Expression of the nodulation gene nod C of Rhizobium meliloti ... - NCBI

2 downloads 34 Views 2MB Size Report
... C fusion protein as well as cI repressor. M.John et al. 2 ..... Torok,I., Schmidt,J. and. John,M. (1985) in Ludden,P.W. and Burris,I.E. (eds.), Proceedings of the.
The EMBO Journal vol.4 no. 10 pp.2425-2430, 1985

Expression of the nodulation gene nod C of Rhizobium meliloti in Escherichia coli: role of the nod C gene product in nodulation

Michael John1, Jurgen Schmidt1, Ursula Wienekel Eva Kondorosi1l2, Adam Kondorosi1l3 and Jeff Schell1 1Max-Planck-Institut fur Zuchtungsforschung, Abt. Schell, D-5000 Koln 30, FRG, Institutes of 2Biochemistry and 3Genetics, Biological Research Center, Hungarian Academy of Sciences, P.O. Box 521, H-6701 Szeged, Hungary Communicated by J.Schell

The nod C gene of Rhizobium meliloti encodes a protein of mol. wt. 44 000 which is highly conserved in at least three Rhizobium species. In order to overproduce this protein, a gene fusion of X cI repressor sequences to a large fragment of nod C was constructed. The fusion was placed under control of the tac promoter on plasmid pEA305 to yield pJS1035. IPTG-induced Escherichia coli cells harbouring pJS1035 accumulated the cl-nod C hybrid protein up to 19% of total cellular protein. The synthesis of the hybrid protein drastically inhibits the growth rate of the bacterium. The fusion protein was purified by gel and hydroxyapatite chromatography in the presence of SDS. Antibodies raised against the purified fusion protein precipitated the mol. wt. 44 000 nod C proteins of R. meliloti and of the broad-host range Rhizobium strain NGR234, which were both expressed in E. coli minicells. The hybrid protein is associated with the outer membrane of E. coli cells, and the cl-nod C fusion protein appears to be an integral membrane protein. Nodulation of alfalfa by R. meliloti and of clover by R. trifolii was markedly inhibited (- 50%) by the addition of antibodies against the hybrid protein to plant growth medium and inoculum. Key words: antibodies/gene expression/gene fusion/nod C gene/ Rhizobium

predicted amino acid sequence of this protein also shows that the carboxyl-terminal region contains large stretches of hydrophobic amino acids, which led to the assumption that the 44 000dalton nod C protein may be associated with the bacterial membrane (Torok et al., 1984; Rossen et al., 1984). Here we report the overproduction and purification of a hybrid protein synthesized in Escherichia coli by a plasmid carrying a fusion of X cI repressor sequences and a large fragment of nod C from R. meliloti. Antibodies against nod C were produced and we provide evidence that the cl-nod C hybrid protein is a transmembrane protein which is associated with the outer membrane of E. coli cells. EcoRi

EcoRI

Nil2 Poul PtI

tac

ci I ApR

SaoIl

HindN

P'PI

COON

7C ApR

Hindill

pEA30S

Sphl

Pvl

Sph l

nodC

PJS20R

ll ~~~~~~~~~~~~Sst

I

EcoRI

TC5

Pvul

l

-I lj 1S T2Cj

SalI 4 Sal I/CIaI

HindM

Kenow Pol I . 4dNTPs Isolate -kb Isolate 6-kbfragment1i65

Klenow Pol. i

*4dNTP's

t Sal

-Cal Ifragment

T4DAligase Transform E.coliW3110

Introduction Rhizobium meliloti interacts symbiotically with alfalfa (Medicago sativa) by forming root nodules in which the bacteria fix nitrogen. Most of the symbiotic genes of R. meliloti are located on very large plasmids (megaplasmids, Banfalvi et al., 1981; Rosenberg et al., 1981). The genes coding for early nodulation functions which are common to a wide range of plant hosts ('common' nod genes) are clustered on a 4-kb region of the megaplasmid (Kondorosi et al., 1984; Schmidt et al., 1984). Within this nod gene cluster at least three protein-coding regions were mapped by deletion mutations and by transposon mutagenesis and designated as nod A, nod B and nod C (Schmidt et al., 1984; Kondorosi et al., 1985). The nod A, nod B and nod C gene products (mol. wts. 23 000, 28 500 and 44 000, respectively) are transcribed from left to right on a map on which these nod genes are located 25 kb to the left of the structural genes for nitrogenase (nif KDH). The nodulation genes of R. meliloti and R. leguminosarum have been sequenced and a comparison of the nucleotide and predicted amino acid sequences revealed - 70% homology. A region of highest amino acid homology (-95%) was found in the nod C gene product (Torok et al., 1984; Rossen et al., 1984). The -

© IRL Press Limited, Oxford, England.

