ABSTRACT We have determined the nucleotide sequences of the KlebsieUla pneumoniae nifL ( regulation of N2 fixation genes ) and the Escherichia coli ginA ( glutamine synthetase ) promoters .
We compared these sequences with the published sequences of three other promoters that , like the nifL and ginA promoters , are activated by the general nitrogen regulators glnF ( ntrA ) and ginG ( ntrC ) .
The three promoters are the argTr ( arginine transport ) and dhuA ( histidine transport ) promoters of Salmonella typhi - murium and the nifH ( nitrogenase ) promoter of Rhizobium mel - iloti .
All five sequences ( with at most one mismatch ) contain the heptameric consensus sequence T-T-T-T-G-C-A .
In the R. meliloti nifH and K. pneumoniae nifL promoters , in which the transcrip - tion initiation sites have been determined , the consensus sequence is situated in the -15 region .
We recently reported that the K. pneumoniae nifA product , which activates nifgenes , can substitute for the ginG ( ntrC ) product in activating promoters of several genes involved in nitrogen assimilation , including the nifL , the glnA , and the R. meliloti nifH promoters .
It is likely that nifA also activates the S. typhimurium argTr and dhuA promoters .
In contrast , the ginG product can not substitute for the nifA product in the acti - vation of the K. pneumoniae nifH ( nitrogenase ) promoter .
Con - sistent with this latter observation , and supporting the conclusion that the T-T-T-T-G-C-A sequence is a regulatory site for ginG product activation , the K. pneumoniae nifH promoter ( C-C-C-T - G-C-A ) has only partial similarity with the T-T-T-T-G-C-A con - sensus sequence in the-15 region .
Positive control of many nitrogen assimilation genes in enteric bacteria is mediated by the products of genes ginF ( ntrA ) ( 1-5 ) and ginG ( ntrC ) ( 3-7 ) .
Genes regulated in this manner include those governing the utilization of histidine ( hut ) , proline ( put ) , and arginine ( aut ) , the transport of histidine ( dhuA , hisJQMP ) ( 8 ) and arginine ( argT ) ( 8 ) , and anabolic genes involved in the assimilation of NH4 ' .
The latter class is exemplified by the glutamine synthetase gene ginA , which has served as a model gene for the study of nitrogen regulation ( for review , see ref .
The enteric bacterium Klebsiella pneumoniae contains an additional nitrogen assimilation pathway not found in the other enteric species .
This pathway is the enzymatic reduction of atmospheric dinitrogen ( N2 ) to ammonia ( NH4 + ) ( nitrogen fixation ) .
A contiguous cluster of at least 17 nif genes arranged in seven or eight transcription units is involved in nitrogen fixation ( for review , see refs .
Two of these genes , found in a single transcription unit ( the nifLA operon ) , are responsible for regulation ( 12-17 ) .
The transcription of nif genes occurs only under conditions of nitrogen starvation and it has recently been shown that the nifLA operon acts as an intermediate in sensing The publication costs ofthis article were defrayed in part by page charge payment .
This article must therefore be hereby marked `` advertisement '' in accordance with 18 U.S.C. § 1734 solely to indicate this fact .
Specifically , activation of the nifLA operon is initiated under conditions of nitrogen deficiency by the general nitrogen control regulators , the products of glnF and ginG .
Once transcription is initiated from the nifLA op-eron , the nifA product in conjunction with the ginF product activates all other nif promoters .
It is believed that the nifL gene product remains inactive as a repressor unless a fixed source of nitrogen or oxygen is introduced into the system .
We have recently reported that the nifA product can substitute for the ginG product as a nitrogen control regulator , replacing the ginG product in the activation of its own promoter and , in addition , the promoters of several nitrogen assimilation genes , including the hut , put , aut , and glnA genes ( 12 ) .
This fact , along with other observations , led us to the proposal that nifA and glnG are evolutionarily related , and that nifA evolved directly from glnG in K. pneumoniae and thereby mediates stringent-control of the nitrogen fixation pathway .
Consequently , nif promoters ( except for the nifL promoter ) , in contrast to the promoters of other nitrogen assimilation genes , can not be activated by the glnG product .
As a first step in elucidating the mechanism of glnG-and nifA-mediated gene activation , we have determined , and we present here , the DNA sequences of the nifL and glnA promoters , both of which are activated by the ginG or nifA product ( 12 , 18 ) .
