In vivo generation of hybrids between two Bacillus thuringiensis insect-toxin-encoding genes

Gene, 98 (1991) 37-44

37

Elsevier

GENE

03875

In vivo generation of hybrids between two Bacillus thuringiensis (Recombinant

DNA;

b-endotoxin;

parasporal

T. Caramori”,

A.M. Albertini”7b and A. Galizzi”

insect-toxin-encoding

genes

crystal)

‘IDipartimento di Genetica e Microbiologia ‘A. Buzzati-Traverso’, Universitd degli Studi di Pavia, 27100 Pavia (Ita1.v); and ” Facoltci di Agraria, Istituto di Produzione

Animale,

UniversitLj degli Studi di Udine, 33030

Received by H.M. Krisch: 9 November Revised: 22 August 1990 Accepted: 3 October 1990

Udine (Italy) Tel. (39-432)660810

1989

SUMMARY

The parasporal crystal of Bacillus thuringiensis is composed of polypeptides highly toxic to a number of insect larvae. The structural genes (crylA) encoding the Lepidoptera-specific toxin from different bacterial strains diverge primarily in a single hypervariable region, whereas the N-terminal and C-terminal parts of the proteins are highly conserved. In this report, we describe the generation of hybrid genes between two crylA genes. Two truncated crylA genes were cloned in a plasmid vector in such way as to have only the hypervariable region in common. The two truncated cryIA genes were separated by the tetracycline-resistance determinant (or part of it). In vivo recombination between the hypervariable regions of the crylA genes reconstituted an entire hybrid cr_yIA gene. Direct sequence analysis of 17 recombinant plasmids identified eleven different crossover regions which did not alter the reading frame and allowed the production of eight different hybrid proteins. The recombination events were independent from the RecA function of Escherichia coli. Some of the hybrid gene products were more specific in their insecticidal action and one had acquired a new biological activity.

INTRODUCTION

Bacillus thuringiensis is a Gram+, spore-forming bacterium capable of producing proteins that accumulate as crystals during sporulation. The proteins are known as &endotoxin and are toxic to a number of insect larvae. Different strains of B. thuringiensis subsp. kurstaki produce crystal proteins of different aa sequence that can show

Correspondence biologia,

to; Dr. A. Galizzi,

Via S. Epifanio

Dipartimento

14, 27100 Pavia

di Genetica

(Italy)

e Micro-

Tel. (39-382)303852;

Fax (39-382)32234. Abbreviations: pair(s); proteins; ORF,

aa. amino

acid(s);

Cm, chloramphenicol;

Ap, ampicillin;

cry, genes

E., Escherichia; kb, kilobase

open reading

frame;

PAGE,

B., Bacillus; bp, base

encoding

insecticidal

polyacrylamide-gel

electrophoresis;

Pollk, Klenow (large) fragment of E. coli DNA polymerase resistance/resistant; SDS, sodium dodecyl sulfate; Tc, tetracycline; ultraviolet;

0378-l

[ 1, denotes plasmid-carrier

I19~91~$03.50

0

1991

crystal

or 1000 bp; nt, nucleotide(s); I; R, UV,

state.

Elsewer Science

Puhhshers

B.V.

(Biomedical Division)

selective toxicity against larvae of several species of Lepidoptera. The comparison of the nt sequences of a number of cloned genes encoding the toxin protein shows that all genes are homologous, that the degree of homology varies among different genes and, more interesting, that the distribution of differences is not random over the length of the coding sequence but tends to be clustered in a so-called hypervariable region (Geiser et al., 1986; Wabiko et al., 1986; Andrews et al., 1987). Since different toxin proteins present specific biological activities (Hofte et al., 1988) it can be inferred that part of the specificity of action of the crystal proteins resides in the primary structure of the hypervariable region. To generate new protein variants with potentially new biological activities, we have taken advantage of in vivo intramolecular recombination to produce hybrid genes (Weber and Weissmann, 1983; Rey et al., 1986). To create recombinants we constructed a plasmid carrying two truncated toxin genes, overlapping only in their hypervariable

38

pT73 11.2kb

J E

-pUBllD\

13kb

El

ori

-

ori Fig. I. Construction ofpBR322

of plasmid

with EcoRI + Sac1 (blunted). (Kronstad

and Whiteley,

et al., 1985). Plasmid (BarnHI-Hind111 (Schnepf

pT173. Plasmid

pT is a derivative

(heavy black arrow). A 1424-bpEcoRI-AvrrI pT73 is a pT derivative

1986). The fragment

pTI was obtained

blunted)

pESAC

of cryIA/u) as a 5.2-kb fragment, at the Aval site by treatment

obtained

by inserting

as follows: pT was digested and corresponding

was a derivative

obtained

from pES1 (Schnepf

nt 291 to 2215. The direction A, Awl;

B, BarnHI;