Screen or orientation

+ of insert with Pvu

Fig. 1. Construction of plasmid pJS 1035 for the expression of the cl-nod C hybrid protein. The strategy for constructing a gene fusion between portions of the nod C coding region and the cI gene of the expression vector pEA305 (Amann et al., 1983) is outlined. The dotted box corresponds to the cI gene of phage X. The junction between the cI and nod C fragments in pJS1035 is indicated by a vertical wavy line. The location and orientation of the tac and lpp-lac promoters (black boxes) in the plasmids are indicated. Ap (ampicillin) and Tc (tetracycline) mark the antibiotic resistance genes on the plasmids. The expression plasmids carry two copies of the transcription terminators T1 and T2 (Brosius et al., 1981). For other details see text.

2425

M.John et al. 2

PTG 1 0

*/ a

--U-

0

M

lh

2|-

r;

0 0

0 I61,

0

a

C

Z-.;;;

I4 .--

C). 6

.11

I

1

-

0

*i

5

1

me U-

I

.-

-I

ad

-f > 0

o

= a

u

6U

6A.J

4h

.^

9 2.5 _

6 66 !

(D

3

. ,/

._.v

_

i.

/I

LU -j

..

0

I)

10

5

15

Ti M E (h"

21 5 -

1 Fig. 2. Effect of the synthesis of the cl-nod C hybrid protein on the growth rate of E. coli W3110. Symbols: -Oi -, no inducer; -*-, 1 mM IPTG added as indicated (arrow). The inset shows the synthesis of the nod C hybrid protein upon induction with IPTG. Samples were withdrawn at the times indicated after IPTG addition, analyzed by SDS-PAGE and proteins were stained with Coomassie blue. The arrowhead marks the position of the hybrid protein.

Results Construction of expression plasmid pJS]035 To produce sufficient quantities of the nod C protein for biological studies, we constructed a plasmid which expresses a fusion protein that has a portion of the cI repressor of phage X at the amino terminus and the nod C protein at the carboxyl-terminal end. Since the DNA sequences of the cI gene (Sauer, 1978) and of the nod C gene (Torok et al., 1984) are known, it was straightforward to ligate the nod C coding sequence in frame with the X cI initiation codon on the expression vector pEA305 (Figure 1). Plasmid pEA305 (Amann et al., 1983) overproduces the phage X cI repressor under the control of the inducible tac promoter (de Boer et al., 1983). The vector has a single HindIII cloning site within the cI gene and carries strong transcriptional terminators downstream of the cI gene to increase plasmid stability after induction of the tac promoter. Plasmid pEA305 was digested with Hindlll and protruding single-stranded ends were converted to double-stranded flush ends by filling-in with the Klenow fragment of DNA polymerase I and the four deoxyribonucleotide triphosphates. The 6-kb fragment was isolated by gel electrophoresis and used as vector (Figure 1). Protein mapping (Schmidt et al., 1984) and DNA sequence analysis of the nod region (Torok et al., 1984) has shown that the nod C gene is contained within a 1. 8-kb EcoRl fragment (see also Figures 1 and 5). This EcoRI fragment was cloned in pINII-A2 yielding pJS209. Plasmid pJS209 was cut at its unique ClaI and the two Sall sites and the sticky ends were filled-in with the Klenow fragment of DNA polymerase I and all four dNTPs. A 1.05-kb Sall-Clal fragment was isolated that carries the right portion of the nod C gene plus 190 bp to the right of nod C. The ligation of the vector fragment with the nod C-coding frag-

2426

-

(J

2

3

4

Fig. 3. Purification of the cl-nod C hybrid protein from E. coli W31101acPL8 containing pJS1035. Samples from various stages of the purification were analyzed on a 9% SDS-polyacrylamide gel. Lane 1, bacterial pellet; lane 2, cell debris after sonication and centrifugation; lane 3, pooled fractions from Bio-Gel A-1.5 m; lane 4, pooled fractions from the hydroxyapatite column.