In addition , we report a comparison of the nifL and ginA promoters with four other glnG-or nifA-regulated promoters .
MATERIALS AND METHODS Genetic Materials .
K. pneumoniae strains : KP5617 is hisD-hsdR-recA-srl-nifB - , KP5611 is hisD-recA-sr-nifA - , and KP5614 is hisD-hsdR-recA-sr - ( 19 ) .
Plasmids : pGR116 carries the K. pneumoniae nifLABQ genes in pBR322 and has been described ( 20 ) .
pDO503 to pDO509 are described in the legend to Fig. 1 .
pgln8 is derived from pBR322 and contains a small part of the Escherichia coli ginA structural gene , including the 5 ' upstream region , on a 1,200-base-pair ( bp ) EcoRI/Cla I fragment ( 18 ) .
pgln26 , which contains a subregion of pgln8 , is pBR322 carrying a 635-bp Hae III fragment ( with HindIII linkers ) that presumably contains sequences subject to regulation by the products of ginG and nifA ( refs .
pgln8 and pgln26 were gifts from K. Backman and Y. M. Chen .
DNA sequences were determined according to the partial chemical degradation method of Maxam and Gilbert ( 21 ) .
S1-nuclease-mapping was performed t Present address : Dept. of Molecular Biology , Massachusetts General Hospital , Boston , MA 02114 .
§ Present address : Biogen , Inc. , Cambridge , MA 02142 .
RESULTS Structure of the nifLA Promoter .
An 8.0-kilobase-pair ( kb ) EcoRI fragment in the nifcluster contains the two operons nifLA and nifBQ and a portion of the nifF gene ( Fig. 1 ) .
Previous studies indicated that the unique endonuclease site Bgl II lies within the nifL structural gene and that the nifL promoter is probably located within 600 bp upstream ( to the right in Fig. 1 ) of the Bgl II site ( 25 ) .
To delimit the nifL promoter , we defined its location functionally by using exonuclease Bal 31 to generate a set of deletions extending in both directions from the unique Bgl II site in pDO503 ( see Fig .
Deletion de-recircularized with BamHI molec-rivatives of pDO503 were ular linkers and the extent of each deletion was determined by restriction endonuclease mapping analysis .
The resulting set of nifL deletion plasmids was tested for complementation of a nifA mutation [ complementation , which results in the ability to fix N2 ( Nif + phenotype ) , was not possible when this same EcoRI fragment was attached onto high-copy-number plasmid vectors due to inhibition of the chromosomal nif genes ( 20 , 26 ) but was possible in this case because the vector used was the low-copy-number plasmid pRK248 ( 24 ) ] .
Because none of the deletions used for this analysis extended beyond the Sal I site in nifL and thus into the nifA gene , a Nif-phenotype could be attributed to the deletion of regions governing nifL transcription or trans-lation-initiation .
( Alternatively , a Nif-phenotype could also be due to polarity from fusion-generated stop codons .
) All 6 nifL deletion plasmids tested complemented the nifB mutation in KP5617 , but only pDO504 and pDO505 complemented the nifA mutation in KP5611 .
Plasmids pDO506 , pDO507 , pDO508 , and pDO509 have in common deletions extending beyond the Xma I ( Sma I ) site near the 5 ' end of the nifL gene .
This result suggests that the region upstream from the XmaI site contains the nifL promoter .
To define the location of the nifL promoter precisely , we used the Si nuclease mapping method ( 22 ) .
Total RNA extracted from derepressed cultures of K. pneumoniae was hybridized to purified coding strand of the Xma I/EcoRI fragment labeled with 32P at the 5 ' end of the Xma I site .
Both the RNA extracted from KP5614 ( wild type ) and KP5614 harboring the multicopy plasmid pGR116 ( which carries the nifLA operon ) showed specific protection from nuclease SI digestion of approximately 200 bp extending upstream from the Xma I end .
This specifically protected length of DNA was observed only when RNA was extracted from cultures grown in the absence of NH4 ' ( data not shown ; see ref .
This was consistent with the fact that plasmids pDO507 , pDO508 , and pDO509 ( all of which have undergone deletions of this region ) failed to complement the nifA mutation , indicating that this transcript comes specifically from the nifLA op-eron .
In contrast , plasmid pDO506 contains the presumed transcription start point but nevertheless failed to complement .
However , DNA sequence analysis of plasmid pDO506 revealed that it was missing the A-G-G-A-G ribosome binding site ( 28 ) ( data not shown ) .
