172 , No. 2 Transcriptional Analysis of the Flagellar Salmonella typhimurium Regulon of KAZUHIRO KUTSUKAKE ,1 * YOSHIKAZU OHYA ,1 AND TETSUO IIN02 Department of Biology , Faculty of Science , University of Tokyo , Hongo , Bunkyo-ku , Tokyo 113,1 and School of Human Sciences , Waseda University , Tokorozawa , Saitama 359,2 Japan Received 26 September 1989/Accepted 21 November 1989 In Salmonella typhimurium , nearly 50 genes are involved in flagellar formation and function and constitute at least 13 different operons .
In this study , we examined the transcriptional interaction among the flagellar operons by combined use of Mu dl ( Apr Lac ) cts62 and TnlO insertion mutants in the flagellar genes .
The results showed that the flagellar operons can be divided into three classes : class I contains only theflhD operon , which is controlled by the cAMP-CAP complex and is required for expression of all of the other flagellar operons ; class U contains seven operons , flgA , flgB , flhB , fliA , fliE , fliF , and fliL , which are under control of class I and are required for the expression of class Ill ; class III contains five operons , flgK , fliDfliC , motA , and tar .
This ordered cascade of transcription closely paralels the assembly of the flagellar structure .
In addition , we found that the fliD defect enhanced expression of the class III operons .
This suggests that the fliD gene product may be responsible for repression of the class Ill operons in the mutants in the class II genes .
These results are compared with the cascade model of the flagellar regulon of Escherichia coli proposed previously ( Y. Komeda , J. Bacteriol .
The bacterial flagellum is composed of three structural components , a basal body , a hook , and a filament .
The filament extends into the extracellular space and is connected by the hook to the basal body embedded in the cell membrane .
Genetic analysis of flagellar mutants have revealed that there are nearly 50 genes involved in flagellar formation and function in both Salmonella typhimurium and Escherichia coli ( 24 , 25 ) .
Intergeneric complementation analysis revealed functional homology in the flagellar genes between these two organisms ( 8 , 21 , 39 ) .
Genes responsible for flagellar formation are calledflg , flh , fli , orflj .
Except for a few genes , mutants defective in these genes are nonflagellate and some of them produce presumptive precursor structures of the flagellum ( 35 , 36 ) .
There are three kinds of genes responsible for flagellar function , including flagellar rotation ( mot ) , chemotaxis ( che ) , and transmembrane signal transduction of chemotactic stimuli ( tar , trg , tsr , etc. ) Mutants defective in these genes produce flagellar structures indistinguishable from those of the wild-type strain .
Most of the flagellar genes are clustered in three regions of the bacterial chromosome , termed regions I , II , and III .
These clustered genes constitute 14 and 13 different operons in E. coli ( 16 , 32 ) and S. typhimurium ( 22 ) , respectively .
Flagellar genes of S. typhimurium are summarized in Fig. 1 .
In E. coli , Komeda ( 14 ) constructed fusions of most of the lac flagellar operons to the genes by using Mu dl ( Apr Lac ) Cohen This bacteriophage developed by Casadaban.and ( 5 ) .
with and its phage contains the lac genes no promoter , integration in a gene can result in rescue of expression of the lac due the of that By using these genes to promoter gene .
operon fusions , he examined the transcriptional interaction the flagellar and proposed the cascade model among operons of the flagellar regulon .
In his model , the flagellar operons are divided into six classes , depending on their transcriptional hierarchy .
Recently , DNA sequence ( 2 , 10 ) and biochemical ( 1 ) analyses suggested that the expression of the flagellar regulon may involve the participation of an alternative sigma factor which binds to core RNA polymerase .
In S. typhimurium there have been several reports which suggested that expression of the flagellar genes is controlled in a similar cascade mode .
For example , Suzuki and lino ( 34 ) reported that mRNA specific for fliC , which encodes flagellin , the component protein of filament , could not be detected in any nonflagellate mutants examined .
However , an overall structure of the flagellar regulon has not yet been established in this organism .