ment resulted in pJS 1035, which was transformed into the lac

repressor-overproducing strain W3 1 lOacIqL8. Plasmids of ampicillin-resistant colonies were analysed with various restriction enzymes and two of 36 isolates had the PvuI site of the inserted coding fragment in the desired position. Overproduction and purification of the cI-nod C fusion protein Cells of E. coli W31 IOlaclqL8 containing the plasmid pJS 1035 were grown in M9 medium supplemented with casamino acids. Induction of the tac promoter with 1 mM isopropyl /3-D-thiogalactoside (IPTG) resulted in the synthesis of high levels of cInod C fusion protein (Figure 2). The fusion protein, with a mol. wt. of 46 500, constitutes the major protein in the total bacterial lysate. After a 4-h induction period the yield of this protein was 19% of total cellular protein. However, the synthesis of the fusion protein drastically reduces the growth rate of the bacteria. The growth curve without inducer shows a normal shape indicating that the repression of the tac promoter prevents the growth inhibition (Figure 2). Induced cells containing the plasmid pJS1035 were lysed by sonication, followed by solubilization of the collected pellet in SDS-containing buffer. The hybrid protein was purified by a rapid purification procedure (see Materials and methods) involving gel and hydroxyapatite chromatography in the presence of SDS. After the last step the cl-nod C hybrid protein was >95% pure as judged by SDS-PAGE. The progress of a typical purification is shown in Figure 3. Antibodies raised against X cI repressor were used to identify the cl-encoded portion of the 46 500-dalton fusion protein induced in E. coli W3 1 10 cells harbouring pJS 1035. Extracts of induced cells and purified cI-nod C fusion protein as well as cI repressor -

*l a) L

°

_ s

-

A

_~~~

ri

m

,^,

A~~

...

B

fo

92.5

x x

66

92.5 -_

M

I

w

-Z z

-Z _i1 M

B

M.-Z,,..

_~~ =

66 45 CK

4

Ct:

45

.w

c

_AA

-z-

-Z

Expression of the R. meliloti nod C gene

-

_%

r

__--

_

- -w

J

C)

V

_I

3

_

31

2

31

2

-

M

2

1

4

3

1

2

3

4

Fig. 4. Gel and Western blot analysis of cI repressor and tthe cl-nod C hybrid protein. Proteins were analyzed by SDS-PAGE and stained with Coomassie blue (A). The corresponding Western blot (B) Nwas stained with antibodies against cI repressor as described in the text. Laines: 1, IPTGinduced W31 10 containing pEA305 (control); 2, purified cI repressor (control); 3, induced W3110 containing pJS1035; 4, purific nd cl-nod C fusion protein; mol. wt. markers (M). 0

U)

69

_

46

_-

30

-

eRr

0>

Q

(/)

an

< _ "m

_M 14.3

_ do

q

2

3

g_ I M

1

4

5

7

6

8

Fig. 5. Immunoprecipitation of nod C protein expressed in E. coli minicells containing nod genes from Rhizobium meliloti 41 and Rhizcobium sp. NGR234. Minicells containing the plasmids indicated were isolated and labelled with [35S]methionine as described (Schroder et al., extracts were reacted with the IgG fraction of rabbit antiseirum against the cl-nod C hybrid protein (lanes 2,4 and 6,8). Lanes 1,3 an,Id5,7 were cell extracts not precipitated. Cell extracts and immunoprecipita by SDS-PAGE, fluorography and autoradiography. The no C protein is indicated by an arrowhead. Mol. wt. markers (M) are sho%wn as mol. wt. x 10-3. The photograph is a composite of two gels ruin at different

td

times.

SDS-polyacrylamide gel (Figure 4A) and transferred to nitrocellulose. As shown in Figure 4B (lanes 3 and 4) the fusion protein reacts with anti-cI antibodies. Thie X cI repressor was included as a positive control in this experimlent (Figure 4B, lanes 1 and 2). were run on a

3

4

5

Fig. 6. Cellular location of the cl-nod C hybrid protein in E. coli W31 10. Soluble (cytoplasmic) and membrane fractions were prepared as described (Ito et al., 1977) and samples were analyzed by SDS-PAGE. Proteins were visualized by Coomassie blue staining (A) and immunostaining (B) using antibodies against cl-nod C hybrid protein. Use of the cl-nod C hybrid protein

to

raise antibodies against

nod C

The purified fusion protein was used to immunize rabbits, and the resulting IgG fraction of the serum was tested for antibodies specifically directed against the 44 000-dalton nod C protein. As shown in Figure 5 (lanes 2 and 4) the antibodies only precipitated the 44 000-dalton nod C protein from extracts of E. coli minicells harbouring plasmids pJS201 and pJS209. As negative control we used plasmid pJS204, which carries a large zSstII deletion eliminating the nod B and nod C coding sequences (Schmidt et al., 1984). Antibodies against the cl-nod C protein did not react with any of the proteins synthesized from pJS204 (lane 6). As stated above, the nod C gene sequences are highly conserved in a variety of fast-growing rhizobia as well as in the broad host range Rhizobium strain NGR234 (Torok et al., 1984; Rossen et al., 1984; Bachem et al., 1985). Based on these results, the immunologic relationship of the R. meliloti protein with the nod C protein of NGR234 was examined. Plasmids pJS201, pJS204 and pJS209 carry nod genes from R. meliloti 41, whereas pJS2020 carries the nod C gene of the broad host range Rhizobium strain NGR234 (Bachem et al., 1985). Figure 5 (lanes 7 and 8) shows that antibodies raised against the nod C protein of R. meliare able to precipitate the 44 000-dalton nod C protein of NGR234 from 35S-labelled minicell extracts. Cellular localization of the cl-nod C hybrid protein in E. coli

loti

During the purification procedure (see Materials and methods) the hybrid protein was recovered in the insoluble fraction of E.

