The EMBO Journal vol .4 no. 2 pp.539-547 , 1985 Nitrogen regulation in Salmonella typhimurium .
Identification of an ntrC protein-binding site and definition of a consensus binding sequence Giovanna Ferro-Luzzi Ames and Kishiko Nikaido Department of Biochemistry , University of California , Berkeley , CA 94720 , USA Communicated by F.Blasi We have investigated the DNA-binding ability of a nitrogen regulatory protein , the product of the ntrC gene , to several nitrogen-regulated promoters in Salmonella typhimurium .
The ntrC protein is able to bind to the regulatory region ( dhuA ) of an operon coding for genes involved in the active transport of histidine , but not to another transport-related regulatory region , argTr .
It bound to two different sites within the regulatory region of glnA ( glutamine synthetase ) and to one site in the regulatory region for the ntrBC operon .
A consensus sequence has been derived from these four binding sites .
The binding sequence displays dyad symmetry , as expected for the dimeric ntrC protein .
The relationship of the binding sites to regulation of transcription initiation and termination , and to published homologies within the sequences of regulatory sites for nif genes is discussed .
Key words : Salmonella typhimurium/ntrC/protein / DNA-binding Introduction Nitrogen control in Salmonella typhimurium and Escherichia coli is centrally exerted through the concerted action of least at three products : ntrA , ntrB and ntrC ( reviewed in gene Magasanik , 1982 ; Merrick , 1982 ) .
( These genes are also refer-red to as glnF , glnL and glnG , respectively , in E. coli .
) The products of these required genes are to regulate expression of the glutamine synthetase gene ( glnA ) , and of numerous genes involved in active transport and utilization of several amino-acids .
The ntrA and ntrC gene products act as positive activators of transcription of several unlinked nitrogencontrolled genes under conditions of nitrogen-limitation .
The ntrC product also mediates repression of glnA transcription under conditions of nitrogen excess ( Gaillardin and Magasanik , 1978 ) , or in the absence of the ntrA gene product ( Kustu et al. , 1979b ) .
In addition , it represses transcription of the ntrBC genes in either excess or limiting nitrogen in the presence or absence of the ntrA product ( Pahel et al. , 1982 ; Ueno-Nishio et al. , 1983 ) .
The function of the ntrB gene product is less clear .
In the working model proposed by Kustu and her collaborators ( Kustu et al. , 1979a ; McFarland et al. , 1981 ) , the ntrB protein may participate in a complex with the ntrC gene product when the latter functions as a repressor .
The ntrA protein may positively regulate by interacting directly with the ntrC ( and ntrB ) gene product , or be reponsible for the synthesis of a signal-molecule .
S. typhimurium and E. coli possess a periplasmic transport system for histidine and one for lysine , arginine and ornithine ( Ames Ferro-Luzzi and Higgins , 1983 ) .
The histidine transport operon , composed of genes hisJ , hisQ , hisM and hisP , is transcribed from the regulatory region dhuA ; the monocistronic IRL Press Limited , Oxford , England .
operon argT is transcribed from the regulatory region argTr ( Figure 1 ) .
In S. typhimurium these two transport operons are regulated by nitrogen availability ( Kustu et al. , 1979b ; Higgins and Ames Ferro-Luzzi , 1982 ; Stern et al. , 1984 ) .
The entire nucleotide sequence of these transport operons and their regulatory regions have been obtained ( Higgins and Ames Ferro-Luzzi , 1981 , 1982 ; Higgins et al. , 1982 ) .
Here we present evidence that the ntrC protein binds to a specific nucleotide sequence within dhuA and that this sequence bears strong homology to sequences within several other nitrogen-regulated promoters .
Results Nitrogen regulation at argTr and dhuA : effect of regulatory mutations We have shown that argTr and dhuA are regulated by nitrogen availability ( Kustu et al. , 1979b ; Higgins and Ames Ferro-Luzzi , 1982 ; Stern et al. , 1984 ) .
We have also shown that both promoters are under ntrB regulation by demonstrating an elevation of their level of expression when combined with the regulatory mutation ntrBJ39 , which is known to elevate expression of several nitrogen-regulated promoters ( Kustu et al. , 1979b ; Stern et al. , 1984 ) .
To determine how argTr and dhuA respond to the other two known components of the nitrogen regulatory circuit ( ntrA and ntrC ; Kustu et al. , 1979a ; McFarland et al. , 1981 ) we combined a Mud ] ( lac Ap ) operon fusion placed under control of either promoter ( Stern et al. , 1984 ) with mutations eliminating either ntrA or ntrC function , and examined the level of 3-galactosidase produced under either nitrogen-rich or nitrogen-poor growth-conditions .
Table I , lines 1 and 5 , show that the basal level of expression under conditions of nitrogen excess is higher for dhuA than for argTr , in general agreement with previous measurements of the levels of the hisJ and argT gene products ( Kustu et al. , 1979b ) and of Mud ] fusions expression ( Stern et al. , 1984 ) .
However , the experiments presented in Table I used cells in exponential phase which dispaly a much larger difference between dhuA argTr and than stationary-phase cells : 17-fold versus 5-fold determined previously ( Stem et al. , 1984 ) .
Table I also shows that both promoters are under nitrogen regulation and the response to nitrogen-limitation is much larger for the argTr promoter than for the dhuA promoter .
These results are in general agreement with previous determinations , ( Kustu et al. , 1979b ; Stem et al. , 1984 ) ; again , comparison with the argTr response in stationary-phase cells shows that exponential phase cells have a much larger response : 30-fold ( Table I ) versus 6-fold ( in Stern et al. , 1984 ) .