In our previous work ( 22 ) , we isolated Mu dl ( Apr Lac ) cts62 and TnJO insertion mutants in almost all of the flagellar genes in this organism .
In this study , by combined use of these two kinds of insertion mutants , we have examined the transcriptional interaction of flagellar genes .
On the basis of the observed interactions , we proposed a cascade model of the flagellar regulon in which the flagellar operons are divided into three classes with respect to transcriptional hierarchy .
Flagellar gene symbols used in this report follow the new unified nomenclature for S. typhimurium and E. coli ( 11 ) , and each flagellar operon is named after the gene which is transcribed first in that operon ( 22 ) .
MATERIALS AND METHODS Bacteria and phage .
All the bacterial strains used were derivatives of S. typhimurium wild-type strain LT2 .
TnJO insertion mutants in the flagellar genes were derived from and were described LT2 KK1004 or ( LT2 AfljAB ) previously ( 20 , 22 ) .
TnJO insertion mutants in cya and crp were PP1002 and PP1037 ( 29 ) , respectively , which were provided by K. E. Sanderson .
Mu dl insertion mutants in the flagellar genes were derived from KK1005 ( KK1004 galE ) and were described in the previous report ( 22 ) .
Only the Lac ' insertion strains were used in this study .
In order to construct strains carrying both Mu dl and TnJO insertions in two different genes , TnJO insertion mutations were introduced into the Mu dl insertion strains by P22-mediated transduction , which was performed by using P22HTint phage as described previously ( 22 , 23 ) .
74 Media and culture conditions .
Minimal medium , minimal agar plate , L broth , and L broth agar plate were prepared as described previously ( 22 , 23 ) .
Ampicillin and tetracycline were used at final concentrations of 25 and 20 , g/ml , respectively .
5-Bromo-4-chloro-3-indolyl - , B-galactopyra-noside ( X-Gal ) was used at a final concentration of 25 , ug/ml .
Throughout this study , bacterial cells were grown at 30 °C .
, B-Galactosidase activities in tolu-ene-treated cells were assayed by the method of Miller ( 26 ) .
The enzyme units reported here were the average of at least five independent experiments .
The results of the individual assays were all within 20 % of the reported averages .
The endogenous 3-galactosidase activity of the parental strain KK1005 was 1.0 U under the conditions used in this study .
RESULTS Effect of cya and crp mutations on the expression of flagellar genes .
The production of the flagellum in E. coli and S. typhimurium has been shown to be under positive control from the cAMP-CAP complex ( 17 , 30 ) .
In order to confirm this , we introduced a cya : : TnJO or crp : : TnJO mutation into Mu dl insertion mutants in flagellar genes and expression of the lacZ gene in the resulting double-insertion mutants was examined on minimal agar plates containing X-Gal ( Table 1 ) .
As expected , they were all Lac - , indicating that transcription of all the flagellar genes are positively controlled by cAMP-CAP complex .
Transcriptional interaction of the flageliar operons .
TnWO insertion mutations in flagellar genes were introduced into strains carrying fusions of flagellar genes to the lac genes , and expression of the lacZ gene in the resulting double-insertion mutants was examined on minimal agar plate containing X-Gal .
The results are summarized in Table 1 .
Insertions in the different genes in the same operons gave the same results , which suggests that the internal signals of transcription within the polycistronic operons , if any , may not be so significant in our experimental system used here .
Based on the positive interaction of transcription , the fla-gellar operons can be divided into three classes .
Class I comprises only theflhD operon .
This operon did not require functions of any other flagellar genes for its transcription , and genes in it were required for expression of all of the other flagellar operons .
Class II includes seven operons , flgA , flgB , flhB , fliA , fliE , fliF , and fliL .
Their expression required the functions of genes in the class I operon , and their functions are required for expression of the class III operons .
Class III comprises five operons , flgK , fliD , fliC , motA , and tar .
Their expression needed functions of the class I and class II operons , and their functions were not responsible for expression of the other operons .