This may result in premature termination of transcription .
We determined the DNA sequence of the region upstream from the Xma I site .
The sequencing strategy consisted of end labeling the Xma I site of a purified Xma I/EcoRI fragment and the BamHI sites present on deletion clones pDO506 and pDO507 ( see Fig .
The sequence was read upstream toward the EcoRI site until overlapping sequences were found .
The sequences of both strands were determined for 420 bp upstream from the Xma I site .
In Fig. 2 , we show the DNA sequence of the noncoding strand of regions of interest ( see below ) .
When the nuclease Sl-resistant ( mRNA-protected ) coding strand ( labeled at the Xma I 5 ' end ) was subjected to electro-phoresis next to a sequencing ladder ( labeled at the same Xma I 5 ' end ) , several mRNA start sites were found clustered at approximately 200 bp upstream from the Xma I site ( data not shown ; see ref .
The band with the strongest radioactive signal corresponded to the mRNA start point shown in Fig. 2 .
RNA extracted from cultures grown aerobically or at high-temperatures ( but in the absence of NH4 + ) also initiated transcription at this site ( data not shown ; see ref .
At 64-68 bp downstream from the mRNA start site there is a presumptive Shine-Dalgarno ( 28 ) ribosome binding site followed closely by several possible ATG protein initiation co-dons .
By aligning the DNA sequence with the NH2-terminal amino-acid sequence of the nifL protein [ kindly provided by M. Chance and W. H. Orme-Johnson ( personal communication ) ] , we assigned the first ATG codon to code for the NH2-terminal methionine of the nifL protein ( Fig. 2 ) .
The -10 and -35 regions of the nifL promoter show poor homology to the Pribnow/Rosenberg and Court E. coli consensus promoter sequence ( 30 , 31 ) .
This is consistent with the fact that the nifL promoter is inactive in the absence of positive control ( 12 ) .
Structure of the ginA Promoter .
Backman et al. have cloned and characterized the E. coli glnA gene ( 18 ) .
They identified the ginA promoter ( see Fig .
lb ) by showing that the insertion of a 635-bp Hae III fragment into the promoter probe vector A132 ( 32 ) resulted in the regulated expression of,3-galactosid-ase by the ginG product .
We have recently reported that the expression of f-galactosidase from the cloned ginA promoter in AglnlOl was also positively regulated by the K. pneumoniae nifA product ( 12 ) , indicating that regulatory sequences recognized by nifA also reside on this 635-bp fragment .
We determined the DNA sequence for about 200 bp on each side of the Hae III site at the `` right '' end of the 635-bp Hae III fragment as drawn in Fig .
A DNA sequence beginning about 60 bp to the right of the Hae III site corresponds to the primary protein sequence reported for the first 21 amino-acids of glutamine synthetase ( 33 ) [ except for three differences : the previously reported amino-acid sequence lacked methionine at the 1st position ( Fig. 2 ) , placed histidine rather than serine at the 4th position ( Fig. 2 ) , and lacked histidine at the 12th position between glutamate and glutamine ( data not shown ) ] .
About 12 bp upstream from the ginA ATG start codon is a Shine-Dalgarno sequence ( 28 ) .
Further upstream and within the Hae III fragment are three consensus promoter-like -10 region sequences ( at -80 , -130 , and -180 bp from the ATG initiation codon , underscored with dots in Fig. 2 ) that are possible candidates for the in-vivo promoter ( s ) .
Interestingly , the second upstream sequence is 10 bp downstream from a presumptive glnG/nifA regulatory sequence ( see below ) .
The DN sequence of the ginA promoter region relevant to this paper is shown in Fig. 2 .
Comparison of Promoters Regulated by glnG , nifA , or Both .
The DNA sequences of three promoters additional to the ones reported in this paper that require both ginG and glnF for activation have been reported .
They are the dhuA ( histidine transport ) and argTr ( arginine transport ) promoters of Salmo-nella typhimurium ( 29 ) , and the nifH ( nitrogenase reductase ) promoter of R. meliloti ( 23 , 34 ) .
We have shown that the latter promoter , like the ginA and nifL promoters , also responds to K. pneumoniae nifA-mediated activation ( 23 ) .
It is likely that the nifA product can also activate the dhuA and argTr promoters because we have previously shown that the nifA product can substitute for the ginG product in the activation of the aut ( arginine utilization ) , hut ( histidine utilization ) , and put ( proline utilization ) genes in E. coli and Klebsiella aerogenes ( 12 ) .
