Appl Microbiol Biotechnol (1999) 51: 79±84
ORIGINAL PAPER
P. Khampang á W. Chungjatupornchai
P. Luxananil á S. Panyim
Ef®cient expression of mosquito-larvicidal proteins
in a gram-negative bacterium capable of recolonization
in the guts of Anopheles dirus larva
Received: 20 February 1998 / Received last revision: 5 October 1998 / Accepted: 11 October 1998
Abstract The gram-negative bacterium, An11/2 G1,
isolated from the guts of Anopheles dirus mosquito lar-
vae, was identi®ed as Enterobacter amnigenus. The
E. amnigenus was able to recolonize in the gut of An.
dirus larva but not in those of Aedes aegypti and Culex
quinquefasciatus larvae. It was able to ¯oat in water for a
longer period than Bacillus thuringiensis subsp. israelen-
sis and Bacillus sphaericus. These are desirable charac-
teristics for a delivery vehicle of mosquito-larvicidal
toxins for the control of mosquito larvae, and E. amni-
genus was therefore used as a host to express the cryIVB
gene of B. thuringiensis subsp. israelensis and the binary
toxin genes of B. sphaericus. The recombinant E. amni-
genus produced a high level of CryIVB protein, which
was toxic to larvae of Ae. aegypti and An. dirus. Another
E. amnigenus producing the 51-kDa protein of
B. sphaericus was toxic to larvae of An. dirus and Cx.
quinquefasciatus. The recombinant plasmids were stable
in E. amnigenus without the presence of selective pressure
for at least 23 generations. The recombinant E. amnigenus
should represent a desirable biological agent for con-
trolling mosquito larvae.
Introduction
The gram-positive bacteria Bacillus thuringiensis subsp.
israelensis and Bacillus sphaericus produce parasporal
crystal proteins that are highly toxic to mosquito larvae.
W. Chungjatupornchai (&) á S. Panyim
Institute of Molecular Biology and
Genetics, Mahidol University, Salaya Campus,
Nakornpathom 73170, Thailand
e-mail: stwcj@mucc.mahidol.ac.th
Fax: +662-441-9906
Tel.: +662-441-9003-7, ext. 1240, 1235
P. Khampang á P. Luxananil
National Centre for Genetic Engineering and Biotechnology,
Bangkok, Thailand
Ó Springer-Verlag 1999
The crystal proteins of B. thuringiensis subsp. israelensis
are composed of at least four major polypeptides of 135,
128, 72, and 28 kDa, which are encoded by the cryIVA,
cryIVB, cryIVD, and cytA genes respectively (for a re-
view, see Hofte and Whiteley 1989). These polypeptides
are all toxic, although to di€erent extents (Chilcott and
Ellar 1988) and can also act in synergy (Wu and Chang
1985; Delecluse et al. 1993). The 128-kDa polypeptide
(CryIVB) is highly toxic to Aedes aegypti larva
(Chungjatupornchai et al. 1988). The parasporal crystal
of B. sphaericus, consisting of two polypeptides of
51 kDa and 42 kDa, is toxic to mosquito larvae of the
Culex and Anopheles species (Yousten 1984). Both
polypeptides are required for larvicidal activity (Broad-
well et al. 1990). The binary toxin genes have been
cloned, sequenced (Baumann et al. 1988; for a review see
Baumann et al. 1991) and eciently expressed in E. coli
(Ahmad et al. 1998).
B. thuringiensis subsp. israelensis and B. sphaericus
have been integrated into a vector for the control of
mosquitos. However, the currently available B. thurin-
giensis subsp. israelensis and B. sphaericus often sink
when sprayed on water, removing them from the feeding
area of the mosquito larvae (Priest 1992; Porter et al.
1993). Thus, for an e€ective control of mosquito larvae,
many applications of the bacteria are required.
In order to overcome these problems, several