coli cell extracts (Figure 3, lane 2), which suggested that the fusion protein may be membrane bound. To determine more precisely the location of the cI-nod C protein in the membrane, we fractionated cells into soluble and membrane fractions by two successive centrifugations through appropriate bilayers of sucrose solution (Ito et al., 1977). SDS-gel electrophoresis of the cell fractions (Figure 6A) and subsequent immunostaining of the corresponding Western blot (Figure 6B) with antibodies against the 2427

M.John et al.

D

A

--

9 2.5 66

A

D

A

D

.,

-w~

0

x

-4 5

^,,,kw

I-

I

Lu

(A

3 31

a

0

z

0

a

c) 21.5 .4 0

z

-

14.4

M

1

2

3

Fig. 7. Phase separation of the cl-nod C hybrid protein compared with hydrophilic proteins. Phase separation was performed as described (Bordier, 1981). Detergent phases (D) and aqueous phases (A) were analyzed on a 12% SDS-polyacrylamide gel stained with Coomassie blue. The samples contained the following proteins: 1, cl-nod C hybrid protein; 2, cI repressor; 3, bovine serum albumin and cytochrome c; M, marker proteins.

fusion protein revealed that the cl-nod C hybrid protein is associated with the outer membrane. The pattern of the outer membrane proteins (lane 4) shows that the hybrid protein is the major band in this fraction. Since the cI repressor was found in the soluble fraction of the overproducing E. coli strain carrying plasmid pEA305 (data not shown), it can be concluded that the nod C portion in the hybrid protein is responsible for the association with the outer membrane. In a further experiment we explored the use of various detergents to solubilize the hybrid protein from the outer membrane fraction. Commonly used detergents such as Triton X-100 (at 1 %), Triton X-1 14 (at 1 %), CHAPS (at 2%) and octyl glucoside (at 3 %) were ineffective in solubilizing this protein. The hybrid protein could, however, be solubilized in the presence of SDS at a concentration of at least 0.05 %. Since the complete removal of SDS from the solubilized membrane protein results in aggregation the presence of small amounts of detergent is necessary to maintain the hybrid protein in soluble form. Integral membrane proteins which are characterized by the presence of hydrophobic domains have the property of partitioning into the detergent phase during phase separation in solutions of Triton X-1 14 (Bordier, 1981). As shown in Figure 7 the purified cl-nod C hybrid protein also partitions into the detergent phase (lane 1D) whereas the cI repressor (lane 2A) and other hydrophilic proteins (lane 3A) were found in the aqueous phase. This suggests that the hybrid protein has a hydrophobic domain which is responsible for the insertion into the membrane. Effect of antibodies raised against the cl-nod C hybrid protein on nodule formation When R. meliloti or R. trifolii wild-type strains were inoculated onto their host plants together with antibodies directed against the nod C protein, nodule formation was reduced to -50% in comparison with control experiments which contained antibodies against cI repressor (Figure 8). Normal nodulation occurred in 2428

Fig. 8. Inhibition of nodulation by antibodies raised against nod C. The nodules were induced by R. meliloti AK631 on alfalfa (A) and R. trifolii Resh4O3 on red clover (B). Seeds of alfalfa and red clover were germinated for - 1 week in 175-ml tubes containing 20 ml of sterile agar medium (Kondorosi et al., 1977) supplemented with IgG (30 ,ug/ml). Ten plants were used for each experiment. The plants were inoculated with 108 bacteria and the number of nodules was scored daily over a 22-day incubation period. At the intervals indicated (arrows), fresh antibodies (240 4g IgG) were added to each tube. R. meliloti with antibodies against cl-nod C hybrid protein (-0-); R. meliloti with anti-cl repressor antibodies (control, -O-). R. trifolii with antibodies against cl-nod C hybrid protein (-*-); R. trifolii with anti-cl repressor antibodies (control, -O-).

the control experiments and there was no difference in the number of nodules formed in comparison with plant tests in which no antibodies were added. Immunodiffusion tests showed that the antibodies raised against the cl-nod C hybrid protein and the cI repressor did not react with root extracts of alfalfa. These experiments suggest that the nod C protein may be located on the cell surface of Rhizobium, so that antibodies directed against nod C are able to bind to this outer membrane protein causing a reduction in the level of nodulation. The observed reduction of nodule formation after the addition of antibodies against nod C of R. meliloti to R. trifolii provides evidence that in R. trifolii also the nod C protein is involved in nodulation. Discussion To elucidate how the nod C gene product is involved in the initiation of nodule formation, we overproduced a large portion of this protein together with sequences of the X cI repressor in E. coli and we used this fusion protein to raise antibodies against nod C. Previous studies have shown that the nod C gene of R. meliloti is poorly expressed in minicells of E. coli DS410 (Schmidt et al., 1984). E. coli cells containing plasmids in which the nod C gene was under the control of a plasmid vector promoter showed a reduced growth rate, which indicates that the product of the cloned gene could be toxic to the cells. This effect made it necessary to regulate the expression of the nod C gene.