TnJO insertion mutations in some of the class II genes , such as flgA , flgH , flgI , flgJ , and fliK , were found to allow the class III operon to be expressed partially .
Even the null mutants in these genes , including TnJO insertion ones , produce flagella at a low frequency , probably because these gene products are not absolutely required for flagellar formation and function ( our unpublished results ) .
Therefore , it is likely that this partial expression may be due to leakiness of the mutations .
One surprising finding from this experiment was that the fliD mutation enhanced expression of the other class III operons .
This suggests that the gene product of the fliD operon may have an activity to repress the class III operons , although all the genes responsible for flagellar formation , including the fliD gene , have been identified as the positive factors for the appearance of the flagellar structures .
We propose that this repressible activity be designated as rflA ( repression of flagellar operons ) .
Quantitative assay of I-galactosidase .
In order to obtain quantitative data on expression of the lacZ gene , the activity of 1B-galactosidase was measured in representative fusion strains .
For this purpose , we used strains which carried Mu dl insertions in most promoter-proximal genes available in the individual operons .
For representatives of lac fusion strains in the fliL and tar operons , we used Mu dl insertion mutants in the fliM and cheR genes , respectively , because we had not established Mu dl insertion mutants in eitherfliL or tar .
For those of the other operons , we used strains carrying Mu dl insertions in the genes which are transcribed first in the corresponding operons .
First , the activities in the strains carrying no TnWO insertion mutation were compared .
In general , the operons , gene products of which have been known to constitute flagellar structures ( the flgB , fliD , flgK , and fliC operons ) or to be involved in flagellar function ( the motA and tar operons ) , had higher activity , although the flhD operon showed an exceptionally high activity .
We also measured the activity in the double-insertion mutants .
The results shown in Table 2 are essentially equivalent with those obtained by the qualitative experiment i which minimal agar plate containing X-Gal was used .
The fliD mutation was found to enhance two to six times the expression of the class III operons .
742 A cya Region I crp trg 50 Region II H2 region Region III B Region I Region II Region III H2 region flgA flhA flhB fljA fliA  fljB f lgB flgC fliB cheZ hin flgD flgE flIgF flgG f lgH flgI * flgJ cheY cheB cheR tar fliC fliD fliE cheW cheA motB motA f liF f1liG fliH f iIi fiJi 8 f1iK f1liL f1iiM fliN filgK I flgL tflhC flhD fliP fliQ FIG. 1 .
Flagellar genes in S. typhimurium ( 11 , 21 , 24 , 25 , 29 ) .
( A ) Flagellar genes are shown on the chromosomal map .
Genes responsible for catabolite gene activation , cya and crp , are also indicated .
( B ) The operon structure of clustered flagellar genes are indicated by arrows .
Expression of the lacZ gene fused to promoters of the flagellar operonsa Mudl ( AprLac ) cts62 insertion TnlO insertion * o in .
- GO0 0o , ' o - n sa 0 - .
eC ee c. - e M-t cnc-n-e -- sen-sn nm n c-sm -- n-c * - c '' X o or 4 .
uO - ) : M. * aa. , * aa , .
m M : 4 4 4 h cents 4 e X 14 4 4 o-e e4 - .
o r-0 o M M * - - s -- ne ^ ns-e ms-tsn-csn-ms-csn-s J tD K Y4 an-39 go > co cn o Ha 50 0 0 ) 0 4 O O = = C ¬ .
¬ - W N M Da * 0 0000000 o o o - s. s.s ~ ~ ~ ~ wt ~ ~ - m c-tm ^ - n - ^ n cn e g. C mo a I. .
= 0 n t on o on oo onooco * ccco-4 I h 4 4 e ` irq.5-n ~ ~ - s-m cn oCe .
C e ¬ 5 ~ C -4 0 n.-o 0 4 + + .