B. thuringiensis subsp. israelensis and B. sphaericus genes
have been expressed in cyanobacteria (Angsuthana-
sombat and Panyim 1989; Chungjatupornchai 1990;
Murphy and Stephens 1992; Sangthongpitag et al. 1997),
Caulobacter crescentus (Thanabalu et al. 1992), Ancylo-
bacter aquaticus (Yap et al. 1994), and Asticcacaulis
excentricus (Liu et al. 1996). However, all these micro-
organisms used as hosts for the gene expression are
laboratory strains. The association between these micro-
organisms and mosquito larvae is not known, nor
whether they are present and how long they can survive
in the mosquito gut.
In this work, we present a gram-negative bacterium
that was isolated from the guts of mosquito larvae and is
80
able to recolonize in the gut of An. dirus larvae. This
bacterium, also able to ¯oat on water, was used as a host
to synthesize the CryIVB protein of B. thuringiensis
subsp. israelensis or the binary toxins of B. sphaericus.
The toxicities of the recombinant hosts against three
species of mosquito larvae and the stabilities of their
plasmids were determined.
Materials and methods
Bacterial strains and plasmids
Bacillus sphaericus 2297 was a gift from A. A. Yousten, Depart-
ment of Biology, Virginia Polytechnic and State University. An11/2
G1, isolated from the guts of mosquito larvae as decribed below,
was deposited in the Bangkok MIRCEN Culture Collection (code
number TISTR1343). Escherichia coli K1±14 contains a pBR322
derivative with the tetracycline-resistance gene. E. coli JM109 was
used as a recipient strain for cloning experiments. Plasmid pBR322
has been described previously (Bolivar et al. 1977). Plasmid
pMU388, which is the pUC12 vector carrying the 3.8-kb XbaI
fragment of the cryIVB gene of B. thuringiensis subsp. israelensis,
was generated previously (Angsuthanasombat et al. 1987). Plasmid
pMEx-5142, which is the pMEx8 vector containing the 2.6-kb
SmaI/HindIII fragment of the binary genes in a single tran-
scriptional operon encoding the 51-kDa and 42-kDa proteins
of B. sphaericus strain 2297, has been previously constructed
(Boonserm 1996). Both pMU388 and pMEx-5142 carry the ampi-
cillin-resistance gene.
Isolation and identi®cation of bacteria from the gut
of mosquito larvae
Mosquito larvae of An. dirus were washed several times with sterile
distilled water to eliminate surface micro-organisms before the guts
were dissected. The contents from each gut and water sample col-
lected from the same habitat were spread separately on Luria-
Bertani (LB) agar (1% tryptone, 0.5% NaCl, 0.5% yeast extract,
2% Bacto agar) plates. After incubation at 30 °C for 24 h, bacterial
colonies on the agar plates were counted. The isolated bacteria
were di€erentiated by morphology. One of the predominant bac-
teria, designated An11/2 G1, was further identi®ed by biochemical
tests. Biochemical reactions were performed using an automated
process and identi®cation kits (ATB plus and API identi®cation
program, Bio Merieux Vitek Inc., Mo., USA).
Electroporation of An11/2 G1
Competent cell preparations and electroporation for An11/2 G1
were slightly modi®ed from the previously described method
(Mahillon et al. 1989) as follows. Competent cells were washed and
resuspended in 10% glycerol with 1±2 lg DNA and the competent
cells were submitted to a single electrical pulse; the electroporation
apparatus was set at 1600 V, 25 lF, and the parallel resistor at
400 X. An11/2 G1 transformants were selected on LB agar con-
taining either tetracycline (25 lg/ml) or ampicillin (100 lg/ml).