Expression of the R. meliloti nod C gene

For this purpose we used the inducible tac promoter system of the plasmid pEA305 which can be stably maintained in the lac repressor overproducing strain W3 1 10lacIqL8. Overproduction of this protein in Rhizobium was not possible due to the lack of a suitable expression vector which allows the controlled overproduction of a protein that is toxic to the host. Plasmid pEA305 directs the synthesis of high levels of the X cI repressor (26% of total cellular protein) upon induction of the tac promoter with IPTG (Amann et al., 1983). We inserted 70% of the coding region of nod C including the C-terminal end of this protein at the HindIII site of the cI gene (Figure 1). This plasmid construction (pJS 1035) reads from the X cI gene in frame into the nod C gene and produces a high level of hybrid protein when the promoter is induced (Figure 2). We observed a band of the predicted size (mol. wt. 46 500), amounting to 19% of total cellular protein. When the tac promoter is induced in cells containing pJS 1035 by 1 mM IPTG, the cells reduce their growth rate significantly (Figure 2). This may be due to the hydrophobicity of the nod C portion, which causes the hybrid protein to stick to the bacterial membrane. As shown in Figure 6 (lane 4) this protein is the most abundant band in the outer membrane fraction. Western blot analysis revealed that the cl-nod C hybrid protein reacts with antibodies raised against cI repressor, which indicates that the fusion protein carries sequences of the cI protein. We purified the fusion protein to homogeneity by a simple procedure involving two chromatographic steps. Since the hybrid protein could be solubilized by SDS only, the purification steps we used were restricted to methods which allow protein separations in the presence of SDS. The purified hybrid protein was used as an antigen to produce antibodies in rabbits. Polyclonal antibodies against the hybrid protein are able to immunoprecipitate the 44 000-dalton nod C proteins of R. meliloti and of the broad host range Rhizobium sp. strain NGR234, which were both expressed in E. coli minicells (Figure 5). This indicates that both nod C proteins are antigenically related polypeptides. These data extend genetic studies which have recently shown that the nod C genes of R. meliloti, R. leguminosarum and Rhizobium NGR 234 are closely related (Torok et al., 1984; Schmidt et al., 1984; Rossen et al., 1984; Bachem et al., 1985). Several lines of evidence suggest that the cl-nod C hybrid protein is an integral membrane protein which is associated with the outer membrane of E. coli. The first line of evidence suggesting hydrophobic domains in the nod C protein was provided by the predicted amino acid sequence. It could be deduced from the hydrophobicity profile (Torok et al., 1984) that the carboxylterminal part of the nod C protein, which we had fused to the cI repressor sequences, contained the most hydrophobic region of this polypeptide. Furthermore, the solubilization of the hybrid protein requires an ionic detergent. Another independent line of evidence for a hydrophobic character of the fusion protein is provided by experiments in which this protein was localized in the outer membrane fraction of E. coli (Figure 6). According to recent findings (Tommassen et al., 1985) the localization of hybrid proteins which are only based on cell fractionation experiments have to be interpreted with caution. But further evidence for hydrophobic regions in the nod C protein is provided by experiments in which the purified hybrid protein was subjected to phase separation in Triton X-1 14 solution. As shown by Bordier (1981) hydrophilic proteins are found exclusively in the aqueous phase and integral membrane proteins are recovered in the detergent phase. After separation the hybrid protein was found in the detergent phase whereas the cI repressor and other hydrophilic proteins were recovered in the aqueous -