E, EcoRI;

of transcription H, HindIII;

is indicated

with SmaI,

and Whiteley,

partially

obtained

from pJWK20

from nt 1383 of the sequenced

region (Adang

with BumHI

1981) following of the 2500-bp

HincII + AvaI digestion. BumHI

fragment

The dashed

and ligated to a 1.9-kb fragment

The fragment

of pT1. comprising

region

arrows

indicate

box represents

the region of partial

was made blunt

the first part of the

The open box is the cryIA(c) toxin-coding

from the toxin gene. The hatched

by the arrows.

the TcR determinant

was inserted in PBS 19 digested

1985) in which, in the SmuI site, was cloned the entire gene

cleaved with BarnHI.

downstream

containing

of the cryIA(a) gene from nt 291 to 2215 of the sequenced

(Fort and Errington,

from the insertion

the HD-73 sequences

ofpBR322

in the EcoRI site of a pT a 5.4-kb EcoRI fragment

to completion

cr~‘lA(u) gene and the last two thirds ofthc TcR gene. in pT73 completely

unpublished),

bearing the TcR determinant

to the first portion

of pSGMU2

with the PolIk. pT173 derived

from nt 1383. The heavy line represents

(Wells et al., 1983; G. Gray,

the last two thirds of the cryIA(c) gene starting

comprises

derived from pESAC

et al., 1985). Plasmid

ofpBS19

(blunted inAi’a1) fragment

sequence,

starting

the crylA(aJ sequence,

from

homology.

S, Srrc,I; Sm. S,,lcrI.

Fig. 2. Restriction maps of plasmids pT173 and pGEM-173. pGEM-173 was derivative of pGEM-42 (Promega, Madison, WI, USA). The cryIA(c/ truncated gene was first cloned in pGEM-4Z as an EcoRI fragment obtained from pJWK20. The cryIA(uJ moiety was derived from pTl73 as a BomHI fragment,

carrying

the promoter

region, the proximal

the plasmid derived from the vector For other symbols see Fig. I.

(pBSI9

for pTl73

portion

of the crqIA(a) gene and part of TcR determinant.

and pGEM-42

for pGEM-173).

Restriction

The thin line represent

site abbreviations:

Hi, HincII;

the portion

of

N, NrruI: K, KI”II.

39 region and separated by an nt sequence containing a unique restriction site. The two truncated genes were cryZA(u)

derived from strain HD-lDipe1,

consisted

region, the 5’-coding

and a large portion

(Schnepf et al., 1985) and crylA(c) (Kronstad and Whiteley, 1981; Adang et al., 1985) that show 63% homology at the nt level in the hypervariable region. Both genes belong to the cryZA subgroup, according to the nomenclature of Hiifte and Whiteley (1989).

hypervariable region. The second gene, cryIA(c), derived from strain HD-73, consisted of the hypervariable region and the 3’-terminal coding sequence. The two truncated genes were separated by the TcR determinant or part of it. The residual homology of the hypervariable region (63% at

RESULTS

AND

(a) Bacterial strains and plasmid construction E. coli HBlOl (F- h&S20 recA13 am-14 proA leuB6 lacy1 galK2 rpsL20 xyl-5 mtl-1 srpE44); JM103 {A(/ac-

of the

through several steps as described in Fig. 1. Plasmids pT1 and pT173 coded for a truncated polypeptide in the range of 65-68 kDa which immunoreacted with antibodies raised

proAl?) supE thi strA endA hspR4 [F’ traD36 proA’B” lacfQ ZA M 151) ; 294 (endA thi pro hsdR hsdM hsm); 294recA (endA thi pro hsdR hsd~ hsm recA f were used for transformation according to Hanahan (1985). The medium used was LB medium (Difco tryptone, 10 g/yeast extract, 5 g/NaCl, 10 g/water to 1 liter, pH 7.1). For growth of strains with plasmid pT173, 12.5 pg Tc/ml or 10 pg Cm/ml were added to the medium; for growth of strains with plasmid pGEM-I 73 we used 100 pg Ap/ml. The recA mutation was tested by streaking cells on M9 minimal plates (Miller, 1972) and irradiating with a wavelength of 254 nm for 120 and 180 s at a distance of 20 cm (UV lamp model UVSL-58; UVP Inc., San Gabriel, CA). As a control, half the agar plate was masked with cardboard during UV exposure. We constructed two plasmids for the in vivo generation of recombinants between the two related B. thuringiensis genes, cryIA{ai and cryIA(c). Both plasmids carried two truncated genes of B. th~ri~~gie~z~~~s; one gene, qvfA(o),

200

of the promoter

the nt level) should be enough to promote in vivo recombination Plasmid pT173 was a derivative of pBS19 (Wells et al., 1983; G. Gray, unpubl.), contained the entire TcR gene of pBR322 inserted in an inverted orientation with respect to the promoter region of the cryIA(u) gene. It was constructed