Re-introduction of An11/2 G1 into the guts of mosquito larvae
An11/2 G1 harbouring pBR322 was re-introduced into the guts of
third-instar laboratory-reared Ae. aegypti, An. dirus and Cx.
quinquefasciatus larvae as follows. An11/2 G1 harbouring pBR322
and E. coli K1±14 (used as an internal marker) were grown sepa-
rately in LB broth at 30 °C for 16 h. Cells were harvested by
centrifugation, washed and adjusted to 4 ´ 108 cells/ml with sterile
distilled water. A 10-ml sample of An11/2 G1 harbouring pBR322
was mixed with 10 ml E. coli K1-14. The 20-ml mixture was fed to
50 third-instar larvae for 16 h. The mosquito larvae were subse-
quently washed many times with sterile distilled water before their
guts were dissected. The remaining larvae were further maintained
in sterile distilled water either with or without food supplement.
After a peroid of 2 or 7 days, the mosquito guts were dissected.
Each gut content was diluted with LB broth and spread on both LB
agar plates, to obtain the total number of bacterial cells per gut,
and on LB agar plates containing tetracycline (25 lg/ml) to detect
the re-introduced bacteria. The plates were incubated at 30 °C for
24 h, and the bacterial colonies were counted and identi®ed. An11/
2 G1 and E. coli K1-14 showed distinct colony morphology. Tests
were performed in triplicate.
Flotation determination
Bacteria were grown in LB broth at 30 °C for 16 h. The cells were
harvested by centrifugation, washed and adjusted to A600 ˆ 1.0
with water. A 2-ml sample of cell suspension was placed in a
12 ´ 75-mm sterile glass tube. The tubes were kept at room tem-
perature and photographed every 2 days.
Western blot analysis
Bacterial cells, grown in LB broth at 30 °C to A600 ˆ 1, were
harvested by centrifugation and subjected to 10% sodium dodecyl
sulphate/polyacrylamide gel electrophoresis (SDS-PAGE) by a
previously described method (Laemmli and Favre 1973). E. coli
JM109 harbouring pMU388, grown in LB broth at 37 °C to
A600 ˆ 0.3, were induced with 1 mM isopropyl-b-D-thiogalacto-
pyranoside (IPTG) for 4 h.
Antisera against CryIVB protein of B. thuringiensis subsp.
israelensis or the 51-kDa polypeptide of B. sphaericus were ob-
tained by subcutaneous injection of the proteins into rabbits.
Speci®city of the antisera was con®rmed by Western blotting,
carried out essentially as described (Towbin et al. 1979). Antibody
binding was detected by alkaline-phosphatase-conjugated goat
anti-(rabbit immunoglobulin G) (Sigma).
Mosquito larvicidal assays
Bacterial cells were grown in LB broth at 30°C for 16 h. Cells
were harvested and washed three times and a series of decimal
dilutions (1 ´ 108 to 1 ´ 103 cells/ml) was made in sterile distilled
water. Ten second-instar larvae of mosquitoes were added to 2 ml
cell suspension. Tests were performed in duplicate. The 50% lethal
concentration (LC50) values were calculated by the log probit
analysis method (Finney 1971) after 24 h incubation at room
temperature.
Stability of An11/2 G1 transformants
Single colonies of An11/2 G1, carrying recombinant plasmids, were
obtained by plating on LB agar and incubated at 30 °C overnight.
Fifty single colonies were then subcultured weekly on LB agar with
and without ampicillin (100 lg/ml). The presence of the plasmids
was determined by comparing the number of colonies on the two
LB agar plates.
Results
Identi®cation of the isolated An11/2 G1
One of the dominant bacteria, isolated from the guts of
An. dirus larvae, was designated An11/2 G1. On the