+ + + + - * + + + + + + + + + + + + None ------------ ¬ cya crp f lhD ( 2040 ) ------------- f 1 hC ( 2009 ) + ------- ¬ ¬ ----------- ¬ -------- ¬ flgA ( 2640 ) + + + + + + + + + + + + + + + + + + + + + + + + + + ± t t ± t t t t t t t flgB ( 2085 ) f lgC ( 2705 ) f lgD ( 2043 ) f lgF ( 2708 ) f lgG ( 2008 ) flgH ( 2059 ) fIgI ( 2141 ) flgJ ( 2017 ) + + + .
¬ - ¬ ¬ -- ¬ ¬ ¬ ¬ + + + + + + + + + + + + + + + + + + + + + ± ± t ± t ± t t ¬ ¬ ¬ ± ± ± flhB ( 2019 ) + + + .
+ + + + + + + + + + + + + + + + + + + + + fliE ( 2024 ) + + + + + + + + + + + + + + + + + + + + + + + + + + ---------- ¬ fliF ( 2092 ) fliG ( 2119 ) fliH ( 2135 ) fliI ( 2046 ) f liJ ( 2700 ) fliK ( 2143 ) + + + + + + + + + + + + + .
+ + + fliM ( 2071 ) fliN ( 2736 ) f liO ( 2873 ) fliP ( 2053 ) fliQ ( 2130 ) fliR ( 2110 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + .
-- ¬ fliD ( 2601 ) flgK ( 2944 ) 4 flgL ( 2087 ) fliC ( 2604 ) + + + + + + + + + + + + + + + + + + + + + + + + + + + ** * ** * * * * + + + + + + + + + + + + + + + + + + + + .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + -- ¬ + + + + + + + + -- + + ¬ cheR ( 2811 ) cheB ( 2078 ) cheY ( 2014 ) cheZ ( 2668 ) + + + .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a Single colonies of each strain grown on L broth agar plates were transferred onto minimal agar plates containing X-Gal , and the Lac phenotype was examined after incubation at 30 °C .
Although the incubation periods were varied ( from 5 to 24 h ) depending on the transcriptional activity of the lac fusion strains , they were the same among the strains derived from the same lac fusion strains .
Blue color developed in and around the colony was categorized by using the following symbols : + , blue color development ; ± , less blue color development ; - , no color development ; * , dark-blue color development .
The transcriptional orientation of polycistronic operon is indicated by the arrow .
Numbers in parentheses indicate KK numbers of the strains used in this experiment .
DISCUSSION By using fusions of flagellar operons to the lac genes , we have analyzed the effects of mutations in flagellar genes on the expression of each flagellar operon .
The operon fusions were produced by insertion of Mu dl at various flagellar genes , and TnJO insertion mutations were introduced into these fusion strains .
In our experimental system , we could not examine the autogenous regulation of flagellar operons because the lac fusion strains used were defective in the genes at which Mu dl inserted and in the downstream genes in the same operons .
However , the results shown in Tables 1 and 2 suggest the positive fashion of regulation and categorize flagellar operons into three classes .
On the basis of these results , we constructed a hypothetical scheme for the interaction of flagellar operons in the flagellar regulon of S. typhimurium ( Fig. 2A ) .
Only the flhD operon belongs to class I .
This operon contains two genes , flhD and flhC , both of which have been postulated to regulate in a positive fashion expression of all of the other flagellar genes ( 31 ) .
Furthermore , expression of this operon has been thought to be controlled positively by cAMP-CAP complex ( 17 , 30 ) .
Our results clearly confirmed these presumptions .
Except forfliA , which is responsible for filament formation , all of the genes in the class II operons have been shown to be involved in hook-basal body complexes in the morphogenic pathway of flagellar structure ( 35 , 36 ) .
On the other hand , genes in th class III operons are responsible for formation of the filament , the final step of flagellar assembly ( the flgK , fliD , and fliC operons ) or function of the complete flagellum ( the motA and tar operons ) .
These facts indicate that the cascade of the flagellar regulon closely parallels the assembly hierarchy of the flagellar structure .
Therefore , we propose that classes I , II , and III be called the early , middle , and late flagellar operons , respectively ( Fig. 2A ) .