Introduction of either ntrC an or an ntrA mutation has only a minor effect on the expression of either promoter under conditions of nitrogen excess , but completely eliminates the response to nitrogen-limitation 1 ( compare line with 2 and 3 , and line 5 with 6 and 7 ) .
The results obtained with strain TA3880 show that eliminating ntrB in addition to ntrC ( by polarity of a TnJO insertion in transposon ntrB ) , does not have any additional effect .
These argTr data demonstrate that and dhuA are positively regulated by ntrA and ntrC .
Therefore , it is likely that they will interact with one or more of those gene products .
Table I. Response of transport promoters to nitrogen regulation Promoter Strain Relevanta ( - galactosidase unitsb N-poor genotype N-rich N-poor N-rich TA3741 ntrA + ntrC + 309 ( 100 % ) 850 2.8 TA3879 ntrA + ntrC-305 ( 99 % ) 180 0.6 TA3876 ntrA ntrC + 214 ( 69 % ) 134 0.6 TA3880C ntrB-glnAp 283 ( 92 % ) 218 0.8 TA3738 ntrA + ntrC + 18 ( 100 % ) 528 30d 1.4 TA3863 ntrA + ntrC-15 ( 83 % ) 21 TA3864 ntrA ntrC + 19 ( 105 % ) 15 0.8 dhuA argTr aAll strains contain a stabilized MudJ ( lac , AmpR ) insertion and are described in Table III .
bAssayed and expressed as described by Miller ( 1972 ) in cells grown to exponential phase ( A650 of 0.8 ) in the salt medium described by Kustu et al. ( 1979b ) , witWT ` H4Cl ( 20 mM ) or L-glutamine ( 3 mM ) as The carbon source was rich and poor nitrogen sources , respectively .
0.4 % glucose in all cases .
cThis strain contains a TnJO insertion in ntrB , thus it is phenotypically ntrC-because of polarity of the TnOO insertion mutation .
The glnAp mutation is a promoter-up mutation ( McCarter et al. , 1984 ) .
This dramatic difference is dIn stationary-phase cultures this due both to a higher ( - galactosidase level in N-rich conditions ( 85 units ) and a lower level in N-poor conditions ( 329 units ) in stationary-phase culture .
The ntrC gene product binds to dhuA We attempted to identify proteins which might be involved in regulating activity at either or both dhuA and argTr by analyzing cell extracts for DNA-binding activities using plasmid pFA9 as the source of labelled DNA .
This plasmid contains a 1.4-kbp fragment of chromosomal DNA containing the entire argTr , argT and dhuA region ( DNA fragment sa-2 1 : Higgins and Ames Ferro-Luzzi , 1982 ; the 3 ' end of sa-21 is 15 bp downstream from the transcription initiation site ) cloned between the EcoRI and BamHI sites of pBR322 ( Figure 1 ) .
The search was performed by a filter-binding assay and was meant to identify any protein which would bind to either promoter .
A crude protein extract from TA3010 , a strain lacking both the argT and histidine transport operons ( Table III ) , was passed through a phosphocellulose column and the eluted fractions were assayed for binding activity with nicktranslated pFA9 and pBR322 DNAs .
The latter DNA constitutes a control for proteins binding DNA unspecifically or specifically to the vector portion of pFA9 .
Numerous peaks of activity were detected , several of higher activity toward pFA9 than pBR322 DNA ( data not shown ) , as expected since the extract presumably contains other possible regulatory proteins in addition to those involved in nitrogen regulation .
One of the peaks of activity was enriched for a protein with a mol .
of-54 kd , which is the mol .
reported for the product of the ntrC gene ( McFarland et al. , 1981 ; Backman et al. , 1981 ) .
We investigated whether this peak of activity was due to the ntrC protein by utilizing a strain ( NCM302 ) which was obtained from Dr Kustu 's laboratory , and which overproduces the ntrC protein .
This strain carries a plasmid containing the ntrBC operon under temperature-sensitive control of the lambda PL promoter .
A comparison between SDS-gel electrophoretograms of crude extracts from NCM302 grown either at 30 °C or at 42 °C showed a clear increase in a 54-kd protein at 42 °C ( Figure 2A ) ; this protein has been identified genetically as the ntrC product ( J.Hirschman and S.Kustu , in preparation ) .
Phosphocellulose chromatography of an extract of NCM302 grown at 42 °C shows a peak binding activity in fractions amounts containing large of a 54-kd protein ( data not shown ) .
These fractions were pooled and chromatographed on hydroxylapatite ( Figure 2B ) .
Several fractions , greatly enriched in a 54-kd protein , exhibited binding activity specific for pFA9 .
The peak fraction ( fraction 34 ) was estimated to contain the presumed ntrC protein at a concentration of 0.16 mg/ml , at -50 % purity by densitometer analysis of the electrophoretogram , and was used for all further experiments .
The specificity of DNA binding was determined by two methods .
Firstly , pFA9 DNA was nick-translated and then divided into four fragments by with RsaI digestion ( Figure 1 ) .
Three fragments carrying dhuA , argTr or pBR322 promoter P4 , respectively , were purified and used as substrates in binding assays .
Table II shows that the presumed ntrC protein bound only to the dhuA fragment containing but not the others .
Interestingly , no binding to the fragment carrying argTr was detected under the conditions of the assay .