The structure of the K. pneumoniae nifH promoter has also been reported recently ( 23 ) .
Functionally , it differs from the above promoters in that the nifA product , but not the ginG product , can activate its transcription ( 34 ) .
Among the six promoters whose sequences have been determined , the mRNA transcriptional start sites have been determined for the nifL , the R. meliloti nifH , and the K. pneumoniae nifH promoters ( 23 , 27 ) .
A comparison of the nifL and the R. meliloti nifH promoters reveals a common 6-bp sequence ( T-T-T-G-C-A ) in their respective -15 regions ( Fig. 2 ) .
The probability of encountering the same 6-bp sequence randomly is about 1 in 4,000 .
Similar sequences were found in the ginA , dhuA , and argTr promoters ( see Fig. 2 ) , about 130-190 bp away from the ATG translational start codons .
The sequence T-T-T-T-G-C-A appears to be a consensus sequence for all of these promoter regions ( except the K. pneumoniae nifH promoter ) because each of the five promoters contains at least 6 out of the 7 bp of T-T-T-T-G-C-A .
In addition , the dhuA and the R. meliloti nifH sequences show 8 bp ( T-T-T-T-G-C-A-C ) in common .
Most striking is the 12-bp sequence ( C-T-T-T-T-G-C-A-C-G-A-T ) shared between the R. meliloti nifH promoter and the ginA promoter region .
In the case of the nifL and the R. meliloti nifH promoters , it is known from SI nuclease mapping experiments that this consensus sequence lies within a critical region for regulation .
In the case of ginA , argTr , and dhuA , however , it is not known whether the consensus sequence is actually within the promoter .
We can not rule out the possibility that the sequence merely reflects a fortuitous evolutionary similarity .
The -35 regions of the K. pneumoniae nifL and the R. mel-iloti nifH promoters give only weak evidence of homology .
When the -35 regions of these promoters were compared to the corresponding regions ( upstream from the presumptive consensus T-T-T-T-G-C-A sequence ) of the dhuA , argTr , and ginA promoters , homology was observed only between the nifL and dhuA sequences .
These promoters share the 6-bp sequence A-T-C-A-C-G .
Interestingly , we found a stretch of sequence about 150 bp upstream from the K. pneumoniae nifL promoter that is very similar to the R. meliloti nifH promoter sequence .
This presumptive second nifL ( `` nifL2 '' in Fig. 2 ) promoter may respond to glnG-or nifA-mediated activation , although transcripts originating from this region were not observed by using the SI nuclease mapping technique described above , at least not under conditions in which transcription was activated by the ginG product .
Among the six promoters compared in Fig. 2 , only the K. pneumoniae nifH promoter fails to respond to the ginG product .
The -15 region of this promoter , which responds only to the nifA product , does not contain T-T-T-T-G-C-A but rather C-C-C-T-G-C-A .
Because all six promoters probably respond to the nifA product , it is possible that the sequence T-G-C-A , common to all six promoter regions , is necessary for nifA-me-diated gene activation , although this sequence alone is most likely insufficient .
This conclusion is based on the argument that the probability of encountering a 4-bp sequence is relatively high .
A more plausible explanation would be that nifAmediated transcription requires either the presumptive ginG/nifA recognition sequence T-T-T-T-G-C-A or T-G-C-A in conjunction with another 8-bp sequence , A-C-G-G-C-T-G-G ( Fig. 2 , overscored ) , that is found in the -30 region of both the K. pneumoniae and R. nwliloti nifH promoters ( 23 ) .
DISCUSSION Transcriptional start points for the dhuA and argTr promoters that lie 120 and 60 bp , respectively , downstream from the T-T-T-T-G-C-A sequence have been suggested previously .
In the case of the argTr promoter , Higgins and Ames ( 29 ) assigned the -10 region on the basis of an E. coli consensus promoter-like sequence .
Preliminary results indicate that in-vivo transcription does start near this previously proposed promoter ( G. F.-L .
In the case of the dhuA promoter , the -10 region was assigned on the basis of both an E. coli consensus promoter-like sequence and on a promoter-up point mutation ( 29 ) .
Consensus promoter-like sequences can also be found at 10 and 65 bp downstream from the T-T-T-T-G-C-A region of the ginA promoter to the left of the Hae III site ( Fig. 2b , underscored with dots ) .