basis of biochemical reactions, the gram-negative An11/
2 G1 was identi®ed as Enterobacter amnigenus.
Recolonization of E. amnigenus in the guts
of mosquito larvae
In order to investigate whether E. amnigenus was able to
recolonize in the larval gut environment, third-instar
laboratory-reared Ae. aegypti, An. dirus and Cx. quin-
quefasciatus larvae were fed with a cell mixture of
E. amnigenus harbouring pBR322 and E. coli K1-14,
which was used as an internal control. The tetracycline-
resistance gene in pBR322 was used as a marker to
distinguish between the re-introduced and the resident
E. amnigenus. Results showed no E. coli K1-14 growing
on plates containing tetracycline since day 1, indicating
that it was not able to survive in the guts of the three
species of mosquito larvae tested. The number of
E. amnigenus in the guts of An. dirus was stable for at
least 7 days after the feeding, whereas that in the guts of
Ae. aegypti and Cx. quinquefasciatus decreased dramat-
ically after day 1 and continued to fall until day 7
(Fig. 1). These results showed that E. amnigenus was
able to recolonize in the guts of An. dirus but not in
those of Ae. aegypti and Cx. quinquefasciatus larvae.
However, since the experiments were carried out under
conditions in which the mosquito larvae were without
food supplement after feeding with the bacteria, we
further tested whether the recolonization was due to the
e€ects of fasting. Thus, after being fed with the bacteria,
the mosquito larvae were continuously fed a mosquito
larval diet. Results revealed that, although the number
of the bacteria reduced slightly, E. amnigenus was able to
survive in the larval gut (Fig. 2). This indicated that
Fig. 1 Recolonization of Enterobactor amnigenus in the guts of Aedes
aegypti, Anopheles dirus and Culex quinquefasciatus larvae. The
mosquito larvae were fed with E. amnigenus harbouring pBR322 and
Escherichia coli K1-14. After a period of 1, 2, or 7 days without food
supplement, the mosquito guts were dissected to determine the
number of bacteria (see Materials and methods). The numbers
represent E. amnigenus harbouring pBR322 in triplicate tests. Vertical
lines standard deviation. No E.coli K1-14 was detected from day 1
81
Fig. 2 Recolonization of E. amnigenus in the guts of An. dirus larvae
in conditions with or without food supplement. An. dirus larvae were
fed with E. amnigenus harbouring pBR322. After a period of 1, 2, or 7
days with or without food supplement, the mosquito guts were
dissected to determine the number of bacteria (see Materials and
methods). Vertical lines standard deviations of the triplicate tests. No
E. amnigenus harbouring pBR322 was detected in unfed control
cultures
recolonization of E. amnigenus was not an e€ect of
fasting.
Floating characteristic of E. amnigenus
A comparison of the abilities of E. amnigenus, B. thurin-
giensis subsp. israelensis and B. sphaericus to ¯oat is
shown in Fig. 3. The cells of B. thuringiensis subsp.
israelensis sedimented completely in 2 days. Half of the
B. sphaericus cells sedimented in 6 days, while E. amni-
genus cells only slightly sedimented. Good ¯otation was
an evident characteristic of E. amnigenus and allowed the
bacterium to remain near the water surface in the larval
feeding zone.
Expression of mosquito larvicidal protein genes
in E. amnigenus
Since E. amnigenus was able to ¯oat and to recolonize in
the guts of An. dirus larvae, it is a potential host for
producing mosquito-larvicidal proteins. To achieve this
potential, plasmid pMU388 harbouring the cryIVB gene
of B. thuringiensis subsp. israelensis and plasmid pMEx-
5142 harbouring the binary toxin genes encoding the
51-kDa and 42-kDa mosquito-larvicidal proteins of
B. sphaericus were separately electroporated into
E. amnigenus cells. Figure 4A shows that E. amnigenus
harbouring plasmid pMU388 produced a protein of
128 kDa at a high level that was detectable by SDS-
PAGE and Coomassie blue staining, and reacted
strongly with antiserum raised against the CryIVB pro-
tein (Fig. 4B, lane 2). The level of the 128-kDa protein in
82