DISCUSSION

kDa

sequence

against the pure toxin crystal of strain HD-73 (data not shown). Plasmid pT173 was designed with the purpose of having a shuttle vector, capable of replication in E. coli and B. subtilis (and potentially 3. thuringie~sis). and the possibility of selecting for Tc-sensitive recombinants in E. co& It was used in a set of experiments and a number of recombinants were obtained (see below). We encountered some difficulties in the selection of CmK colonies of E. coli[pT173] and its recombinant derivatives. We thus constructed a second plasmid, pGEM-173, based on pGEM-4Z, allowing selection for ApR in E. coli (Fig. 2). For purpose of construction the latter plasmid lacked a complete TcR determinant (Fig. 2). Recombinant clones derived from pT173 or pGEM-173 could be enriched, by digestion with iVru1, for which there is a unique site in the TcR gene or the portion of it conserved in pGEM-173.

A

6

123456789

123456

7

8

9

-

97684326-

Fig. 3. Enzyme immunologically.

immunoassay

of crystal

proteins

produced

in E. coli. The method

of Towbin

et al. (1979) was used to detect

the crystal

protein

of E. cofi cells resolved by 0.1 T0 SDS-8% PAGE were transferred eiectrophoretically to nitrocellulose sheets washed with sheets 50 mM Tris HCl/X)O mM NaCIIO.1 5, Nonidet P-40, and then incubated with the antiserum. After a wash with the same buffer, the nitroceilulose were incubated

Extracts

with peroxidase-conjugated

immunocomplexes were then visualized NY, Bethesda Research Laboratories, extracts

of E. coli containing

Lane 9 contains

a purified

plasmids sample

sheep anti-rabbit

immunoglobuhn

G antiserum

(United

States

Biochemical

Co., Cleveland,

Ohio). The

in the presence ofhydrogen peroxide and 4-chloro-I-naphthol as substrates (GIBCO Laboratories, Grand Island, Inc., Gaithersburg, MD). Panel A: immunoblots; panel B: Coomassie blue staining. Lanes 1 to 7 contain cell from pHyl

crystal

antigen

to pHy7,

respectively.

Lane 8 contains

an extract

from B. thuringiensis subsp. kurstuki HD-73.

of E. coli containing

the parental

plasmid

pT173.

40 (b) Generation of hybrid genes and selection of in vivo recombinant plasmids To generate recombinants, plasmid pT 173 or pGEM- 173 were introduced into E. coli 294 RecA + by transformation. One colony was inoculated in LB and grown overnight in

the presence of Cm in the case of pT173, or Ap in the case of pGEM-173. Plasmid DNA was extracted, digested with NruI to linearize the nonrecombinant molecules and used to transform E. coli recA cells. A number of transformants were obtained for each plasmid preparation. The transfor64-53-107

6-15 cryIA(a) cryIA(c)

1521 1383

A-~~CTTTTCC~TATATG~CT * t*t * ** *t l ** t**

*

*

***

l

C~~GTATTGTTG ** * l *

TGCAGCTCCA **********t*

cryIA(a)

1580

TT-AACTGGTTTGGGG-ATTTTTAGAACATTATTATCTTCACCTTTATATA~G~TTATAC

cryIA(c)

1442

CTCAACTAGGTCAGGGCGTGTATAGAACATTATCGTCCACTTTATATAG~GACCT---T *************t * l *** * * *** * * ********t*** **

*

cryIA(a)

1538

cryIA(c)

1499

TTGGTTCAGGCCCAAATAATCA~CTGTTTGT~CTTGATGG~~~AGTTTTCTTTTG TTAATATAGGGATAAATAATCAACAACTATCTATCTGTTCTTGACG~ACAG~TTTGCTTATG ** ** ** *** *** ** * *** ********* **** * l *t *****

cryIA(a)

1698

CCTCCCTAACGACCAACTTGCCTTCCACTATATATAGACAAAGGGGTACAGTCGATTCAC

cryIA(c)

1559

GAACCTCCTCA---AATTTGCCATCCGCTGTATACAGAAAAAGCGGAACGGTAGATTCGC *t * ** **et* *** l * t*** *** **** l * ** ** *****

**

*

127 2 TAGATG’&ATACCGCCACAGjXTAATAGT@$CCACCTCGTGCG@%AT=?j&~ATCGAT GATTTA CATCGAT TGGAT TAACAAC~CCACCTAGGC~ TACCGCCACA ******** ******* l *tt* *******t****t **** t l * ****** *

5

cryIA( a) cryIA(c)

1758 1616

cryIA(a)

1818

cryIA(c)

1676

32

21 T T l

**

******

*****

*

*

66-45

**

**

**

**

***

*

**

f

****

122

104

cryIA( a) cryIA(s)

1872 1736

cryIA(a)

1932

AAATTACACAAAThCCTTTAACAkAATCTACTAATCTTGGCTCTGGAACTTCTGTCGTTA

cryIA(c)

1796

GTATTACTCAAATCCCTGCAGTGAAG---GGAARCTTTCTTTTTAATGGTTCTGTAATTT ***et* ***** ***** *** t **

*i

3-4-7 cryIA(a) cryIA(c)

