangop (SK-01); track 18, Taiwanese PlxyGV; Track 19, λ 19-Mix molecular size standard.

Results from the discriminate dose assay showed every Kenyan isolate to be significantly more potent than the PlxyGV-Tw with average % mortality ranging from 26.2% to 40.3% compared with 5.2% for the PlxyGV-Tw (Figure 2). However in the dose response bioassays, no significant differences in LC50 values between Kenyan isolates and the PlxyGV-Tw isolate were observed. Average LC50 values for II instar DBM larvae varied from 2.36 x 106 OB/ml for Nya-01 PlxyGV to 3.95 x 107 OB/ml for Nya-40 PlxyGV. In comparison, the LC50 for the PlxyGV-Tw was 1.55 x 107 OB/ml.

Figure 2. Average percent mortality in discriminate dose bioassays of Kenyan PlxyGV and Taiwanese PlxyGV.

The field trials carried out at JKUAT showed that the PlxyGV when sprayed using standard farmer application equipment was highly infectious to DBM, spreading rapidly in trial plots and infecting 80-90% of larvae within two to three weeks of application (Figure 3). Very little occurrence of infected insects was recorded from the control or insecticide treated plots. Both the high dose rate of 3.0 x 1014 OB/ha and the lower dose of 3.0 x 1013 OB/ha reduced DBM damage to crops to below that seen in either unsprayed controls or insecticide treated plots (Figure 4).

Figure 3. The level of PlxyGV infection observed in treatments from the first field trial in Kenya.

Figure 4. The level of crop damage observed in treatments from the first field trial in Kenya.

In the second trial at NARL, the yield data (Figure 5) showed that the highest application rate dose gave significantly higher yield than the no treatment control (37% higher, P<0.001 df=4 and 28, F=6.25) or the insecticide treatment (17% higher, P<0.001 df=4 and 28, F=6.25). The average DBM numbers in each treatment showed an application-rate effect with the lowest numbers occurring in the highest virus rate treatment (Figure 6). In the second trial, average observed DBM infection rates in virus treated plots also showed a clear application-rate trend with the highest dose producing an average of 40% (Figure 7). In this trial there was some infection observed in the control and insecticide plots. From insects sampled from the PlxyGV application-rate plots, the true infection rate was much higher than that observed in the field and Table 1 shows the percent virus mortality recorded from insects taken from the plot treated at 3 x 1013 OB/ha.

Figure 5. Average kale yield per hectare for each of the treatments from the NARL field site.

Figure 6. Average DBM population per hectare for each of the treatments at the NARL field site.

Figure 7. Average PlxyGV infection-rate observed in DBM larvae for each of the treatments at the NARL site.

Table 1. Percentage GV infections that developed in laboratory-reared larvae sampled from virus plots sprayed with 3 x 1013 OB/ha

Larval instar

DBM Infection Rate









The maximum productivity of the PlxyGV was found to be 4.0 0.44 x 1010 OB/larva obtained from II instars inoculated with 2.0 x 108 OB/ml.


The pathogen survey revealed that the GV of DBM occurred on 50% of the farms surveyed, though, in all cases, with a relatively low incidence. On no farm were widespread epizootics observed or reported by local farmers questioned. The discovery of so many different genetic isolates (14) in the small number of infected larvae collected is therefore striking. Previously reported work (Abdul Kadir et al. 1999) has characterised only two genetically distinct isolates one from China and one from Taiwan. Other studies of DBM pathogens have also only reported finding a single genetically distinct isolate from India (Rabindra et al. 1997) and Japan (Yamada & Yamaguchi 1985).

The GV isolates from Kenya are genetically similar to, though genetically distinct from, the previously reported Taiwanese isolate. This isolate we now know is itself similar to and closely related to the Chinese isolate (Abdul Kadir et al. 1999). The two isolates studied differed from each other by one to three major bands in the EcoR1, BamH1 and HindIII profiles. The differences in the Kenyan isolates studied here were greater at two to six bands with only two profiles EcoR1 and Pst1, even amongst isolates collected from the same farm. This genetic diversity amongst isolates of PlxyGV from Kenya could be extremely useful as a diverse genetic resource that could be exploited in the development of a GV for DBM control. The high level of variation in the PlxyGV isolates could indicate a long association between PlxyGV and DBM in the region and could have a bearing on the debate concerning the origin of DBM. This was generally considered to be somewhere in Mediterranean Europe having evolved on cultivated brassicas also believed to have European origin (Hardy 1938). Recently however, the Mediterranean origin of DBM has been brought into question by Kfir (1998) who hypothesised a southern African origin for DBM on the basis of the diversity of wild hosts and endemic parasitoids found in South Africa. The genetic variation in PlxyGV isolates discovered in Kenya during the present study and apparent lack of diversity in isolates from other regions of the world might be interpreted as providing additional support to the theory that the origin of DBM lies in Sub-Saharan Africa.