product was too low to be detected in SDS-PAGE

stained with Coomassie blue (Fig. 5A).

Mosquito-larvicidal activity of recombinant

E. amnigenus

E. amnigenus harbouring pMU388 was toxic to larvae of

Ae. aegypti and An. dirus. Its LC50 was not signi®cantly
di€erent from that of E. coli (Table 1). This result
agreed well with the similar level of CryIVB protein
produced in E. amnigenus and in E. coli (Fig. 4).
E. amnigenus harbouring pMEx-5142 was toxic to larvae
of An. dirus and Cx. quinquefasciatus, although its tox-
icity was slightly less than that of B. sphaericus
(Table 1).

Stability of recombinant plasmids in E. amnigenus

Fifty single colonies of E. amnigenus carrying recombi-

nant plasmids were subcultured weekly on LB agar with
and without ampicillin. After 23 weekly subcultures, the
numbers of E. amnigenus colonies containing plasmid
pMU388 or pMEx-5142 on the two agar media were the
same (data not shown). The result indicated that the
plasmids were stable in E. amnigenus for at least 23
generations.

Fig. 3a±c Flotation characteristics of E. amnigenus, Bacillus thurin-

giensis subsp. israelensis and Bacillus sphaericus. A cell suspension of
(a) E. amnigenus, (b), B. thuringiensis subsp. israelensis and (c)
B. sphaericus was placed in glass tubes and allowed to stand for the
indicated number of days
E. amnigenus was as high as that in E.coli induced with
IPTG (Fig. 4, lanes 2 and 5).
E. amnigenus carrying plasmid pMEx-5142 produced
a protein of 51 kDa that was recognized by antiserum
raised against the 51-kDa protein of B. sphaericus
(Fig. 5B, lanes 2 and 3). However, the 51-kDa protein
Discussion
We found that E. amnigenus was a dominant bacterium
in the guts of Anopheles mosquito larvae. The genus
Enterobacter is found widely distributed in nature, oc-
curring in fresh water, soil, plants and animal and hu-
man faeces (Holt et al. 1994). The presence of
E. amnigenus in a mosquito larva gut has not been
previously reported, and no data on bacterial survival in
the guts of living mosquito larvae are available. It has
been shown that the spores of B. thuringiensis subsp.
israelensis can germinate in the midgut of mosquito

Fig. 4A, B Expression of the

cryIVB gene of B. thuringiensis
subsp. israelensis in E. amni-
genus. A 10% sodium dodecyl
sulphate polyacrylamide gel
stained with Coomassie blue.
B Western blotting, using an-
tiserum raised against CryIVB
protein. Lanes: 1, 2, proteins
of E. Calimnigenus harbouring
pUC12 vector and pMU388
respectively; 3 proteins of
E. coli JM109 harbouring
pUC12; 4, 5 proteins of E. coli,
harbouring pMU388 without
and with induction respectively
 83

Fig. 5A, B Expression of the

51-kDa protein gene of B.
sphaericus in E. amnigenus. A
10% sodium dodecyl sulphate
polyacrylamide gel stained with
Coomassie blue. B Western
blotting, using antiserum raised
against the 51-kDa protein of
B. sphaericus. Lanes: 1 proteins
of E. amnigenus harbouring
pMEx-8 vector; 2, 3 proteins of
E. amnigenus harbouring
pMEx-5142; 4 proteins of
B. sphaericus 2297

Table 1 Mosquito-larvicidal activity of recombinant Enterobacter 24 h). These numbers represent two independent experiments. ND
amnigenus. LC50 50% lethal concentration (expressed as number of not determined
cells per millilitre that had killed 50% of the second-instar larvae at

Bacterial strain (plasmid) Larvicidal component (kDa) LC50 (cells/ml)

Ae. Aegypti An. dirus Cx. quinquefasciatus

E. amnigenus (no plasmid) None Non-toxic Non-toxic Non-toxic

E. amnigenus (pMU388) 128 7.0 ´ 105 8.7 ´ 105 ND
E. coli JM109 (pMU388)a 128 3.5 ´ 106 7.5 ´ 105 ND
E. amnigenus (pMEx-5142) 51 + 42 ND 3.2 ´ 106 6.0 ´ 106
B. sphaericus 2297 51 + 42 >108 1.0 ´ 105 <103
a
Induced with isopropyl b-D-thiogalactopyranoside