Our scheme presented here is slightly different from that of E. coli described by Komeda ( 14 ) .
In our model , the flagellar regulon consists of only one sequential pathway with three steps of transcription ( Fig. 2A ) .
On the other hand , according to the Komeda scheme , the flagellar regulon shows a branching cascade in which the flagellar operons are divided into six classes , classes 1 through 6 ( Fig. 2B ) .
It is possible that this difference may reflect species specificity between S. typhimurium and E. coli , because it is conceivable that the organisms differ in the details of the regulatory hierarchy .
However , according to the Komeda model , the flgJ gene ( class 4 ) should be expressed only after the operons responsible for formation of hook-basal body structures ( class 3 ) were expressed , which is inconsistent with the results showing that function of the gene is required for flgJ the very early stage in the morphogenic pathway of hookbasal body ( 35 , 36 ) .
In our earlier work ( 22 ) , we showed that the flgJ gene of S. typhimurium does not form an independent operon but belongs to theflgB operon .
This result was supported further by a recent DNA sequence analysis of the upstream region of the flgJ gene ( 12 ) .
If we assume that the flgJ gene might belong to theflgB operon in E. coli as well as in S. typhimurium , the schemes of the flagellar regulon would be very similar between these two organisms .
Correspondences of the classes between the two schemes are as follows ( the former is of the Komeda scheme , and the latter is of ours ) : class 1 to class I , classes 3 and 4 to class II , classes 5 and 6 to class III .
Of the class 2 operons , flgA belongs to class II andfliD belongs to class III .
However , we believe that our classification is correct even in E. coli , because at least oneflgA mutation of E. coli has been shown to reduce expression of the class 5 and class 6 operons ( 15 ) and the upstream region of fliD of E. coli displays the consensus sequence for the class III promoter ( see Fig. 3 ) .
Recently , DNA sequence data on the upstream regions of E. coli and S. typhimurium flagellat operons have been accumulated .
Comparison of these sequences revealed the promoter structures specific for flagellar operons which should be recognized by RNA polymerase containing an alternative sigma factor ( 2 , 10 ) .
In the present work , we divided the flagellar operons into three classes .
This suggests that there may be three different types of regulation of gene expression specific for each class .
Because we can assume that this difference should be reflected by the difference i the sequence of the upstream regulatory sites , we compared once again the preceding sequences of flagellar operons on the basis of our classification .
The results are summarized in Fig. 3 .
The sequences TAAA and GCCGATAA with 15-base-pair spacers , which have been supposed to be the -35 and -10 promoter consensus sequences specific for flagellar operons ( 10 ) , were found to be characteristic of the upstream regions of the class III operons .
Recently , Arnosti and Chamberlin ( 1 ) isolated from E. coli cells an alternative sigma factor ( UF ) which restores specific transcription in-vitro of the tar operon of E. coli when added to core RNA polymerase .
More recently , we obtained several lines of evidence supporting the theory that ( F is encoded by thefliA gene ( manuscript in preparation ) .
These results suggest that aF may be specific for transcription in-vivo of the class III promoter .
On the other hand , the preceding regions of the class II operons , except the fliL operon of E. coli , contain only the putative -10 sequence , which is consistent with the result reported by Bartlett et al. ( 2 ) .
The preceding region of thefliL operon of E. coli seems to contain both the putative -35 and -10 sequences .
Although Arnosti and Chamberlin ( 1 ) showed that RNA polymerase containing uF can transcribe in-vitro thefliL operon of E. coli , it is unlikely that this class III promoterlike sequence might be functional in the transcription in-vivo , because the initiation site of transcription in-vivo of this operon is different from that of transcription in-vitro ( Fig. 3 ) and the transcription in-vivo of this operon does not require the function of the fliA gene ( 14 ) .
+ + + a Single colonies of each strain grown on L broth agar plates were transferred onto minimal agar plates containing X-Gal , and the Lac phenotype was examined after incubation at 30 °C .