1992 1853

cryIA(a)

2050

ARCCTT~GAGTAART----ATTAC--TGCAC----CATTATTAT~A-~~--AGATATCGGG

cryIA(c)

1913

AGAATAGAGGGTATATTGRAGTTCCARTTCACTTCCCATCGACATCTACCAGATATCGAG ** * * *** *** ** * * ******** * * ** *** **

cryIA(

Fig. 4. Optimal indicate

a)

**

TAAGAATTCGCTACGCTTCTACTACAAATTTACATTCCATACATCAATTGACGG~G--

1973

TTCGTGTACGGTATGCTTCTGT~CCCCGATTCACCTCM--CGTTAATTGGGGTAATTC * ** ** * * * ***** * t * * *t ** *et*** **

cryIA(a)

2155

ACCTATT~TCAGW;TAATTTTTCAGCAACTATGA-GTAGTGGGAGT~TTTACAGT~CG

cryIA(c)

2031

ATCCATTTTTTCCAATACAGTACCAGCTACAGCTACGTCATTAGA-TAATCTACAATCAA

cryIA(a) cryIA(c)

2214 2090

Matches Length

= 456 = 722

alignment ofcryIA(u)

* * ***

*

**

Matches/length

(GenBank

AC No. M 11250) and

and to Adang et al. (1985)for

cryIA(c).

pGEM-4Z

and subsequently (Chen

A band corresponding digested

and Seeburg,

**

f **

*

** ****

**

****

**

with EcoRV.

1985). Sequcnase

percent

cryIAic) (GenBank

in the recombinant

to DNA

of approximately

A band of 0.7-kb Sequencing

(USB

DNA

genes

‘The numbering

from recombinant

Ohio,

2.9 kb was separated

region. Asterisks to Schnepf

by agarose gel eiectrophoresis, electroeluted.

by the chain termination USA)

in the hypervariable

of the nt is according

et al.

plasmids derived from pT173 and pGEM-173,

was purified by 67; PAGE,

was performed

Corp. Cleveland,

AC No. Ml 1068)

plasmids.

Fragments of DNA, originated

digested with SnrnI and treated with phosphatase. DNA

****

= 63.2

(1985) for ccvIA(u]

by Hind111 digestion.

*

*

GA GT

regions identified

to plasmid

*

2097

boxes indicate the crossover

were obtained

*

cryIA(c)

matches,

by electroelution

**

recovered

and ligated to the plasmid

method of Sanger et al. (1977) adapted

was used in the dideoxy-chain

elongation

reaction.

41 mation

efficiency

the expected restriction

was very low, in the range of 30 to 60

blotting,

transformants per pg of pGEM-173 DNA. For each transformant one single colony was used to prepare plasmid DNA that was screened for the presence of recombinants by BumHI and Hind111 digestion. Of 134 clones examined - the outcome of two independent experiments, one carried out with pT173 and the other with pGEM-173 - ten gave the BumHI and Hind111 restriction

Fig. 5. Optimal

393

CryIA( a) HY3,4,7 Hy104,122 HY66,45 HY32 HY127,21 HY2 HY5 HY126 HY6,53,64,107 CryIA( c)

452

Cr IA(a) HYY ,4,1 Hy104,122 HY66,45 HY32 HY127,21 HY2 HY5 HY126 HY6,53,64,107 CryIA( a)

513

due to the presence

react-

The exception was represented by culture samples containing the recombinant plasmid pHyl5 (see section c). (c) Sequence analysis of the recombinant plasmids Seventeen recombinant plasmids were sequenced between the two EcoRV sites in the hypervariable region. The 17 recombinant plasmids idenulied eleven crossover regions, distributed over the length of the hypervariable region (Fig. 4). Only in one case (pHyl5)

the crossover

was

-~---~-~----_gQ~~-~Q~-Q-VY-_-__-_-P^F~~

LTTNLPSTIYRQRGTVDSLDVIPPQDNSVPPRAGFSHRLSHVTM~LSQA~AGAVYTLRAPT _______________________-____-__--_---_--__--__-_______~___~__ __~~___________________-___--_---_--__-__________~~__~~__~-__ __~~___________________-___--__-__--__-___-______~~_~~~__~~__ ________--___---___-___-___--__-___-__-_____________________~ __-~____-____---__--___-___--__-__--__-___S_FR_GFSNSS~SII~~_~ __-~____--___-____--___-___--__-Q__-__-___S_FR_GFS~SS~SII~~_~ ----_---________-____-___N-N_--_Q-_______~S_FR~GFSNSS~SII-~~~ ~SS----AV_-KS___-___E-___N-N-_N_N_-__Q___~~__~~S~FR~GFSNSS-SII--~~ ^SS----AV--KS---__--E__--N_N-__-Q_--------S-FR-G~SNSS-SII---M ~~~----~~_-~~--_----E-_------Q-----__Q--_~~-~~~~~~~~~~~~~~-~~~-~~~