larvae cadavers (Aly et al. 1985). E. amnigenus does not
form spores, its presence and long-lasting survival in the
gut supported its ability to recolonize.
The cryIVB gene in pMU388 was under the native
cryIVB promoter (Angsuthanasombat et al. 1987).
While the binary toxin gene in plasmid pMEx-5142 was
under the tac promoter (Boonserm 1996). Together,
these results indicated that the native cryIVB promoter
of B. thuringiensis subsp. israelensis and the tac pro-
moter of E. coli functioned eciently in E. amnigenus.
However, there were numerous bands in addition to the
128-kDa protein (Fig. 4B). These bands were degrada-
tion products, which had been observed in E. coli (An-
gsuthanasombat et al. 1987).
Although the expression of the 42-kDa protein in
E. amnigenus carrying pMEx-5142 was not determined,
the 51-kDa protein of B. sphaericus was detectable. Since
the binary toxin genes of B. sphaericus were in a single
transcriptional operon, this suggested that the 42-kDa
protein, down stream of the 51-kDa protein, was also
expressed. Since both polypeptides of B. sphaericus are
required for larvicidal activity (Broadwell et al. 1990),
the toxicity of the recombinant E. amnigenus implied
that the 42 kDa polypeptide was also produced.
The recolonization of E. amnigenus was speci®c to
An. dirus. It did not survive in the gut of Ae. aegypti nor
in Cx. quinquefasciatus larvae. This property, although
possibly limiting its wide applications, should allow a
speci®c application in An. dirus, a vector of malaria
parasites (Baimai et al. 1988). The bacteria capable of
recolonizing the guts of all three mosquito species are
being isolated. Whether the recolonization of E. amni-
genus plays any role in larvicidal activity will be further
investigated by feeding a sublethal concentration and
observing the mortality after 3±7 days.
Acknowledgements We thank the Miscellaneous Bacteriology
Section, National Institute of Health, Nonthaburi, for identifying
An11/2 G1, and Ms. Somsri Saijam for technical assistance. This
research was supported by a grant from the National Centre for
Genetic Engineering and Biotechnology, National Science and
Technology Development Agency, Bangkok, Thailand. W. Ch. is a
recipient of a grant from the Thailand Research Fund (TRF). S.P.
is a TRF Senior Research Scholar.
References
Ahmad S, Selvapandiyan A, Bhatnagar RK (1998) Increased tox-
icity of modi®ed mosquitocidal binary toxins of Bacillus
sphaericus expressed in Escherichia coli. Appl Microbiol Bio-
technol 49: 164±167
Aly C, Mullar MS, Federici BA (1985) Sporulating and toxin
production by Bacillus thuringiensis var. israelensis in cadavers
84
of mosquito larvae (Diptera: Culicidae). J Invert Pathol 46:
251±258
Angsuthanasombat C, Panyim S (1989) Biosynthesis of 130-kilo-
dalton mosquito larvicidal in the cyanobacterium Agmenellum
quadruplicatum PR-6. Appl Environ Microbiol 55: 2428±2430
Angsuthanasombat C, Chungjatupornchai W, Kertbundit S, Lux-
ananil P, Settasatian C, Wilairat P, Panyim S (1987) Cloning
and expression of 130-kd mosquito-larvidal d-endotoxin gene of
Bacillus thuringiensis var. israelensis in Escherichia coli. Mol
Gen Genet 208: 384±389
Baimai V, Kijchalao U, Sawadwongporn P, Green CA (1988)
Geographic distribution and biting behaviour of four species of
Anopheles dirus complex in Thailand. Southeast Asian J Trop
Med Public Health 19: 151±161
Baumann L, Broadwell AH, Baumann P (1988) Sequence analysis
of the mosquitocidal toxin genes encoding 51.4- and 41.9-kilo-
dalton protein from Bacillus sphaericus 2362 and 2297. J Bac-
teriol 170: 2045±2050
Baumann P, Clark MA, Baumann L, Broadwell AH (1991) Bacillus
sphaericus as a mosquito pathogen: properties of the organism
and its toxins. Microbiol Rev 55: 425±436
Bolivar F, Rodriguez RL, Greene JP, Betlach MC, Heyneker HL,
Boyer HW, Crosa J, Falkow S (1977) Construction and char-
acterization of new cloning vehicles. II. A multipurpose cloning
