0851996884.pdf

should also be kept and reviewed on a
monthly basis. Even so, large-scale produc-
tion of larvae with synchronized predictable
growth is difficult to achieve in practice
(Shieh, 1989). Ensuring an adequate, sus-
tained supply of appropriate larvae for infec-
tion is probably the most demanding part of
any BV production process and is where
serious problems are most common.

Yields

The numbers of OB in BV products should
be counted by light microscopy and bioas-
sayed regularly, either at weekly intervals or
on every batch produced, to check for yield.
These counts should be made when insects
are harvested so that any production prob-
lems are identified quickly. Microscope
counts of NPV suspensions (Wigley, 1980;
Hunter-Fujita et al., 1998) are preferred over
indirect or non-visual methods, such as the
enzyme-linked immunosorbent assay, pro-
tein assay or use of particle counters (Jones,
2000). Counting is best done on the technical
product – that is, the processed BV suspen-
sion prior to final formulation. These proto-
cols for regular counting can be
time-consuming but they are at the core of
effective quality control, and failure to
implement an effective system may lead to
production problems later.

DNA profiling

When a new stock of inoculum is made (see
section on culture maintenance), DNA profil-
ing of BV produced should be carried out to
monitor the strain consistency, as genetic
diversity is common in BV (Gettig and
McCarthy, 1982). Routine DNA profiling of
selected production batches can also serve to
detect any viral contamination by unwanted
NPV, GV or even CPV. Samples for DNA
should be processed within a week or two at
most and not saved up for later analysis; oth-
erwise contamination by a foreign virus may
be detected too late to prevent it spreading
throughout the production system. When
this occurs, a major shut-down and disinfec-
tion of the facility is often the only solution.

Harvesting

Processing of NPV for formulation as micro-

bial control agents may be relatively simple,

involving no more than the preparation of

crude suspensions simply by macerating the

infected insect cadavers to release the BV.

Insect debris, such as skin and mandibles, is

then filtered out to prevent blockages in

sprayers. This is often as far as processing

goes for many products. More advanced pro-

cessing protocols include settling or the use

of centrifugation to concentrate the NPV and

remove lipid and insect-derived cellular

debris. This procedure also removes some of

the by-products that provide contaminant

microflora with substrates on which to mul-

tiply. However, simple centrifugation in

water does not remove many microbial cont-

aminants, as bacterial spores, etc., tend to

pellet with NPV OB (Grzywacz et al., 1997).

Bacteriostatic agents and pH buffers can be

added to stabilize formulations by reducing

the multiplication of bacterial contaminants.

Safety

BV are obligate parasites that can replicate

only in arthropods or arthropod cell lines

and so are inherently safe for humans. Many

BV are also highly host-specific; therefore the

main safety issue arises from possible conta-

mination of production systems by human

or veterinary pathogenic bacteria and fungi.

Studies have shown that a variety of species

may contaminate BV production systems

and multiply either in host insects during life

or on the cadavers or faeces (Podgwaite et al.,

1983).

The setting of arbitrary standards for non-

pathogenic contaminant bacteria is problem-

atic. Early BV commercial products, such as

ELCAR, were established with the vigorous

limit of 1 × 107 cfu g^1 non-pathogenic

microbial contaminants (Burges, 1981). It can

be argued that this was unnecessarily

demanding, contributing nothing significant

to improving real safety while burdening the

product with an excessively expensive qual-

ity control system (H.D. Burges, personal

communication).

Quality Control of Fungal and Viral Biocontrol Agents 257