Consistent with the possibility that transcription of these three promoters does not start near the T-T-T-T-G-C-A sequence is the fact that the T-T-T-T-G-C-A sequence in each promoter is quite far from the ATG translation initiation codons ( see Fig. 2 ) .
On the other hand , consensus promoter-like sequences are located at approximately 75 , 60 , and 60 bp from the ATG translation initiation codons of the argTr , dhuA , and ginA promoters , respectively .
The T-T-T-T-G-C-A sequences in the nifL and R. meliloti nifH promoters are only about 80 bp away from the ATG start codons .
In the following paragraph , we propose a model for the transcriptional control of the five promoters discussed above that takes into account the apparent differences in the structures of the nifL and R. meliloti nifH promoters on the one hand and the ginA , dhuA , and argTr promoters on the other .
The nifL and R. meliloti nifH promoters share the feature that both promoters exhibit an absolute requirement for glnG-or nifA-mediated activation .
In contrast , the ginA , dhuA , and argTr promoters constitute a second group of ginG/nifA-reg-ulated promoters because they are transcriptionally active at a basal ( lower ) level even when cells are grown in excess NH4 ' and glutamine , suggesting that positive regulation is not required for basal activity .
It is possible that the basal level of transcription of the latter three promoters is due to initiation at the poor but recognizable consensus promoter-like sequences and that the glnG/glnF-or nifA/gInF-induced levels of transcription are due to initiation near the T-T-T-T-G-C-A sequences .
Alternatively , under activating conditions , transcription may still be initiated at the presumptive consensus promoter-like sequence .
However , recognition of the T-T-T-T-G-C-A sequence by ginG ( or nifA ) and glnF products may enhance transcriptional efficiency at the downstream transcriptional start site .
For example , transcriptional factors could be shuttled downstream from the T-T-T-T-G-C-A site or transcriptional factors binding at the T-T-T-T-G-C-A sequence could promote cooperative binding of transcriptional factors at the downstream site The latter hypothesis is interesting in light of the second T-T-T-T-G-C-A sequence in the K. pneumoniae nifL promoter at -170 bp upstream from the transcriptional start site .
Because this region of the nifL promoter shows remarkable similarities to the R. meliloti nifH promoter , which is activated by nifA product , it is possible that auto-activation of the nWfLA operon by the nWfA product occurs-at this upstream sequence .
On the other hand , mRNA transcripts have not been mapped to this region .
An alternative explanation for a function for this region may be analogous to that described for the glnA , dhuA , and argTr promoters .
That is , the nifL2 region may function as an initial binding site to shuttle transcriptional factors to the downstream promoter , or the binding of transcriptional factors to the nifL2 locus may promote cooperative binding at the downstream promoter .
Finally , it is interesting that the sequence homology that exists among the K. pneumoniae nWfL , the R. meliloti nifH , and the K. pneumoniae nifH promoters occurs at the -15 and -30 regions .
Prokaryotic RNA polymerases are believed to recognize promoters at the -10 and -35 regions ( 30 , 31 , 35 ) .
This suggests that the nifA and glnF or.glnG and glnF products may be modifying RNA polymerase affect-its ability recognize to to these particular -10 and -35 sequences .
In closing , we would like to note that Drummond et al. have independently determined the nucleotide sequence and the start point transcription of the nifL promoter ( 36 ) , which are in complete agreement with the results presented here .
A preliminary report of this work was presented at the North Atlantic Treaty Organization symposium in Porto Portese , Italy , August 1982 .
We thank Orme-Johnson and Chance for preliminary W. H. M. results on the NH2-terminal amino-acid sequence of the nifL protein , G. F.-L .
Ames for preliminary results on the transcriptional of start site the argTr promoter , K. Backman and Y. M. Chen for plasmids pgln8 and pgln26 , and B. Magasanik for helpful discussions and for critically reading the manuscript .
We also thank R. Hyde for preparation of the manuscript .
This work was supported by National Science Foundation Grant PCM-8104193 , awarded to F.M.A. ; by U.S. Department of Ag-riculture Grant 59-2253-1-1-722-0 , awarded to W. H. Orme-Johnson at the Massachusetts Institute of Technology , with a subcontract to F.M.A. ; and by National Science Foundation Grant PCM78-08576 awarded to B. Magasanik at the Massachusetts Institute of Technology .