Secondly , the dhuA-containing fragment was further digested to give a variety of smaller fragments and the of ability the presumed ntrC protein to retain them on a filter was determined by the method of Hsieh and Brutlag ( 1979 ) .
Figure 3 shows that all fragments of DNA covering the region between -146 and -217 , which is enclosed by the smallest fragment used ( HhaI , 71 bp long ) , were significantly retained .
Definition of the protein-binding site in dhuA Footprinting experiments indicated a clearly protected region I P4 Fig. 1 .
Transport operons structure and relevant physical map of pFA9 .
Thin line : pBR322 DNA .
Thick line : S. typhimurium DNA ; open sections : regulatory and untranslated regions ; solid sections : structural genes .
P4 = pBR322 promoter 4 ( Stuber and Bujard , 1981 ) .
argT codes for an arginine - , lysine - , ornithine-binding protein ; hisJ , hisQ , hisM and hisP code for a histidine-binding protein and three membrane-bound components of histidine transport ( Ferro-Luzzi Ames and Higgins , 1983 ) .
The openreading frame upstream of argTr is presumed to be gene ubiX .
A B 130K 94K - ¬ ... m ... W ~ 0 ¬ e v Binding was assayed as described in Materials and methods .
The total number of counts added reflects the different recoveries during fragment purification .
polymerase-binding sites ( Higgins and Ames Ferro-Luzzi , e v - O ¬ m 60K lwwo .
-- 53K - 'm = 0 41 K NfIl 36K - .
, ^ a oil 0 ll l g II m m 25K ~ ¬ Jim f ¬ 420 300 30 32 34 36 38 40 42 44 46 48 50 MW 52 54 56 58 60 62 MW Fig. 2 .
Purification of the ntrC protein .
( A ) Crude extracts from strain NCM302 cultured at 30 °C or at 42 °C .
( B ) Pattern of protein elution from a hydroxylapatite column .
The numbers are the fraction numbers .
The arrows point to : ( A ) a band of increased intensity in cells cultured at 42 °C and to : ( B ) a major protein band in fraction 34 corresponding to the peak of DNA-binding activity .
Other bands of increased intensity in A are presumably due to heat-shock proteins .
in dhuA including bases from-183 to -164 when the protein was added to the transcribed ( bottom ) strand ( Figure 4A , compare lane 4 with 5 ) .
Some enhancement of digestion was also obvious at several sites .
Footprinting also allowed a preliminary analysis of the strength of binding .
Decreasing the protein concentration resulted in a loss of protection in the region between bases -168 and -164 ( lanes 5-8 ) ; 50 % protection is afforded in this region at the lowest concentration of protein used , which we estimate to be -4 x 10-8 M taking into consideration its partial purity .
Protection in the region between bases -183 and-172 was not decreased concomitantly with the decrease in protein concentration , indicating a tighter binding of the protein in this region with an estimated affinity considerably better than 4 x 10-8 M. Protection of the non-transcribed ( top ) strand also detected in the region , although the effect was same was somewhat weaker ( data shown ) .
not Figure 4A shows additional protected region between an bases -149 and 136 , which will be discussed below.-Confirmation ofthe DNA-binding protein as the ntrC protein An identical pattern of protection between bases -183 and -164 was obtained using partially purified authentic ntrC protein ( K.Sei and S.Kustu , in preparation ) ( Figure 4A , lanes 1-3 ) .
Antibodies to purified ntrC protein ( K.Sei and S.Kustu , in preparation ) antagonize the protection due to fraction 34 Table II .
Specificity of ntrC protein binding RsaI fragment nM Size Counts ( kb ) Total added Bound % of total 0.05 0.09 0.09 2.17 1.35 1.57 14.3 1.54 1.02 dhuA 23-570-3370 argTr 25-620-395 P4 ( pBR322 ) 31-560-320 ( Figure 4B ) .
These two pieces of evidence confirm that the 54-kd protein purified by us and binding to dhuA DNA is indeed the ntrC protein .
Interestingly , an important difference emerged when we compared the rest of the footprinting patterns for the two protein preparations : a strong protection is evident for bases -149 to -136 only when using fraction 34 ( Figure 4A ) .
Since neither ntrC protein preparation is completely pure , we tentatively ascribe this difference to the presence of an additional binding protein in our preparation which is absent from that obtained from Sei and Kustu .
Transcription initiation sites in dhuA and argTr Both dhuA and argTr have several potential -10 , -35 RNA 54 1982 ) .
However , a likely in-vivo acting site in dhuA was defin-ed genetically ( Higgins and Ames Ferro-Luzzi , 1982 ; Lee and Ames Ferro-Luzzi , 1984 ) .
RNA was prepared from cells grown either in nitrogen excess or nitrogen-limitation , and the transcriptional start sites for both argTr and dhuA were determined by Si nuclease digestion and probing with appropriate 5 ' - labelled DNA fragments .
Figure 5 shows that start sites occur at position -48 for dhuA ( arrow A ) and approximately -60 for argTr ( arrow B ) immediately downstream from the previously hypothesized TATA boxes ( Higgins and Ames Ferro-Luzzi , 1982 ; Lee and Ames Ferro-Luzzi , 1984 ) .
The dhuA transcriptional start site for TA271 , a strain carrying dhuAl , a promoter-up mutation raising the level of expression of this operon-10-fold ( Ames Ferro-Luzzi and Lever , 1972 ) , was the same , but the level of RNA was 10-fold elevated ( first two lanes ) .