Although the incubation periods were varied ( from 5 to 24 h ) depending on the transcriptional activity of the lac fusion strains , they were the same among the strains derived from the same lac fusion strains .
Blue color developed in and around the colony was categorized by using the following symbols : + , blue color development ; ± , less blue color development ; - , no color development ; * , dark-blue color development .
The transcriptional orientation of polycistronic operon is indicated by the arrow .
Numbers in parentheses indicate KK numbers of the strains used in this experiment .
The actual value of , B-galactosidase enzyme units is shown in parentheses .
Strains used were the same as in Table 1 .
A Flagellar regulon ... Class I Class II Class III ( Early ) ( Middle ) ( Late ) flgA flgB flgK flhB fliA fliE fliF fliL fliD ( rflA ) -  fliC cAMP-CAP > - flhD motA Ltar : ... B Flagellar regulon ... I. ... .
Class 2 [ flgA / fliD Class 1 Class 3 [ flgB flhB fliA fliE fliF Class 6 > fliC Class 4 EflgJ ¬ cAMP-CAP-a -LCB- flhD Class 5 [ flgK LfliL LmotA ... .
Hypothetical schemes of the flagellar regulon showing the positive interaction of flagellar which drawn operons , were based on the results obtained in this study ( A ) or from descriptions by Komeda ( 14 ) ( B ) .
The positive interaction for transcription is shown by the arrow .
Each flagellar operon is named after the gene which is transcribed first in that operon .
Details are described in Discussion .
Therefore , we believe that the absence of the putative -35 sequence is characteristic of the class II promoter .
The sequence data suggested that the gene products offlhD and act as factors If this is it is flhC may sigma ( 2 ) .
correct , possible that the presumptive sigma factors encoded by these genes may be specific for transcription of the class II promoter .
The preceding region of the class I operon also contains the sequence similar to the putative -10 sequence .
However , because occurrence of A and C at nucleotides 1 and 5 of the -10 sequence is quite unusual , as compared with the putative -10 sequences of the other flagellar operons , it seems unlikely that this sequence might participate in the promoter structure of this operon .
For the final decision , it is necessary to determine the transcription start sites of all the flagellar operons .
The preceding region of the class I operon also contains the potential cAMP-CAP binding site ( Fig. 3 ) .
This feature is consistent with the result indicating that the transcription of this operon is positively regulated by cya and crp ( Tables 1 and 2 ) .
In the present study , we did not analyze the expression of five flagella-related operons , the trg , tsr , fljB , fliB , and hin operons , because we had not established fusions of these operons to the lac genes .
The upstream sequences of the former three contain both TAAA and GCCGATAA operons consensus sequences with 15-base-pair spacers , which is characteristic of the class III promoters ( Fig. 3 ) .
Because trg and tsr are responsible for flagellar function and fljB is responsible for filament formation ( 24 , 25 ) , the products of these three genes should function in the late stage of flagellar morphogenesis .
These facts suggest that these three genes may belong to the class III operons .
ThefliB gene is involved in methylation of certain lysine residues of flagellin ( 33 ) and residues in III it is region ( Fig. 1 ) .
Therefore , likely thatfliB may also be one of the members of the flagellar regulon .
The hin gene is involved in site-specific inversion of the promoter segment of the fljB operon , which is responsible for phase variation ( 40 ) .
The inversion has been shown to occur even in the flhD mutants ( our unpublished results ) , which indicates that the expression of hin is controlled independently of the flagellar regulon .
Our scheme shown in Fig. 2 indicates that the expression of class III requires the functions of all of 27 genes of class II .
However , it is unlikely that all of these gene products might act as activators for transcription of class III .
In E. coli , Komeda ( 15 ) presented evidence that the positive controller of the later operons ( classes 4 , 5 , and 6 ) is only the fliA gene among the class 3 genes and the flgA gene product functions as repressor whose action can be masked by its interaction with gene products of the class 3 operons .
It is possible that onlyfliA may be the activator gene for class III in S. typhimurium as well as in E. coli , because among the mutants in the class II genes , only thefliA mutants produce a complete hook-basal body lacking a filament portion ( 35 , 36 ) .