FSWQHRSAEFNNIIPSSQITQIPLTKSTNLGSGTSVVKGPGFTGGDILRRTSPGQISTLRV ---__---________-____-___-___--__-_______~__~~~~~~~~~-~~~Q~~~

------------_-A-DS_--__AV_GNF_FN-^--IS_-IS~~~~~~--LV-LN-S-NNIQN-G ---I----_---__A-DS_____AV_GNF_FN_^--IS__~~_~~~_~-~~~-~~-~-~~~Q~~~ ---~_--~__-~__~_~~__-__~V-GNF-FN-^--IS__~~_~~~_~~~~~-~~-~-~~~Q~-~ ---I----_---__A-DS_____AV_GNF_FN-^--IS__~~_~~~~~-~~~-~~-~-~~~Q~-~ __-I________--A_DS-__--AV_GNF_FN_^--IS__~~___~__~~~~~~~~~-~~~Q~~~

---I--------_-A-DS_-___AV_GNF_FN-^--IS--IS--------LV-LN-S-NNIQN-G ---I----_---__A-DS_____AV_GNF_FN-^--IS_-IS~~-~~--~LV-LN-S-NNIQN-G ---I----------A-DS-----AV-GNF-FN-^--IS--IS--------LV-LN-S-NNIQN-G _--I-__--_----A_DS--___AV_G~F_FN-^--IS--IS--_~--~~LV~~~S-~IQR~G

of polypeptides

NITAPLSQRYRV~RIRYASTTNLQFHTSIDGRPINQGNFSATMSSGSN*LQSGS Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWC--SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPAATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--S-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIHLNVNWG-SSI--NTVPATATS-DNLQ Y-EV-IHFPSTST-Y-VRVRYASVTPIRLNVNWG-SSI--~ATATS-DNLQ deduced

from the nt sequences

The recombinant

proteins

of cryZA(u) (NBRF are indicated

shown in Fig. 4. The aa residues identical in CryIA(a)

aa gaps introduced

by immuno-

polypeptide

FAFPLFGNAG~PV~LVSLTGLGIFRTLSSPLYRRIILGSGPNNQELPVLM;TEPSF~ __~~____________________-___-___-__-_____~__~~~_~~~~~--~~--~~ ________________________-___-_________-_______~__~~_~~-_~~-_~ ~~~~___________________--__--___-__-__-_______~__~~_~~-_~~-_~ ~~~~___~__________~~____~___~_________~___~______~~~~~~~~~~_~ _--__________________-_________________~__~~__~~~~~-~~~--~---~~ ~--~~~~~_____________---_---___-__-__________~__~~~_~~~_~~-__ __-____________________-__-____-__-__________~__~~~_~~~__~-__ --------__--_________--__-____--_____P*FNI_I_~_Q~S~_~~~_~AYGT --------_---__QQRI_AQL_Q_VY-___-T--_T-__-P^FNI_I---Q-S-------AYGT ----------KCSSTTTYCCSTRSGRV*

of their recombinants.

in the nt sequence

Carets represent is truncated

CryIA( a) HY3,4,7 HY104,122 HY66,45 HY32 HY127,21 HY2 HY5 HY126 HY6,53,64,107 CryIA(c)

alignment

from the nt sequences recombination

333

were analyzed

of a full-length

ing with polyclonal antibodies raised against purified &endotoxin. In all instances but one, the cell extracts contained a polypeptide antigen having an electrophoretic mobility similar to the pure 135-kDa crystal protein (Fig. 3).

pattern expected for recombinants. The majority of the other clones (105 out of 134) gave the same pattern of the parental plasmids and could be explained by incomplete NruI digestion prior to the transformation. A smaller fraction (13 out of 134) showed a complex restriction pattern and they were not analyzed any further. Samples of cultures of E. coli containing the plasmid with

CryIA( a) HY3,4,7 HY104,122 HY66,45 HY32 HY127,21 HY2 HY5 HY126 HY6,53,64,107 HY15 CryIA(c)

pattern

for the presence

as to obtain best alignment. The asterisk indicates

of a stop codon.

Numbering

of aa residues

of protein

EC No. A22617),

by Hy (HY)

followed

and in CryIA(c) the position CryIA

564

crJ,ZA(c) (NBRF by a number

or in the hybrid product

at which the protein coded

(a) is reported

for reference.

EC No. A23962)

and

that refers to the site of are denoted

by dashes.

by the plasmid pHyl5

42 unequal, giving rise to the addition of one base in a sequence of three contiguous G residues. In all other instances the

Hy122) the deduced though the crossover

recombination product was the exact juxtaposition of the two parental sequences. The regions of crossover resolution were distributed over the length of the partially homologous sequence, but were not completely random. In some regions of relatively long uninterrupted homology (15 or more con-

aa sequences were the same, even regions were different.