The level of the dhuA transcript is not significantly changed during nitrogen starvation : this result is compatible with the expectation of only a 2-to 3-fold increase in expression for this promoter ( see Table I ) which is within the reproducibility of our assay .
The level of the argTr transcript is increased during nitrogen starvation > 20-fold , as expected ( see Table I ) .
Interestingly , a transcript originating approximately in the middle of the argT structural gene is present under conditions of nitrogen starvation ( arrow C in Figure 5 ) .
An additional transcript ( arrow D ) shows no response to nitrogen .
Discussion The ntrC protein binds to dhuA but not to argTr We have shown that the ntrC gene product , a nitrogen regulatory protein , binds to a 20-bp region within dhuA , a nitrogen-regulated region .
This region is located-180 bp 5 ' to the translational start site for the hisJ gene ( Figure 6 ) .
A -10 , -35 RNA polymerase-binding site in dhuA has been identified ( i ) by homology with the consensus sequence , ( ii ) as having a location 7 bp 5 ' to the transcriptional start , and ( iii ) genetically , as the locus of promoter-up mutations , ( in the -10 site ) ( Higgins and Ames Ferro-Luzzi , 1982 ; Lee and Ames Ferro-Luzzi , 1984 ) .
The ntrC protein appears to be an activator of transcription from dhuA , since this promoter is stimulated albeit slightly ( 2-to 3-fold ) , by nitrogen deprivation ( Table I , Stern et al. , 1984 ; Higgins and Ames Ferro-Luzzi , 1982 ) .
The possibility that it stimulates transcription by interacting directly with RNA polymerase seems unlikely in the case of the dhuA promoter , because of the large distance between the ntrC protein-and RNA polymerase-binding sites .
However , it is possible that an interaction between these two proteins occurs at the ntrC protein-binding site , followed by the sliding of an ` activated ' form of RNA polymerase to the -10 , -35 region , or that one of the other gene products known to be involved in nitrogen regulation participates as a means of communication between them .
It is also possible that the ntrC protein-binding site is close to an alternative transcriptional start site which is activated only under nitrogen-limitation .
Since no classical-10 , -35 region is found close to the ntrC protein-binding site , we would have to postulate that a modified RNA polymerase initiates transcription at this site .
Finally , it may be that when the ntrC protein functions as an activator of transcription it binds to a site different from the one we uncovered ; this implies that the latter site is involved in a mechanism other than that of stimulation of transcription ( see below ) .
Since the argT gene and the hisJ gene ( the first gene of the histidine transport operon ) originated from an ancestral gene by duplication ( Higgins and Ames Ferro-Luzzi , 1981 ) , we expected to find similar responses of their regulatory regions to the ntrC protein .
However , no binding of the ntrC protein to argTr was detected under our assay conditions .
An important difference between argTr and dhuA is that dhuA is expressed constitutively , at a level-17-fold higher than argTr .
The basal levels of expression of both argTr and dhuA are independent of ntrC .
Therefore , we need not invoke the differential binding of the ntrC protein to explain this difference .
Some possible explanations for the differential binding are : ( i ) The binding affinity of the ntrC protein may be much better for dhuA than for argTr and this differential af-finity may be a distinguishing and important physiological characteristic only under nitrogen-limitation , when dhuA may have to compete for the ntrC protein with argTr and other nitrogen-regulated promoters .
The ntrC protein itself is regulated by nitrogen availability ( Pahel et al. , 1982 ) and may be present in limiting amounts under some conditions .
( ii ) Binding to argTr may require additional factors .
( iii ) The ntrC protein may perform an additional , as yet undiscovered , function in dhuA , but not in argTr .
Since ntrC is a gene with pleiotropic activity , we expected to find homologous ntrC protein-binding sites in other nitrogenregulated promoters .
Evidence that the E. coli ntrC protein binds to the regulatory region of ntrBC ( glnLG ) has been presented ( Reitzer and Magasanik , 1983 ; Ueno-Nishio et al. , 1984 ) .
The existence of a promoter for the ntrBC genes ( ntrBCr ) in S. typhimurium has been established ( Krajewska ¬ -81 -34 -267 -240 -217 m -146 -126 dhuA 375 2282 631 -- Hinf I L. Bg/I a @ F ¬ roF I HhoI * - LLOa Fig. 3 .
Top : RsaI fragment carrying dhuA .
Bottom : only those fragments which are retained on the filter by the presence of ntrC protein are shown .
The negative numbers above the RsaI fragment are from Higgins and Ames Ferro-Luzzi ( 1982 ) and are relative to translational initiation as being + 1 .
The numbers below are from the pBR322 standard numbering system .
172 G G -16 8 ¬ T G-16e C A T -164 A G C fw Esa -164 / - ¬ 4mu w .
- p % W '' `` m : 0 qf vap .
% M W * bb 44 wo , ¬ qf % ... rw A A C A-149 G G - ¬ C G C A T A ` op A A T G C-136 A T A Fig. 4 .
Footprinting analysis of binding site .
All lanes contain -0.04 pmol ( 50 000 c.p.m. ) of end-labelled DNA ( 140-bp fragment between RsaI at position -266 to ClaI at position -126 in dhuA ) in a final volume of 55 jd .
( A ) Protection by ntrC ( lanes 1-3 ) and by fraction 34 ( lanes 5-8 ) .
Lanes 1-3 : 1 , 3 and 10 tl , respectively , of ntrC protein ( 0.11 mg/ml ) .