In this study , we used anflgA : : TnJO mutant to analyze the effect offlgA defect on the expression of other operons .
If the gene product of flgA has dual functions ( one is a repressor for the expression of class III , and the other is an essential factor for formation of basal body ) , introduction of anflgA : : TnJO mutation should enhance the expression of the class III operons because the TnJO insertion mutant is expected to show a null phenotype .
However , the results shown in Tables 1 and 2 indicate that it also reduced expression of the class III operons .
Therefore , we do not believe at present that the gene product of flgA might function as a repressor in S. typhimurium .
On the other hand , we found that introduction of anfliD : : TnJO mutatio enhanced expression of the class III operons .
This indicates that the fliD gene or an unidentified gene downstream in the fliD operon may be involved in repression of the class III operons in the mutants in the class II genes .
Expression of the flagellar genes in Caulobacter crescentus is also regulated in a cascade fashion ( 6 ) .
Recently , negative regulation was found to be superimposed on the positive control in the flagellar hierarchy of this organism ( 27 , 38 ) .
Therefore , it is plausible that the negative regulation is a general mechanism to coordinate expression of flagellar genes and flagellar assembly in bacteria .
In the quantitative assay of 0-galactosidase , we found that the operons whose gene products have been shown to constitute flageilar structures have activities about 10 times as high as those of the operons whose gene products have not yet been identified in the isolated flagellar structures ( 24 ) .
It is reasonable to assume that the flagellar genes whose products are required in larger quantities for flagellar assembly may be transcribed at higher rates .
If so , transcription rates of flagellar genes could be coordinated with the amounts of protein products required for flagellar morphogenesis .
On the other hand , the activity of the flhD operon was found to be exceptionally high , although neither of the gene products of this operon is the structural constituent of flagellar structure ( 24 ) .
It is possible that an unidentified gene encoding a hypothetical repressor for expression of theflhD operon might exist in the flagellar regulon .
If so , the , Bgalactosidase activity in the flhD-lac fusion strain reported here might represent that of a derepressed state , because the hypothetical repressor gene should not be expressed in that fusion strain .
Examination of autogenous regulation will give some clues to know the manner of transcriptional control of the flhD operon .
DNA sequence comparison of the upstream regions of flagellar operons of S. typhimurium and E. coli .
Sequences are aligned according to Helmann and Chamberlin ( 10 ) and Bartlett et al. ( 2 ) .
E. and S. preceding each gene symbol indicate that the gene originated from E. coli and S. typhimurium , respectively .
Sequence data were adapted from the following references : E. flhD , E. fliL , E. fliF , and E. flhB ( 2 ) ; S. fliL ( 13 ) ; S. fliF and S. flgK ( 12 ) ; E. fliD , E. fliC , S. fliD , S. fliC , and S. fljB ( 37 ) ; E. motA ( 2 and 7 ) ; E. tar ( 18 ) ; S. tar ( 28 ) ; E. trg ( 3 ) ; E. tsr ( 4 ) ; S.flgA , S.flgB , and S.fliA from our unpublished results ( our nucleotide sequence data will appear in the DDBJ , EMBL , and GenBank nucleotide sequence databases under the accession numbers D00497 for fliA and D00498 for flgA and flgB ) .
The double-underlined sequence in E. flhD is a potential cAMP-CAP binding site ( 2 ) .
The boxed or underlined nucleotides indicate the start sites of transcription in-vivo or in-vitro , respectively , whose data were adapted from the following references : E. fliL ( 1 and 19 ) ; E.fliC and E.fliD ( 9 ) ; E. and E. ( 1 ) .
Although tar tsr we have not analyzed expression of the trg , tsr , or fljB operons in this study , these operons were included in class III in this figure .
Details are described in Discussion .
We thank S. Harayama and R. M. Macnab for their encouragement .
This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education , Science and Culture , Japan .
Arnosti , D. N. , and M. J. Chamberlin .