phase (Figure 7). Until now we were not able to detect the natural nod C protein either in free-living rhizobia or in nodules by Western blot analysis using the anti-nod C antibodies. Furthermore, it is interesting to note that nod gene transcripts have not been detected in freeliving rhizobia or in nodules. In plant nodulation experiments with R. meliloti and R. trifolii (Figure 8) it was shown that the nod C protein is expressed and accessible to antibodies resulting in a reduced nodule formation. These data suggest that the nod C protein may also be located in the cell envelope of Rhizobium and that a substance of plant origin may be necessary for the induction of this protein. Assuming that the nod C protein is a transmembrane protein in Rhizobium, it is possible that this protein plays a role in transmembrane signalling. The reduction of nodulation after the addition of antibodies against nod C could be explained by the inhibition of a pore protein or blocking of a possible receptor site. Electron microscopy of immunogold-stained thin sections of bacterial cells will be used as an alternative technique to localize further the nod C protein in E. coli and, if possible, also in Rhizobium. The availability of monospecific, polyclonal antibodies prepared against the nod C protein, as well as sufficient quantities of pure nod C protein could greatly aid in elucidating the biological role of the nod C protein in the early steps of the nodulation process. Materials and methods Bacterial strains and plasmids The lac repressor-overproducing strain E. coli W3 101acIqL8 (Brent and Ptashne, 1981) used as a host for tac promoter-containing plasmids was kindly provided by Mark Ptashne. E. coli DS410 (Dougan and Sherratt, 1977) was used as minicell-producing strain. Rhizobium meliloti AK631 (Nod+,Fix+) is a compact colony variant of the wild-type R. meliloti 41. R. trifolii Resh4O3 was provided by Z.Banfalvi. Unless otherwise stated, E. coli was grown at 37°C in LB medium (Miller, 1972) containing the appropriate antibiotic. Rhizobium strains were cultured at 28°C in TY medium (Beringer, 1974). Plasmid pEA305 carrying the tac promoter and the cI gene of phage X (Amman et al., 1983) was kindly provided by Mark Ptashne. Plasmids pJS201 and pJS204 have already been described (Schmidt et al., 1984). pJS209 which contains a 1.8-kb EcoRI fragment carrying the nod C gene of R. meliloti in pIN-II-A2 was derived from pJS201 by insertion of an EcoRI linker - 100 bp upstream of the SphI site of the nod C gene. In pJS2020 a 3.7-kb EcoRI fragment carrying the nod C gene of the broad-host range Rhizobium sp. strain NGR234 (Bachem et al., 1985) was inserted into the EcoRI site of pIN-II-A2 (Nakamura and Inouye, 1982) so that the promoter was proximal to the PstI site of the insert.

Molecular cloning Plasmid cloning procedures, including restriction endonuclease digestions, incubation with Klenow fragment of DNA polymerase I, attachment of linkers, ligation, bacterial transformation, screening recombinant plasmids, and agarose gel electrophoresis of DNA fragments were essentially as described by Maniatis et al. (1982). Restriction enzymes were purchased from Boehringer, Mannheim and New England Biolabs and were used according to the manufacturers' instructions. Klenow fragment, T4 DNA ligase and EcoRI linkers (dGGAATTCC) were from Boehringer. Protein purification E. coli W31 IOlacJqL8 containing pJS1035 was grown in M9 salts (Miller, 1972) supplemented with 0.2% casamino acids, 0.5% glycerol and 50 Atg/ml ampicillin. At a cell density of A600 =0.6 the tac promoter was induced by addition of 1 mM IPTG (Sigma). 4 h after induction, cells were harvested by centrifugation and washed with cold 50 mM Tris-Cl (pH 7.5) containing 100 mM NaCl. Cells were suspended in buffer containing 100 mM Tris-Cl (pH 8.0), 200 mM KCI, 10 mM MgCl2, 2 mM CaCl2, 1 mM EDTA, 0.5 mM DTT, 5% glycerol, and 20 mg/l phenylmethylsulfonyl fluoride (PMSF) and were lysed by sonication. The cell debris was collected by centrifugation (20 min at 14 000 g). The pellet was resuspended and washed in the same buffer. After centrifugation the cell debris was lysed in 0.1 M sodium phosphate buffer (pH 6.8) containing 3 % SDS and 100 mM DTT. The sample was boiled for 15 min, centrifuged and applied to a Bio-Gel A-1.5 m column (2.6 x 84 cm, 20°C) equilibrated and eluted with 0.1 M sodium phosphate (pH 6.8) plus 1 mM DTT and 0.1 % SDS. Fractions

2429

M.John et al.