system.Gene 2: 95±113
Boonserm P (1996) Genetically engineering an active hybrid mos-
quito-larvicidal toxin. M.Sc. thesis in Biochemistry, Faculty
Graduate Studies, Mahidol University, Bangkok, Thailand
Broadwell AH, Baumann L, Baumann P (1990) The 42- and 51-
kilodalton mosquitocidal proteins of Bacillus sphaericus 2362:
construction of recombinants with enhanced expression and
in vivo studies of processing and toxicity. J Bacteriol 172: 2217±
2223
Chilcott CN, Ellar DJ (1988) Comparative toxicity of Bacillus
thuringiensis var. israelensis crystal proteins in vivo and in vitro.
J Gen Microbiol 134: 2551±2558
Chungjatupornchai W (1990) Expression of mosquitocidal-protein
genes of Bacillus thuringiensis subsp. israelensis and herbicide
resistance gene bar in Synechocystis PCC6803. Curr Microbiol
21: 283±288
Chungjatupornchai W, Hofte H, Seurinck C, Angsuthanasombat
C, Vaeck M (1988) Common features of Bacillus thuringiensis
toxins speci®c for Diptera and Lepidoptera. Eur J Biochemistry
173: 9±16
Delecluse A, Poncet S, Klier A, Rapaport G (1993) Expression of
cryIVA and cryIVB genes, independently or in combination, in
a crystal-negative strain of Bacillus thuringiensis subsp. israel-
ensis. Appl Environ Microbiol 59: 3922±3927
Finney D (1971) Probit analysis, 3rd edn. Cambridge University
Press, London
Hofte H, Whiteley HR (1989) Insecticidal crystal proteins of
Bacillus thuringiensis. Microbiol Rev 53: 242±255
Holt JG, Krieg NR, Sneath PHA, Staley JT, Williams ST (eds)
(1994) Bergey's manual of determinative bacteriology, 9th edn.
Williams & Wilkins, Baltimore, Md, pp 178
Laemmli UK, Favre M (1973) Maturation of the head of bacte-
riophage T4. I. DNA packaging events. J Mol Biol 80: 575±599
Liu J-W, Yap WH, Thanabulu T, Porter AG (1996) Eciency
synthesis of mosquitocidal toxins in Asticcacaulis excentricus
demonstrates potential of gram-negative bacteria in mosquito
control. Nat Biotechnol 14: 343±347
Mahillon J, Chungjatupornchai W, Decock J, Dierickx S, Michiels
F, Peferoen M, Joos H (1989) Transformation of Bacillus
thuringiensis by electroporation. FEMS Microbiol Lett 60: 205±
210
Murphy RC, Stephens SE Jr (1992) Cloning and expression of the
cryIVD gene of Bacillus thuringiensis subsp. israelensis in the
cyanobacterium Agmenellum quadruplicatum PR-6 and its re-
sulting larvicidal activity. Appl Environ Microbiol 58: 1650±
1655
Priest FG (1992) Biological control of mosquitoes and other biting
¯ies by Bacillus sphaericus and Bacillus thuringiensis. J Appl
Bacteriol 72: 357±369
Porter AG, Davidson EW, Liu J-W (1993) Mosquitocidal toxins of
bacilli and their genetic manipulation for e€ective biological
control of mosquitoes. Microbiol Rev. 57: 838±861
Sangthongpitag K, Penfold RJ, Delaney SF, Rogers PL (1997)
Cloning and expression of the Bacillus sphaericus 2362 mosqui-
tocidal genes in a non-toxic unicellular cyanobacterium,
Synechococcus PCC6301. Appl Microbiol Biotechnol 47: 379±
384
Thanabalu T, Hindley J, Brenner S, Oei C, Berry C (1992) Ex-
pression of the mosquitocidal toxins of Bacillus sphaericus and
Bacillus thuringiensis subsp. israelensis by recombinant
Caulobacter crescentus, a vehicle for biological control of
aquatic insect larvae. Appl Environ Microbiol 58: 905±910
Towbin H, Staehelin T, Gordon J (1979) Electrophoresis transfer
of proteins from polyacrylamide gels to nitrocellulose sheets:
procedure and some application. Proc Natl Acad Sci USA 76:
4350±4354
Wu D, Chang FN (1985) Synergism in mosquitocidal activity of 26
and 65 kDa proteins from Bacillus thuringiensis subsp. israel-
ensis crystal. FEBS Lett 190: 232±236
Yap WH, Thanabalu T, Porter AG (1994) Expression of mosqui-
tocidal toxin genes in a gas-vacuolated strain of Ancylobacter
aquaticus. Appl Environ Microbiol 60: 4199±4202
Yousten AA (1984) Bacillus sphaericus: microbiological factors
related to its potential as a mosquito larvicide. Adv Biotechnol
Process 3: 315±343