Lanes 5-8 : 10 , 6 , 3 and 1 Itl , respectively of fraction 34 .
( B ) Effect of ntrC antiserum .
Lane 1 : 5 Al of fraction 34 was mixed with 6 A1 of antiserum in a final volume of 50 Al , adjusted with binding buffer , and incubated at room temperature for 60 min and then subjected to footprinting as usual .
Lane 2 : same as lane 1 , with 6 pl of pre-immune antiserum replacing the antiserum .
Lane 3 : same as lane 1 , omitting the antiserum .
We have obtained evidence by footprinting experiments that the S. typhimurium ntrC protein binds to both regulatory regions : glnAr ( two sites ) and ntrBCr ( data not shown ) .
A comparison of the sequences of these protected regions ( Figure 6 ) , shows that there is a remarkable degree of overall homology , with two subsets of perfect homology : all four sites contain 5 ' GCAC followed can be deduced from the four available binding sequences , on the basis of the conservation of homology plus the ability to form a dyad symmetry as : 5 ' TGCACC ( four bases , AT-rich ) TGGTGCA .
The argTr sequence contains a region with only poor homology to the consensus sequence , in agreement with our failure to detect ntrC protein-binding to argTr .
% M W * bb 44 wo , ¬ qf % ... rw A A C A-149 G G - ¬ C G C A T A ` op A A T G C-136 A T A Fig. 4 .
Footprinting analysis of binding site .
All lanes contain -0.04 pmol ( 50 000 c.p.m. ) of end-labelled DNA ( 140-bp fragment between RsaI at position -266 to ClaI at position -126 in dhuA ) in a final volume of 55 jd .
( A ) Protection by ntrC ( lanes 1-3 ) and by fraction 34 ( lanes 5-8 ) .
Lanes 1-3 : 1 , 3 and 10 tl , respectively , of ntrC protein ( 0.11 mg/ml ) .
Lanes 5-8 : 10 , 6 , 3 and 1 Itl , respectively of fraction 34 .
( B ) Effect of ntrC antiserum .
Lane 1 : 5 Al of fraction 34 was mixed with 6 A1 of antiserum in a final volume of 50 Al , adjusted with binding buffer , and incubated at room temperature for 60 min and then subjected to footprinting as usual .
Lane 2 : same as lane 1 , with 6 pl of pre-immune antiserum replacing the antiserum .
Lane 3 : same as lane 1 , omitting the antiserum .
by TGGTGC ( underlined in Figure 6 ) and separated from it by five bases ( AT-rich ) .
Since the binding site in dhuA contains a region of perfect dyad symmetry ( thin overlying arrows on dhuA in Figure 6A ) and since the E. coli ntrC protein is a dimer ( Reitzer and Magasanik , 1983 ) , we propose that its dimeric state is related to its need and ability to bind to this region of dyad symmetry .
A preliminary consensus sequence A B 114 Z z B Ibrf-r 2 Cr CL - , - Cr2 Z ( I ) * r-4 az.L cr z z Z * t ' B .
RNA was prepared , hybridized to labelled probe , and digested with SI nuclease as described in Materials and methods .
Strain TA83 1 with wild-type promoter was used as a source of RNA for all experiments , except for the first two lanes which used RNA from strain TA271 .
RNA was obtained from cells grown in nitrogen-rich-medium , NR , or nitrogen-poor medium , NP .
( A ) An Alul-Hinf I ( 42 1-bp ) probe labelled at the 5 ' end of the bottom strand was used for determination of transcriptional starts in dhuA .
This probe extends from base pair -340 to base pair +81 , thus covering the entire dhuA region ( -240 to -1 ; Higgins and Ames Ferro-Luzzi , 1982 ) .
Bands visible immediately below the intact probe in A may be due to unspecific degradation of the probe and thus be artifactual .
A m-inor band above band A in dhuAl grown on nitrogen-poor medium appeared unreproducibly .
Autoradiography was for the same length of time for all lanes .
( B ) A HpaII-Sau96 ( 830-bp ) probe labelled at the 5 ' end of the bottom strand was used for the determination of transcriptional starts in argTr , in duplicate RNA preparations .
This probe extends from base pair -407 to base pair +421 , thus covering almost the entire argTr region ( Higgins and Ames Ferro ¬ Luzzi , 1982 ) .
Arrows A and B point respectively to the dhuA ( base pair -48 ) and argTr ( base pair -60 ) transcript starts .
The C arrow points to a 5 ' transcript end which is present at high level under nitrogen-limitation and is located approximately at base pair +270 , i.e. , in the middle of the argT structural gene .
Arrow D points to an RNA 5 ' end which is not affected by the nitrogen source as evident in films exposed for the same length of time .
The film exposure for the first three lanes was 10 times the exposure of the last three lanes .
Quantitation by densitometry of appropriately exposed films indicated that : ( i ) the argTr transcript is increased at least 20-fold under nitrogen-limitation ; ( ii ) the dhuA transcript is not significantly changed ; ( iii ) the dhuA transcript is 10-fold elevated by the presence of the dhuAl mutation ( densitometry data not shown ) .
Sequencing ladders for both probes were used in other experiments to place the initiation sites .
The control lanes , minus RNA and minus S1 nuclease , show no bands and the intact probe , respectively , as expected .
RNA from TA3010 , a deletion removing both transport operons , shows no transcripts with either probe ( data not shown ) .
Figure 6B shows the location of each binding site relative to the-10 , -35 region of each promoter .
The large distance between the-10 , -35 and the ntrC protein-binding sites in dhuA has already been discussed .