(d) The generation of recomhinants is RecA-independent Our work was originally based on the assumption that recombination between the partially homologous sequences of the hypervariable region would ensue by a RecAdependent mechanism. For this reason all experiments

tiguous nt), no recombinants were observed. Some crossover regions were represented more than once. We cannot draw any conclusion regarding the presence of possible hot spots of recombination since the experimental conditions did not ensure that all recombinants were independent iso-

were carried out on plasmid DNA obtained after a passage through an E. coli strain proficient in homologous recombination. We tested this assumption making plasmid preparations of pGEM- 173 from a single colony of a RecA _ (H B 10 I ) and from a single colony of a RecA + (JM 103) E. coli transformant. After digestion with NvuI and transformation we obtained a comparable number of transformants for the two plasmid preparations: 33 colonies per pg of DNA in the strain and case of plasmid DNA derived from the RecA

lates. As already observed by Weber and Weissmann (1983) the final resolution of the crossing-over event could be found in regions with as few as 2 nt of uninterrupted homology (e.g., pHy2 and pHy21). The aa sequences of the protein products, as deduced from the nt sequences, are reported in Fig. 5; the eleven recombinant sequences encoded nine different proteins. One, coded by the plasmid generated by unequal crossingover, was truncated early in the hypervariable region, due to a frame-shift. The remaining eight proteins were of full size and were hybrids with novel aa sequences, never reported for 6-endotoxins obtained from natural strains of B. thuringiensis. In three occurrences (Hy6 and Hy64; Hy 127 and Hy2 1; Hy 104 and

60 colonies per pg of DNA in the case of the RecA ’ -derived plasmid. Upon extraction and HirzdIII restriction analysis the two DNA samples gave the following result: five out of 23 DNA samples obtained from the RecA strain were recombinants, the rest gave the same restriction pattern as the original plasmid. In the sample derived from a RecA’ strain, three out of 26 plasmids were recombi-

\ Transformation

-Nrul digestion \

/

-0

I Transformation __.

Homologous

recombination Exonucleolytic

I

_______

__~ :::zo digestion

I Nrul digestion and transformation -

oBR322

.OriL$& Recombinants

Fig. 6. Two possible

pathways

for the generation

of in viva recombinants.

According

to one model, homologous

recombination

takes place during the

multiplication of plasmids. The Nrul digestion would thus act as a selective factor. In the second model the NrruI cleavage produces linearized plasmids that, upon introduction into the recipient cells. could be subjected to exonucleolytic digestion. The single stranded complementary ends could be substrates for recombination. The heavy closed boxes represent the partially homologous The dotted lines represent single-stranded DNA derived from exonucleolytic

regions, open boxes represent digestion.

CU~IA~CJ, and hatched

represent

crvlA(tr/.

43 nants

with the expected

complex

rearrangements

restriction and

pattern,

19 were

four showed

as the

parental

plasmid. From these data we are led to conclude that the formation of hybrid genes in our system does not require the RecA function of E. coli. Depending on the experimental design, recombination could occur at two stages; during multiplication of the plasmid, prior to the NruI digestion, or following transfection with the linearized plasmid (Fig. 6). In the former situation, restriction with NruI would serve the purpose of lowering the background. In the latter case the linearization of the plasmid would provide the recipient cells with a substrate that could be converted into single-stranded form and eventually undergo recombination. At present the experimental data do not allow to discriminate between the two alternative models. Nevertheless, it is noteworthy that in the RecA- background we only observed correct recombinants, whereas in the Ret-proficient strains, we often obtained plasmids showing complex rearrangements, in addition to the expected recombinants.

TABLE

1

Toxicity

of parental

and recombinant

clones against

four insect species

LDSO’

Plasmid a

Ephestia

Trichoplusin

Helioris

Spodoptera

kuehnieila

ni.

sp.

littorcrlis

pES1

-

4.78

1.0

pJWK20

43.80 -

0.25

0.48

-

-

-

-

-

-

PHYI~ pHy127

-

-

-

-

-

-

-

-

PHY6

42.90

< 0.7

<0.2

PHY64

49.18

PHY~

-

PHYLA pHy45 (66) pHy32 pHy104

-

pHy122

-

PHY2 PHY~ (4.7)