The initial discriminate single dose bioassay results showed all the Kenyan isolates to be significantly more pathogenic than the Taiwanese isolate. However the LC50 data from the subsequent dose response assays showed no significant differences, even though the mean LC50 for Taiwanese isolate was 6.5 times higher that of the most active Kenyan isolate (Nya-01). This result reflects the high variability in response seen with some Kenyan isolates including Nya-01. These were originally in vivo propagated, but not cloned, which might have reduced this variability. These isolates have since been cloned and the assays are currently being repeated on these cloned isolates.

The productivity of the Kenyan isolates is high at 4.0 0.44 x 1010 OB/larva, equivalent to 8.0 x 109 OB/mg. This may be compared with between 1.9 x 1010 and 4.5 x 109/larva reported with other GVs produced in Lepidoptera (Evans 1986). High productivity is a valuable asset in a potential biopesticide as it reduces the number of insects needed to produce the desired application rate. At this rate of production, the highest application rate used in these trials, 3.0 x 1014 OB/ha would be equivalent to 7,500 infected larvae/ha. In comparison, most existing commercial baculovirus products are applied at rates of between 50-500 larval equivalents/ha (Moscardi 1999).

The first field trial showed that application of PlxyGV at 3 x 1013 OB/ha could reduce DBM damage much better than either the use of the standard chemical insecticide or the no treatment control. The very limited effectiveness of the standard insecticide, lambda-cyhalothrin, suggests that significant resistance has developed in DBM. Resistance has since been confirmed by other work in Kenya (Cooper 2001) and lambda-cyhalothrin is now no longer recommended for DBM control.

The speed with which weekly sprays of PlxyGV initiated infection rates of 90% could indicate that one or two applications of PlxyGV at the start of the season might be sufficient to start an epizootic infection in resident DBM populations. However whether augmentative approach alone would be sufficient to produce control of DBM numbers and damage, would need testing under field conditions. While collection of a high percentage of infected insects in virus treated plots suggests that recycling of PlxyGV is very important, its precise contribution to control remains to be quantified.

In the second trial, the yield results showed that again the PlxyGV performed significantly better than the chemical insecticide at the highest application rate used 3 x 1014 OB/ha. A similar result in terms of controlling DBM numbers has been reported by Su (1989) using a Taiwanese isolate applied as here at seven day intervals. However direct comparisons are difficult, as in Su's (1989) trial, the PlxyGV was quantified in terms of larval equivalents per litre and no direct enumeration of the GV was carried out.

Glasshouse trials again have showed that application of the Taiwanese isolate can reduce DBM numbers and that there is a dose response over the range 9 x 1011 to 9 x 1013 and at the highest dose the PlxyGV reduced damage as effectively as application of Bacillus thuringiensis (Abdul Kadir 1992). In addition it was shown that the addition of molasses to a formulation could increase the viruses efficacy by a factor of ten and allow for a consequent reduction in the application rate of PlxyGV. This finding closely mirrors that of Ballard et al. (2000) who found that addition of 10% molasses produced a similar 10 fold increase in efficacy with the codling moth (Cydia pomonella) granulovirus (CpGV) on apples.

The two granuloviruses that have been commercialised to date, CpGV and Adoxophyes orana granulovirus, are both sold for application at rates of 1 x 1013 OB/ha. In comparison, the rate of PlxyGV used here which produced a significant increase in yield is 3 x1 014 OB/ha. Even given that the Kenyan PlxyGV seems to be more productive than other GVs, this suggests a need to reduce the application rate by a factor of ten if its use is to be commercially attractive.

The trials reported here did not include formulation ingredients and field trials of such a formulation are underway now in Kenya to evaluate the efficacy of reduced rate formulated PlxyGV. Formulation might also address the short persistence time on field crops seen with GVs. Abdul Kadir (1986) reported that with PlxyGV-Tw exposure of unformulated virus to seven hours of sunlight in Malaysia was sufficient to reduce virus efficacy by 50%. Although the persistence of the Kenya PlxyGV has yet to be quantified it is unlikely to be longer.

In conclusion, while the results of these trials of PlxyGV are promising it has yet to be determined that PlxyGV can be effective or reliable enough for consistent control of DBM.


Abdul Kadir HB. 1986. The granulosis virus of Plutella xylostella. In: Biological Control in the Tropics: Proceedings of the first regional symposium on biological control, Malaysia, September 1986. MARDI, Malaysia, pp. 261-273.

Abdul Kadir HB. 1992. Potential of several baculoviruses for the control of diamondback moth and Crocidolomia binotalis on cabbages. In: Diamondback moth and other crucifer pests (ed NS Talekar). Proceedings of the Second International Workshop, Tainan, Taiwan, 10-14 December 1990, Asian Vegetable Research and Development Center, Shanhua, Taiwan, AVRDC Publication No. 92-368, pp. 185-192.

Abdul Kadir HB, Payne CC, Crook NE & Winstanley D. 1999. Characterization and cross-transmission of baculoviruses infective to the diamondback moth, Plutella xylostella, and some other Lepidopterous pests of brassica crops. Biocontrol Science and Technology 9, 227-238.