It is interesting that an equally large distance exists in glnAr between these two sites , while in ntrBCr the binding site overlaps the-10 region .
In Consensus TGCACCAAAATGGTGCA contrast to dhuA and glnAr , ntrBCr is not activated by ntrC plus ntrA , but is repressed by ntrC under nitrogen starvation ( Ueno-Nishio et al. , 1983 ) .
The dramatic difference in location of the ntrC protein-binding site within ntrBCr may be responsible for this difference in response to nitrogen availability .
Interestingly , inspection of the ntrC protein-binding site in dhuA reveals that it partially overlaps a sequence characteristic of a rho independent transcription terminator which is probably the termination site for the upstream gene , argT ( Higgins and Ames Ferro-Luzzi , 1982 ) .
It seems possible that a ( or the ) function of the ntrC protein at this binding site is to interfere with transcription termination , would allow transcription to continue from the argTgene into the histidine transport operon , explaining at least in part the increased level of expression of dhuA under conditions of poor nitrogen availability .
Interestingly , inspection of the ntrC proteinbinding sites in glnArII and ntrBCr also allows the construction of hypothetical termination sites in both .
The observed in-vitro inhibition of transcription of ntrBCr by the ntrC protein ( Reitzer and Magasanik , 1983 ) may be coincidental : if the ntrC protein ` sits ' on the-10 region for anti-termination it might well interfere with transcription initiation .
purposes , It should be noted that a role in anti-termination does not necessarily exclude additional roles involving transcription initiation .
The regulation of nitrogen assimilation in the nitrogen-fixing species Klebsiella and pneumoniae Rhizobium meliloti bears strong similarities to that of E. coli and S. typhimuinum .
Several contiguous operons including 17 nitrogen fixation genes ( nif ) are regulated by the products of two of the genes , nijL and nifA , which constitute an operon ( Roberts and Brill , 1981 ) .
The nifA gene product is required for activation of all nif operons except its own and is under the regulation of the central system : ntrA , ntrB and ntrC , which functions similarly to that of S. typhimurium and E. coli ( Alvarez-Morales et al. , 1984 ) .
The promoter sequences of several of these operons have been compared with each other and with those of dhuA , argTr and glnAr ( Ow et al. , 1983 ; Beynon et al. , 1983 ; Brown and Ausubel , 1984 ) .
A seven-base sequence , TTTTGCA , which was observed to be strongly conserved amongst all these promoters corresponds only to the 5 ' TTTTGCA portion of the consensus sequence we are proposing for the ntrC proteinbinding site at -182 .
Neither of our ntrC protein-binding sites in glnAr corresponds at all to that pointed out by Ow et al. ( 1983 ) .
A set of four bases , CTGG , 5 ' to and separated by six bases from the above sequence , was also strongly conserved in nif genes , but not found in dhuA , argTr or glnAr .
The relationship between these sequences and our consensus sequence , if any , remains to be explained .
A B -150 -100 -50 -1 -250 -200 T ¬ I 6 -184 +1 I .
, r ¬ dhuA - TTTTGCACCTTTTTGGTGCATAAGCCGT -41 -19 - ATG ... ~ -4 - ** i 4 -- .
- ~ ~ ntrfCr - TAATGCACTAAAATGGTGCAACCTTTTC ATG me -181 -172 ginAr I TAAAGCACTATTTTGGTGCAACATAGTC-ATG ' -221 -205 me g/nAr GTTGCACCAATGTGGTGCTTAATGTTC-ATG i ~ ~ ~ me .
-70 A -55 lG orgTr - ACCTGCATGAAAAGTCTGCAAACACACA-ATG Fig. 6 .
Nucleotide sequence of DNA-binding site and comparison with other nitrogen-regulated promoters .
( A ) Alignment of ntrC protein-binding site in dhuA with binding sites in ntrBCr , ginArI , glnArII , with a possible binding site in argTr , and with a preliminary consensus sequence .
All sequences are from S. typhimurium .
The reference for the ntrBCr sequence is Hanau et al. ( 1983 ) ; the portion of the glnAr sequence shown in this figure was obtained from J.Brenchley .
Sequences identical to those underlined in this figure are found in the corresponding regions of E coli glnAr and ntrBCr ( Covarrubias and Bastarrachea , 1983 ; Ueno-Nishio et al. , 1984 ) .
Open bars : protected region ; broken ends indicate uncertainty as to precise limits .
Thin arrows : dyad symmetry within the binding site .
Thick arrow followed by dashes : presumed terminator .
All numbering is from translational start sites .
( B ) Location of ntrC protein-binding sites within the respective regulatory regions .
Hatched boxes : presumed-10 , -35 regions .
Vertical arrows : known transcription initiation sites .
Several additional presumed promoters can be visualized in ginAr ; however , no data are available as yet on actual transcription initiation sites in S. typhimurium .
Transcription initiation sites in ginAr have been determined in E. coli as being -75 and -185 from the translational start ( Reitzer and Magasanik , personal communication ) , or at about -43 and -185 by Covarrubias et al. , personal communication .
An initiation site in ntrBCr in E. coli has been determined at -33 from the translational start ( Ueno-Nishio et al. , 1984 ) .
Overlining arrows followed by dashes : possible transcription terminators .
Additional possible terminators are present .
In E. coli , the hypothetical terminator at glnArH has a run of three Ts at its 3 ' end .
A terminator at -90 in ntrBCr in E. coli has been determined by Covarrubias et al. , personal communication .