containing the hybrid protein (determined by SDS-PAGE) were pooled, dialyzed against 10 mM sodium phosphate buffer (pH 6.4), containing 1 mM DTT and 0.1% SDS and applied to a Bio-Gel HTP column (1 x 40 cm, 20°C) equilibrated in this buffer. The hybrid protein was eluted with a linear sodium phosphate gradient from 10 to 500 mM phosphate (Moss and Rosenblum, 1972). Fractions were screened for the 46 500 mol. wt. hybrid protein by SDS-PAGE. The pooled hybrid protein was dialyzed against 2 x phosphate-buffered saline (PBS) and concentrated by ultrafiltration (Amicon). The protein was stored at -20°C at a concentration of - 2 mg/mi in 2 x PBS. For the purification of cI repressor the overproducing E. coli strain W31 101acIL8 carrying the plasmid pEA305 was cultured and induced as described (Amann et al., 1983). The repressor was purified by the procedure of Sauer and Anderegg (1978) except that the repressor was eluted with a linear gradient (50- 600 mM KCl) from the DNA cellulose column. The purified repressor was dialyzed and stored as described above. The purity of the hybrid protein and cI repressor was judged by SDS-PAGE (Figures 3 and 4). Antibodies 1 mg of purified antigen in 2 x PBS was emulsified (Branson sonifier) with an equal volume of complete Freund's adjuvant (Difco) and was injected s.c. at multiple sites along the back of the rabbit. A second injection of 1 mg antigen in incomplete Freund's adjuvant (Difco) was made 21 days later into the upper leg muscles. Ten days after the second injection the rabbit was bled. Rabbit immunoglobulin (IgG) was prepared by ammonium sulfate precipitation (50% saturation) and affinity chromatography on protein A coupled to Sepharose CL-4B (Miller and Stone, 1978; obtained from Pharmacia). The eluted IgG fraction was immediately neutralized with 1 M Tris-Cl (pH 8.0), dialyzed against PBS and concentrated (Amicon, PM30 membrane). For prolonged storage IgG was filter sterilized (Millex-GV, Millipore) and aliquots containing -7 mg/mi of protein were frozen at -20°C. The reactivity of the antisera and IgG against the corresponding antigens was checked by the immunodiffusion method of Ouchterlony (1958).

Immunoprecipitation E. coli minicells were isolated and labelled with [35S]methionine (Amersham) as described (Schroder et al., 1981). The minicell pellet was resuspended in 90 UI of a solution containing 1% SDS, 50 mM Tris-Cl (pH 8.0) and 1 mM EDTA and heated in a boiling water bath for 5 min. The sample was diluted 1:10 with cold TEN buffer (20 mM Tris-Cl, pH 7.4, 1 mM EDTA, 100 mM NaCl) containing 0.1% Nonidet P-40 and centrifuged for 10 min (SS-34 rotor, 20 000 r.p.m.). In order to reduce non-specific binding the supernatant was mixed with 100 Id of a 10% suspension of washed, formalin-fixed Staphylococcus aureus cells (Kessler, 1975; obtained from BRL) and incubated on ice for 1 h. After centrifugation (Eppendorf, 1 min) the supernatant was mixed with 10 /1l of IgG and the mixture was incubated on ice for 30 min. 100 1l of the washed 10% suspension of S. aureus cells was added and the mixture was kept on ice for a further 30 min with occasional mixing. The antigen-antibody-bacteria complex was collected by centrifugation and washed according to the procedure of Shapiro and Young (1981). After the final wash the antigen-antibody-bacteria complex was pelleted through a 10% sucrose cushion in TEN buffer (1 ml cushion in an

Eppendorf tube). The immunoprecipitate was dissolved in electrophoresis sample buffer. The sample was boiled for 3 min followed by centrifugation to remove bacteria. Proteins were electrophoresed on a 12% polyacrylamide slab gel (Laemmli, 1970). The gel was fixed overnight in 50% trichloroacetic acid, treated with En3Hance (NEN), washed, dried, and exposed to Kodak X-Omat S film at -70°C. Other procedures Electrophoretic transfer of proteins from SDS-polyacrylamide gels to nitrocellulose was carried out as described (Towbin et al., 1979). The immobilized antigens were detected on the nitrocellulose sheets with the appropriate antibodies followed by incubation with peroxidase-conjugated goat anti-rabbit second antibodies using the immun-blot assay kit from Bio-Rad. The amount of hybrid protein was determined by laser densitometer scanning (LKB) of a polyacrylamide slab gel stained

with Coomassie blue. Cytoplasm, inner and outer membrane fractions of E. coli were prepared as described (Ito et al., 1977), using the spheroplasting procedure, lysis by sonication and sucrose-gradient centrifugation. Phase separation of hydrophilic and integral membrane proteins in Triton X-1 14 solutions was performed as described by Bordier (1981). In order to test various detergents to solubilize the nod C hybrid protein the outer membrane fraction from E. coli was suspended in 50 mM Tris-Cl (pH 7.6) plus 150 mM KCl at 4°C and adjusted to the specified concentration of detergents. After incubation for 1 h on ice the suspension was centrifuged (100 000 g, 1 h at 4°C). Supernatant fractions were separated from pellets and aliquots were analyzed by SDS-PAGE. Plant nodulation experiments were carried out with alfalfa (Medicago sativa) and red clover (Trifolium pratense) seedlings which were grown on nitrogenfree medium as described (Kondorosi et al., 1977).