‘I E. coli strains nant plasmids

0.45

-

-

0.25

-

-

2.69

-

-

-

0.68

-

-

-

0.15

-

0.27

0.46

> 20.0

0.21

0.45

> 20.0

0.13

HBlOl

or 294recA

-

containing

were grown overnight

the parental

in LB medium

or recombi-

supplemented

with

10 pg Cm/ml (for pHy2, pHy3, pHy5 and pHy6) or 100 p’g Tc;ml (all other

(e) Toxicity of hybrid gene products The hybrid gene products were tested for toxicity against four insect species (Table I). Four of the proteins (Hy2, 3, 15 and 127) did not show any insecticidal activity. In the case of Hy15 the loss of activity against lepidopteran larvae was predicted, since the recombination process generated a stop codon and the deduced gene product was a truncated protein. The other three instances of inactive proteins can be explained with the disruption of structural domains necessary for biological activity. An analogous situation has been described by Ge et al. (1989). Two hybrids (Hy6 and Hy64) gave the same toxicity range of the protein encoded by the parental gene crylA(c). The two hybrid genes Hy6 and Hy64 had arisen by exchange in two adjacent regions and the deduced proteins had the same aa sequences. In addition they were due to exchanges very early in the partially homologous region and the gene product was almost identical to the product of the cryIA(c) gene. Four hybrid proteins showed increased specificity in their toxicity; Hy5, Hy21 and Hy45 were active only against larvae of Heliotis sp., whereas Hy32 was very active but only against T.ni. Finally, two proteins (Hy 104 and Hy 122) with the same aa sequence but encoded by two plasmids generated by two different recombinational events, acquired an entirely new activity, being active, albeit at relatively high concentration, against larvae of Spodoptera littoralis (Noctuidae). This finding is particularly interesting since the N-terminal domain of the toxin active against S. littoralis is significantly different (45 y0 identity) from that of toxins active against most of the Lepidoptera (Sanchis et al., 1989). The hybrid proteins could represent an ideal tool to

clones). Cells were harvested priate concentration. by gel electrophoresis immunoblotting.

and resuspended

Plasmid

content

and the presence

The numbers

biological shown

activities

was checked ascertained

refer to recombinant

as the first listed plasmid

of the coded

proteins

by plas-

(Fig. 4). The

were comparable

to those

for the first plasmid.

’ LD50 is the dose required values

of the toxin

in parenthesis

mids with the same nt sequence

in 0.85”” NaCl at appro-

of each preparation

are reported

diet. LD50 consisted

were calculated

symbols

caused

no mortality.

of cell suspension

with a probit

tested

applied

analysis

program.

and the

to artificial Controls

E. coli strains HB 10 I and 294recA. The minus

ofuntransformed

(dash)

to kill 50”” of the insects

as 0Dsh5,,

indicate

that the highest

Methods:

Neonate

amount

tested

(> lOOOD/g)

larvae of Trichoplusia

ni, Heliotis

sp. and Spodoptera littorulis were reared on a diet described bY Shorey and Hale (1965) supplemented with IO”, of the E. coli suspension. All bioassays were performed above

and below

bioassays)

and mortality were conducted

kuehniella

pension

in duplicate

the approximate

to

was taken

with 18 larvae for each concentration LD50

(determined

adding

2 ml of appropriately

1g wheat flour and 1 g ground maizemeal; to 5 ml by addition

30 min, few drops were spread and air-dryed.

in preliminary

was scored after one week. Bioassays

of distilled

water

ofEphestia

diluted cell susthe final volume

and, after stirring

for

on the bottom of cups of 2 cm of diameter

In each cup 50 eggs were placed and mortality

was scored

after one week of incubation.

locate the &endotoxin protein domains responsible for the specificity of action (Widner and Whiteley, 1990). The in vivo recombination described in the present paper is also suggestive of a possible mechanism by which new toxin genes may be generated in nature, by means of recombination between genes with different variable regions. A similar hypothesis has been proposed by Geiser et al. (1986) to account for the large number of similar genes present in the same or in different strains of B. thuringiemis.

44 (f) Conclusions

Hofte, H., Van Rie, J., Jansens,

S., Van Houtven,

and Vaeck, M.: Monoclonal

(1) Hybrid between two B. thuringiensis insect-toxinencoding genes were produced by in vivo recombination.

antibody

A., Vanderbruggen,

analysis

trum of three types of lepidopteran-specific teins of Bacillus

(2) The recombination was independent of the RecA function of E. coli. (3) The hybrid gene products had new biological specificities.

H.R.:

thuringienensis. Microbial.

Kronstad,

crystal

Laboratory,

crystal

pro-

Microbial.

54 (1988)

proteins

of Bucil1u.s

Bacillus

in Molecular Harbor,

C., Mainzer,

Sanchis.

V., Lereclus,

terminating

and

sequence

E., Lad,

A.T. and Hoch, J.A. (Eds.). Applications.

J., Gud, S. and Lecadet.

and analysis active

of the N-terminal

delta-endotoxin

S. and Coulson,

inhibitors.

Harbor

of BcrciNus and gencra-

Biotechnology

crizuwai. 7.29. Mol. Microbial.

F.. Nicklen,

Cold Spring

1986, pp. 229-239.

of the Spodoprera

thurbqienvis

Sanger,

flank a

( 1984) 95-102.

M.H., Ferrari,

r-amylases

D.. Menou, G.. Chaufaux,

M.-M.: Nucleotide region

Genetics.

S.E., Lamsa,

Genetics

Press. Orlando,

sequences 160

NY, 1972.

in viva. In Ganesan,

Molecular

Academic

repeat

gene. J. Bacterial.