Open bars : protected regions .
Underlining : location of binding consensus sequence .
The ntrBCr protected region is in the same location as the one identified in E. coli ( Ueno-Nishio et al. , 1984 ) .
Strain name Genotype ' Sourceb TA831 TA271 TA3738 TA3741 TA3863 This laboratory This laboratory dhuAl dhuA hisA ( argT hisJ ) 5643 This laboratory This laboratory hiszA ( argT dhuA hisJ ) 5643 ntrC352 : : TnlO This laboratory hisA ( argT dhuA hisJ ) 5643 ntrA209 : : TnlO This laboratory ntrA209 : : TnJO This laboratory ntrC352 : : TnJO This laboratory ntrBJ37 : : TnlOglnAp ( up ) 356 This laboratory hisA ( ubiX argT dhuA hisJ hisQ hisM hisP ) 6704 This laboratory NCM302 lacZam trpam Smr ( Xbio252 cI857 AHI ) / pJES23 Sydney Kustu TA3864 TA3876 TA3879 TA3880 TA3010 aAll TA strains are S. typhimurium and carry a deletion mutation in the histidine biosynthetic oepron , hisA ( F ) 645 ; all , except TA831 and TA271 , also carry a stabilized MudJ ( lac AmpR ) insertion in the hisM/P region of boundary ambiguity in the histidine transport operon : hisMiP et of ( OMudl ) 9033 ( Stern al. , 1984 ) .
Synthesis ( 3-galactosidase is under control of the dhuA promoter unless hisA5643 ( which fuses argT to hisJ ) is present when synthesis is under control of argTr ( Higgins and Ames Ferro-Luzzi , 1982 ) .
NCM302 is an E. coli strain .
bMutations ntrC352 : : TnJO , ntrA209 : : TnJO , ntrBJ37 : : TnJO , and glnAp ( up ) 356 were transferred from strains previously described ( Kustu et al. , 1979a ; McCarter et al. , 1984 ) .
Materials and methods Strains The strains used are listed in Table III .
Purification of ntrC protein TA3010 cells grown to late exponential phase in 1 liter of medium E ( Vogel and Bonner , 1956 ) were harvested and resuspended in 5 ml of 10 mM K phosphate buffer , pH 7.4 , and treated as described for NCM302 .
A saturated overnight culture of NCM302 was diluted to an A650nni of 0.06 in 500 ml of LB-medium ( Miller , 1972 ) containing ampicillin ( 2.5 Ag/ml ) and grown at 30 °C with aeration to an A65onm of 0.3 .
The temperature was shifted to 42 °C and growth allowed to proceed for 90 min ( A650nm-1.0 ) when the cells were separated by centrifugation at 10 000 g for 15 min , washed with 20 ml of 10 mM K phosphate buffer , pH 7.4 , resuspended in 5 ml of the same buffer , and disrupted by one passage in a French Press cell at 8000 p.s.i. Intact cells were removed by centrifugation at 6000 g for 10 min and the supernatant fraction was applied to a phosphocellulose column ( 1 x 7.5 cm ) equilibrated with 25 mM K phosphate buffer , pH 7.4 containing 1 mM EDTA and 1 mM dithiothreitol .
The retained protein was eluted with a gradient of KCI from 0 to 1.5 M in the same buffer .
Fractions were assayed either individually or as pools , using nick-translated pFA9 ( Higgins and Ames Ferro-Luzzi , 1981 ) or pBR322 DNA as substrate .
Fractions with binding activity and containing protein with the mol .
of the ntrC protein were pooled , dialyzed against 20 mM K phosphate buffer , pH 7.0 , 1 mM f-mercaptoethanol , 5 % glycerol , 1 mM EDTA , and then applied to a hydroxylapatite column ( 0.6 x 7.2 cm ) equilibrated with the same buffer .
Proteins were eluted with a K phosphate gradient from 20 mM to 400 mM .
The peak fraction of activity ( tube 34 ) was used unless indicated differently .
Binding assay The binding mixture consists of ( in an Eppendorf tube ) : 100 , ul of binding buffer ( 50 mM Tris-HCl 7.4 , 10 mM MgC12 , 50 mM KCI , 1 mM DTT , 0.1 mM EDTA ) : 1.5 1l of a solution containing nick-translated and unlabell-ed DNA ( 50 ng,-10 000 c.p.m. , for experiments with pFA9 ) , calf thymus DNA ( 660 ng ) , and bovine serum albumin ( 3 , ug ) ; 5 Al of protein solution containing-1 Ag of DNA-binding protein .
After addition of the protein and incubation at room temperature for 10 min , the entire mix is filtered on a nitrocellulose filter ( Schleicher and Schuell , BA85 ) which had been pre-treated with 0.4 M KOH for 40 min , washed repeatedly in distilled water and equilibrated in 0.1 M Tris-HCl pH 7.4 buffer for at least 60 min .
The filter is washed once with 600 A1 of binding buffer immediately before filtration .
After filtration of the sample , the filter is washed rapidly three times with 600 11 each of binding buffer .
Footprinting The method of Galas and Schmitz ( 1978 ) was followed with minor modifications and utilizing bovine pancreas deoxyribonuclease I ( grade I , Boehr-inger Mannheim ) .
After stopping the reaction , the mixture was extracted once with an equal volume of phenol/chloroform ( 1:1 ) , ethanol-precipitated , washed with ethanol and dried .