Asayama T. & Osaki N. 1970. A granulosis of the diamondback moth, Plutella xylostella. Journal of Invertebrate Pathology 15, 284-286.

Ballard J, Ellis DJ & Payne CC. 2000. The role of formulation additives in increasing the potency of Cydia pomonella granulovirus for codling moth larvae in laboratory and field experiments. Biocontrol Science and Technology 10, 627-640.

Cooper J. 2001. Pest management in horticultural crops: an integrated approach to vegetable pest management with the aim of reducing reliance on pesticides in Kenya. NRI Final Technical Report R7403, pp. 36.

Evans HF. 1986. Ecology and epizootiology of Baculoviruses. In: The Biology of Baculoviruses. Vol. II Practical Application for Insect Control (eds RR Granados & BA Federici), pp. 89-132. CRC Press, Boca Raton, Florida.

Hardy JE. 1938. Plutella maculipennis Curt. Its natural and biological control in England. Bulletin of Entomological Research 29, 343-372.

Kariuki CW & McIntosh AH. 1999. Infectivity studies of a new baculovirus isolate for the control of the diamondback moth (Plutellidae: Lepidoptera). Journal of Economic Entomology 92, 1093-1098.

Kfir R. 1998. Origin of the diamondback moth (Lepidoptera: Plutellidae). Annals of the Entomological Society of America 91, 164-167.

Kibata GN. 1997. The diamondback moth: A problem pest of brassica crops in Kenya. In: The management of diamondback moth and other crucifer pests (eds A Sivapragasam, WH Loke, AK Hussan & GS Lim). Proceedings of the Third International Workshop, 29 October - 1 November 1996, Kuala Lumpur, Malaysia, Malaysian Agricultural Research and Development Institute (MARDI), pp. 47-53.

Michalik S. 1994. Report on crop protection measures of Kenyan vegetable farmers and their use of pesticides: a knowledge, attitude and practical survey. KARI/GTZ IPM Horticulture Project, p 7. Published by GTZ.

Moscardi F. 1999. Assessment of the application of baculoviruses for control of Lepidoptera. Annual Review of Entomology 44, 257-289.

Padmavathamma K & Vereesh GK. 1989. Effect of contamination of eggs of the diamondback moth, Plutella xylostella (Linnaeus) with nuclear polyhedrosis virus. Journal of. Biological Control 3, 73-74.

Parnell MA. 1999. The genetic variability and efficacy of baculoviruses for control of diamondback moth on brassica vegetables in Kenya. MSc Thesis, University of Greenwich, Greenwich, UK. pp 72.

Rabindra RJ, Geetha N, Renuka S, Varadharajan S, Regupathy A. 1997. Occurrence of a granulosis virus from two populations of Plutella xylostella (L.) in India. In: The management of diamondback moth and other crucifer pests (eds A Sivapragasam, WH Loke, AK Hussan & GS Lim). Proceedings of the Third International Workshop, 29 October - 1 November 1996, Kuala Lumpur, Malaysia, Malaysian Agricultural Research and Development Institute (MARDI), pp. 113-115.

Roush R. 1997. Insecticide resistance management in diamondback moth: quo vadis? In: The management of diamondback moth and other crucifer pests (eds A Sivapragasam, WH Loke, AK Hussan & GS Lim). Proceedings of the Third International Workshop, 29 October - 1 November 1996, Kuala Lumpur, Malaysia, Malaysian Agricultural Research and Development Institute (MARDI), pp. 21-24.

Smith GE & Summers MD. 1978. Analysis of baculovirus genomes with restriction endonucleases. Virology 89, 517-527.

Su CY. 1989. The evaluation of granulosis viruses for control of three lepidopterous insect pests on cruciferous vegetables. Chinese Journal of Entomology 9, 189-196.

Tabashnik BE. 1994. Evolution of resistance to Bacillus thuringiensis. Annual Review of Entomology 39, 47-79.

Wang CL & Rose RI. 1978. Control of imported cabbageworm Pieris rapae crucivora with granulosis virus in the field. Plant Protection Bulletin 20, 16-20.

Wilding N. 1986. The pathogens of diamondback moth and their potential for its control -a review. In: Diamondback moth management (eds NS Talekar & TD Griggs). Proceedings of the First International Workshop, 11-15 March 1985, Tainan, Taiwan, The Asian Vegetable Research and Development Center, Shanhua, Taiwan, AVRDC Publication No. 86-248, pp. 219-232.

Wright DJ, Iqbal M, Verkerk HJ. 1995. Resistance to Bacillus thuringiensis and abamectin in the diamondback moth Plutella xylostella: A major problem for integrated pest management? Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen Universiteit Gent 60, 927-932.

Yamada H & Yamaguchi T. 1985. Notes on parasites and predators attacking the diamondback moth, Plutella xylostella. Japanese Journal of Applied Entomology and Zoology 29, 170-173.

Top Of PageNext Page