D.W.O. was , supported by a National Institutes of Health Fellowship in Biochemistry and D.R. was supported by a National Institutes of Health Postdoctoral Fellowship .
Garcia , F. , Bancroft , S. , Rhee , S. G. & Kustu , S. ( 1977 ) Proc .
Gailardin , C. M. & Magasanik , B. ( 1978 ) J. Bacteriol .
Pahel , G. & Tyler , B. ( 1979 ) Proc .
Kustu , S. , Burton , D. , Garcia , E. , McCarter , L. & McFarland , N. ( 1979 ) Proc .
Leonardo , J. M. & Goldberg , R. B. ( 1980 ) J. BacterioL 142 , 99-110 6 .
deBruijn , F. J. & Ausubel , F. M. ( 1981 ) Mol Gen. Genet .
Espin , G. , Alvarez-Mornales , A. & Merrick , M. ( 1981 ) Mol Gen. Genet .
Kustu , S. G. , McFarland , N. C. , Hui , S. P. , Esmon , B. & Ames , G. F.-L .
Roberts , G. P. & Brill , W. J. ( 1981 ) Annu .
Kennedy , C. , Cannon , F. , Cannon , M. , Dixon , R. , Hill , S. , Jen-sen , J. , Kumar , S. , McLean , P. , Merrick , M. , Robson , R. & Postgate , J. ( 1981 ) in Current Perspectives in Nitrogen Fixation , eds .
Gibson , A. H. & Newton , W. E. ( Aust .
Ow , D. W. & Ausubel , F. M. ( 1982 ) Nature ( London ) 301 , 307-313 .
Hill , S. , Kennedy , C. , Kavanagh , E. , Goldberg , R. B. & Hanau , R. ( 1981 ) Nature ( London ) 290 , 424-426 .
Merrick , M. , Hill , S. , Hennecke , H. , Hahn , M. , Dixon , R. & Kennedy , C. ( 1982 ) Mol .
Roberts , G. P. , MacNeil , T. , MacNeil , D. & Brill , W. J. ( 1978 ) J. Bacteriol .
Dixon , R. , Eady , R. R. , Espin , G. , Hill , S. , Iaccarino , M. , Kahn , D. & Merrick , M. ( 1980 ) Nature ( London ) 286 , 128-132 .
MacNeil , D. , Zhu , J. & Brill , W. J. ( 1981 ) J. Bacteriol .
Backman , K. , Chen , Y. M. & Magasanik , B. ( 1981 ) Proc .
Ow , D. W. & Ausubel , F. M. ( 1980 ) Mol Gen. Genet .
Riedel , G. E. , Brown , S. & Ausubel , F. M. ( 1983 ) J. Bacteriol 153 , 45-56 .
Maxam , A. & Gilbert , W. ( 1980 ) Methods Enzymot 65 , 499-560 .
Berk , A. J. & Sharp , P. A. ( 1977 ) Cell 12 , 721-732 .
Sundaresan , V. , Jones , J. , Ow , D. W. & Ausubel , F. M. ( 1983 ) Nature ( London ) 301 , 728-732 .
C. R. & Helinski , D. R. Bacteriol .
Thomas , M. , Meyer , ( 1980 ) J. 141 , 213-222 .
Riedel , G. E. , Ausubel , F. M. & Cannon , F. C. ( 1979 ) Proc .
Buchanan-Wollaston , V. , Cannon , M. C. & Cannon , F. C. ( 1981 ) Mol Gen. Genet .
Ow , D. W. ( 1983 ) Dissertation ( Harvard Univ. , Cambridge , MA ) .
Shine , J. & Dalgarno , L. ( 1974 ) Proc .
Higgins , C. F. & Ames , G. F.-L .
Pribnow , D. ( 1979 ) in Biological Regulation and Development : Gene Expression , ed .
Goldberger , R. F. ( Plenum , New York ) , Vol .
Rosenberg , M. & Court , D. ( 1979 ) Annu .
Meyer , B. J. , Maurer , R. & Ptashne , M. ( 1980 ) J. Mol Biol 139 , 163-194 .
Ginsburg , A. ( 1972 ) Adv. Protein Chem .
Sundaresan , V. , Ow , D. W. & Ausubel , F. M. ( 1983 ) Proc .
Losick , R. & Pero , J. ( 1981 ) Cell 25 , 582-584 .
Drummond , M. , Clements , J. , Merrick , M. & Dixon , R. ( 1983 ) Nature ( London ) 301 , 302-307 .