2430

Acknowledgements We appreciate the skilful technical assistance of H.D.Krussmann, and we gratefully acknowledge the other help in this work: E.Amann and M.Ptashne (tac promoter vectors), M.Kalda and D.Bock (photographs), and F.J.de Bruijn for discussions and critical reading of the manuscript. This work was supported by grants from Deutsche Forschungsgemeinschaft (Br 800/1-1), Bundesministerium ftir Forschung und Technologie (BCT 03652/Proj. 13) and by a joint grant of Deutsche Forschungsgemeinschaft and the Hungarian Academy of Sciences (Sche 220/1-2).

References Amann,E., Brosius,J. and Ptashne,M. (1983) Gene, 25, 167-176. Bachem,C., Kondorosi,E., Banfalvi,Z., Horvath,B., Kondorosi,A. and Schell,J. (1985) Mol. Gen. Genet., 199, 271-278. Banfalvi,Z., Sakanyan,V., Koncz,C., Kiss,A., Dushaj. and Kondorosi,A. (1981) Mol. Gen. Genet., 184, 318-325. Beringer,J.E. (1974) J. Gen. Microbiol., 84, 188-198. de Boer,H.A., Comstock,L.J. and Vasser,M. (1983) Proc. Natl. Acad. Sci. USA, 80, 21-25. Bordier,C. (1981) J. Biol. Chem., 256, 1604-1607. Brent,R. and Ptashne,M. (1981) Proc. Natl. Acad. Sci. USA, 78, 4204-4208. Brosius,J., Dull,T.J., Sleeter,D.D. and Noller,H.F. (1981) J. Mol. Biol., 148, 107-127.

Dougan,G. and Sherratt,D. (1977) Mol. Gen. Genet., 151, 151-160. Ito,K., Sato,T. and Yura,T. (1977) Cell, 11, 551-559. Kessler,S.W. (1975) J. Immunol., 115, 1617-1624. Kondorosi,A., Svab,Z., Kiss,G.B. and Dixon,R.A. (1977) Mol. Gen. Genet., 151, 221-226. Kondorosi,E., Banfalvi,Z. and Kondorosi,A. (1984) Mol. Gen. Genet., 193, 445452.

Kondorosi,A., Kondorosi,E., Banfalvi,Z., Putnoky,P., Torok,I., Schmidt,J. and John,M. (1985) in Ludden,P.W. and Burris,I.E. (eds.), Proceedings of the 14th Steenbock Symposium on Nitrogen Fixation and CO2 Metabolism, Elsevier, NY, pp. 91-98. Laemmli,U.K. (1970) Nature, 227, 680-685. Maniatis,T., Fritsch,E.F. and Sambrook,J. (1982) Molecular Cloning, A Laboratory Manual, published by Cold Spring Harbor Laboratory Press, NY. Miller,J.H. (1972) Experiments in Molecular Genetics, published by Cold Spring Harbor Laboratory Press, NY. Miller,T.J. and Stone,H.O. (1978) J. Immunol. Methods, 24, 111-125. Moss,B. and Rosenblum,E. (1972) J. Biol. Chem., 247, 5194-5198. Nakamura,K. and Inouye,M. (1982) EMBO J., 1, 771-775. Ouchterlony,O. (1958) Prog. Allergy, 5, 1-78. Rosenberg,C., Boistard,P., Denarie,J. and Casse-Delbart,F. (1981) Mol. Gen. Genet., 184, 326-333. Rossen,L., Johnston,A.W.B. and Downie,J.A. (1984) Nucleic Acids Res., 12, 9497-9508.

Sauer,R.T. (1978) Nature, 276, 301-302. Sauer,R.T. and Anderegg,R. (1978) Biochemistry (Wash.), 17, 1092-1100. Schmidt,J., John,M., Kondorosi,E., Kondorosi,A., Wieneke,U., Schroder,G., Schroder,J. and Schell,J. (1984) EMBO J., 3, 1705-1711. Schroder,J., Hillebrand,A., Klipp,W. and Piuhler,A. (1981) Nucleic Acids Res., 9, 5187-5202. Shapiro,S.Z. and Young,J.R. (1981) J. Biol. Chem., 256, 1495-1498. Torok,I., Kondorosi,E., Stepkowski,T., Posfai,J. and Kondorosi,A. (1984) Nucleic Acids Res., 12, 9509-9524. Tommassen,J., Leunissen,J., van Damme-Jongsten,M. and Overduin,P. (1985) EMBO J., 4, 1041-1047. Towbin,H., Staehelin,T. and Gordon,T. (1979) Proc. Natl. Acad. Sci. USA, 76, 4350-4354.

Received on 17 May 1985; revised on 1I July 1985