P.J. and Gray. G.L.: Homologous tion of their hybrids

crystal

53 (1989) 242-255.

H.R.: Inverted

protein

Cold Spring

Rey, M.W., Requadt,

Insecticidal

Reviews

J.W. and Whiteley,

Miller, J.N.: Experiments

We wish to thank CRC S.p.A. (S. Giovanni al Natisone, Udine) for performing the bioassays, R. Marzari for immunoblotting, F. Scoffone for expert technical assistance, L. Negri for typing the manuscript and Marco Bianchi for helpful discussions and critical reading of the manuscript. This work was partially supported by a grant from Minister0 della Pubblica Istruzione, Rome.

insecticidal

Environ,

H. spec-

2010-2017. Hiifte, H. and Whiteley,

B. thuringiensis

ACKNOWLEDGEMENTS

Appl.

thuringiensis.

and insectrcidal

Proc.

gene

3 (198’)) 229-23X.

A.R.: DNA sequencing Natl.

coding

of Bacillus

Acad.

Sci.

LISA.

with chain74 (lY77)

5463-5467. Schncpf.

H.E.

and

Whiteley,

R. thuringiexsi.s crystal

REFERENCES

H.R.:

protem

Cloning

and

expression

of the

gene in E. coli. Proc. Natl. Acad. SCI.

USA. 78 (1081) 28Y3-2897. Adang,

M.J., Staver,

and Thompson, mid clones

M.J., Rocheleau,

T.A., Leighton,

D.V.: Characterized

of the crystal

full-length

protein

R.S., Faust,

Biotech. method

for sequencing

analysis

plasmid

of a polycistronic

Ge, A.Z., Shivarova, specificity

Rev. in

operon,

thuringiensis

coding

S. and Grimm, for

.sppoC’A.in BociNus of the Eombyx mori

nucleotide

sequence

of

crystal

protein.

the

kurhdl

region in protein

subspkurstakiHD1. Gene 48 (1986) 109-I 18. Hanahan, D.: Techniques for transformation of E. coli. In Glovcr, (Ed.). Oxford,

DNA

Cloning:

A Practical

1985, pp. 109-132.

Approach,

of

gene

Vol. I. IRL

from the DNA base

on a simple artificial

medium.

of the larvae of nine noctuid J. Econ. Entomol.

58 (1965)

522-524. Towbin, H., Stachelm, T. and Gordon, J.: Electrophorctic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure some

Wabiko.

applications.

I-I., Raymond.

Proc.

Natl.

Acad.

Sci. USA.

76 (1979)

of

D.M. Press.

protoxin

K.C. and Bulla, L.A.: Brrcilhrs rhuriry&wsis gene sequence

and gene product

analysis.

entoDNA 5

(1986) 305-314. Webcr,

H and Weismann,

proteins

C.: The hypervartable

protein from B. thuringiensis deduced J. Biol. Chem. 260 (1985) 6264-6272.

H.H. and Hale, R.L.: Mass-rearing

species

mocidal

&endotoxin

entomopathogenic

sequence.

4350-4354.

Sci. USA. 86 (1989) 4037-4041.

M., Schweitzer,

H.E., Wong, H.C. and Whiteley, H.R.: The amino acid sequence

of a crystal

and

165-170.

and complemcntation

I31 (1985) 1091-I 105.

on a Bacillus

Proc. Natl. Acad.

B. thuringiensis:

DNA 4 (1985)

sequence

sporulation

a fast and simple

NJ. and Dean, D.H.: Location

domain

genes

sequencing:

Schnepf,

Shorey,

K.C. and Bulla,

CRC Critical

thuringiensis,

DNA.

J.: Nucleotide

subtihs. J. Gen. Microbial.

the

H., Raymond,

P.H.: Supercoil

P. and Errington,

Geiser,

kursrnki

6 (1987) 163-232.

Chen, E.Y. and Seeburg, Fort.

ofBacii/us

plas-

serta. Gene 36 (1085) 289-300.

R.M., Wabiko,

L.A.: The biotechnology

R.F.

and truncated

of B. thurirzgiensis subsp.

HD-73 and their toxicity to Manduca Andrews,

J., Barker,

Nucleic

C.: Formation

by recombination

between

E., Hcnner,

for hybrid

D.J., Estell, D.A. and Chcn, E.Y.: Clon-

ing, sequencing, and secretion B. .rrrbti/i.y. Nucleic Acids Res. W.R. and Whiteley,

region

coding

cloned genes in E. co/i.

Acids Rcs. 11 (1983) 5661-5669.

Wells, J.A., Fcrrari,

Widncr.

of gents

related,

of B. an?)Miqurfircirm

subtiltsin

in

11(1983) 7Y1 I-7Y25.

H.R.: Location

of the diptcran

in a lepidopteran-dipteran crystal protein thurirrgiensi.~. J. Bacterial. 172 (1990) 2826-2832.

specificity

from

Baci//u\