The dry sample was dissolved in sequencing sample buffer and electrophoresed on a sequencing acrylamide gel ( 10-20 % as needed ) ( Maniatis et al. , 1982 ) .
SI nuclease mapping of mRNA mRNA was prepared essentially as described ( Mike Gilman , Ph.D. .
Thesis , University of California , Berkeley , 1983 ) from bacterial cultures at A650nn , of 0.8 , grown in either nitrogen-rich or nitrogen-poor minimal-medium ( Stern et al. , 1984 ) with 20 mM NH4CI or 3 mM L-glutamine as a source of nitrogen , respectively , and 0.4 % glucose as a carbon source ; RNase-free DNase I was purchased from Worthington ; the proteinase K treatment was omitted .
The S1 nuclease ( P-L Biochemicals ) treatment was as described by Maniatis et al. ( 1982 ) , using end-labelled DNA fragments to hybridize to the mRNA ( at 54 ' C for 3 h ) , digesting at 30 ' C for 30 min with varying amounts of enzyme , and running the products on a sequencing gel .
DNA labelling DNA fragments were purified , nick-translated , and end-labelled essentially as described by Maniatis et al. ( 1982 ) .
Acknowledgements We thank J.Hirschman and S.Kustu for making available strain NCM302 ; M. Wong and S.Kustu for labelling glnAr and ntrBCr DNA fragments for footprinting experiments ; K.Sei and S.Kustu for making available purified ntrC protein and antibodies to it ; K.Jones for advice on footprinting technology ; R.Jaskot for building strains TA3863 , TA3864 , TA3879 and TA3880 ; J.Brenchley for supplying unpublished portions of the glnAr sequence of S. typhimurium .
This work ws supported by NIH grant AM 12121 to Giovanna Ferro-Luzzi Ames .
References Alvarez-Morales , A. , Dixon , R. and Merrick , M. ( 1984 ) EMBO J. , 3 , 501-507 .
Ames Ferro-Luzzi , G. and Higgins , C.F. ( 1983 ) Trends Biochem .
Ames Ferro-Luzzi , G. and Lever , J.E. ( 1972 ) J. Biol .
and Magasanik , B. ( 1981 ) Proc .
Beynon , J. , Cannon , M. , Buchanan-Wollaston , V. and Cannon , F. ( 1983 ) Cell , 34 , 665-671 .
Brown , S.E. and Ausubel , F.M. ( 1984 ) J. Bacteriol. , 157 , 143-147 .
Covarrubias , A.A. and Bastarrachea , F. ( 1983 ) Mol .
Gaillardin , C.M. and Magasanik , B. ( 1978 ) J. Bacteriol. , 133 , 1329-1338 .
Galas , D.J. and Schmitz , A. ( 1978 ) Nucleic Acids Res. , 5 , 3157-3170 .
Hanau , R. , Koduri , R.K. , Ho , N. and Brenchley , J.E. ( 1983 ) J. Bacteriol. , 155 , 82-89 .
Higgins , C.F. and Ames Ferro-Luzzi , G. ( 1981 ) Proc .
Higgins , C.F. and Ames Ferro-Luzzi , G. ( 1982 ) Proc .
Higgins , C.F. , Haag , P.D. , Nikaido , K. , Ardeshir , F. , Garcia , G. and Ames Ferro-Luzzi , G. ( 1982 ) Nature , 298 , 723-727 .
Hsieh , T. and Brutlag , D. ( 1979 ) Proc .
Kustu , S. , Burton , D. , Garcia , E. , McCarter , L. and McFarland , N. ( 1979a ) Proc .
Kustu , S. , McFarland , N.C. , Hui , S.P. , Esmon , B. and Ames Ferro-Luzzi , G. ( 1979b ) J. Bacteriol. , 138 , 218-234 .
Lee , G.S. and Ames Ferro-Luzzi , G. ( 1984 ) J. Bacteriol. , 159 , 1000-1005 .
Maniatis , T. , Fritsch , E.F. and Sambrook , J. ( 1982 ) Molecular Cloning .
A Laboratory Manual , published by Cold Spring Harbor Laboratory Press , NY .
McCarter , L. , Krajewska-Grynkiewicz , K. , Trinh , D. , Wei , G. and Kustu , S. ( 1984 ) Mol .
McFarland , N. , McCarter , L. , Artz , S. and Kustu , S. ( 1981 ) Proc .
Miller , J.H. ( 1972 ) Experiments in Molecular Genetics , published by Cold Spring Harbor Laboratory Press , NY .
, Sundaresan , V. , Rothstein , D.M. , Brow , S.E. and Ausubel , F.M. ( 1983 ) Proc .
Pahel , G. , Rothstein , D.M. and Magasanik , B. ( 1982 ) J. Bacteriol. , 150 , 202-213 .
Reitzer , L.J. and Magasanik , B. ( 1983 ) Proc .
Roberts , G.P. and Brill , W.J. ( 1981 ) Annu .
Stern , M.J. , Higgins , C.F. and Ames Ferro-Luzzi , G. ( 1984 ) Mol .
Stuber , D. and Bujard , H. ( 1981 ) Proc .
Ueno-Nishio , S. , Backman , K.C. and Magasanik , B. ( 1983 ) J. Bacteriol. , 153 , 1247-1251 .
Ueno-Nishio , S. , Mango , S. , Reitzer , L.J. and Magasanik , B. ( 1984 ) J. Bacteriol. , 160 , 379-384 .
Vogel , H.J. and Bonner , D.M. ( 1956 ) J. Biol .