Secondary v factor controls transcription of flagellar and chemotaxis genes in Escherichia coli .
Bartlett , D. H. , B. B. Frantz , and P. Matsumura .
Flagellar transcriptional activators FlbB and FlaI : gene sequences and 5 ' consensus sequences of operons under FlbB and Flal control .
Bollinger , J. , C. Park , S. Harayama , and G. Hazelbauer .
Structure of the Trg protein : homologies with and differences from other sensory transducers of Escherichia coli .
Boyd , A. , K. Kendall , and M. Simon .
Structure of the serine chemoreceptor in Escherichia coli .
Casadaban , M. J. , and S. N. Cohen .
Lactose genes fused to exogenous promoters in one step using a Mu-lac bacteriophage : in-vivo probe for transcriptional control sequences .
Champer , R. , A. Dingwall , and L. Shapiro .
Cascade regulation of Caulobacter flagellar and chemotaxis genes .
Dean , G. E. , R. M. Macnab , J. Stader , P. Matsumura , and C. Burks .
Gene sequence and predicted amino-acid sequence of the motA protein , a membrane-associated protein required for flagellar rotation in Escherichia coli .
DeFranco , A. L. , J. S. Parkinson , and D. E. Koshland , Jr. 1979 .
Functional homology of chemotaxis genes in Escherichia coli and Salmonella typhimurium .
Hanafusa , T. , A. Sakai , A. Tonmnaga , and M. Enomoto .
Isolation and characterization of Escherichia coli hag operator mutants whose hag48 expression has become repressible by a Salmonella HI Mol .
Hlmann , J. D. , and M. J. Chamberlin .
DNA sequence analysis suggests that expression of flagellar and chemotaxis genes in Escherichia coli and Salmonella typhimurium is controlled by an alternative a factor .
fino , T. , Y. Komeda , K. Kutsukake , R. Macnab , P. Matsumura , J. S. Parkinson , M. I. Simon , and S. Yamaguchi .
New unified nomenclature for the flagellar genes of Escherichia coli and Salmonella typhimurium .
Jones , C. J. , M. Homma , and R. M. Macnab .
L - , P - , and M-ring proteins of the flagellar basal body of Salmonella typhi-murium : gene sequences and deduced protein sequences .
Khhara , M. , M. Homma , K. Kutsukake , and R. M. Macnab .
Flagellar switch of Salmonella typhimurium : gene sequences and deduced protein sequences .
Fusions of flagellar operons to lactose genes on a Mu lac bacteriophage .
Transcriptional control of flagellar genes in Escherichia coli K-12 .
Komeda , Y. , K. Kutsukake , and T. Ilno .
Definition of additional flagellar genes in Escherichia coli K-12 .
Komeda , Y. , H. Suzuki , J. Ishidsu , and T. Iino .
The role of cAMP in flagellation of Salmonella typhimurium .
Krikos , A. , N. Mutoh , A. Boyd , and M. Simon .
Sensory transducers of E. coli are composed of discrete structural and functional domains .
Kuo , S. C. , and D. E. Koshland , Jr. 1986 .
Sequence of flaA ( cheC ) locus of Escherichia coli and discovery of a new gene .
Kutsukake , K. , and T. Iino .
Refined genetic analysis of the region II che mutants in Salmonella typhimurium .
Kutsukake , K. , T. Iino , Y. Komeda , and S. Yamaguchi .
Functional homology offla genes between Salmonella typhimu-rium and Escherichia coli .
Kutsukake , K. , Y. Ohya , S. Yamaguchi , and T. Iino .
Operon structure of flagellar genes in Salmonella typhimurium .
Kutsukake , K. , T. Suzuki , S. Yamaguchi , and T. Iino .
Role of gene flaFV on flagellar hook formation in Salmonella typhi-murium .
In F. C. Neidhardt , J. Ingraham , K. B. Low , B. Magasanik , M. Schaechter , and H. E. Umbarger ( ed .
) , Escherichia coli and Salmonella typhimurium : cellular and molecular biology , vol .
American Society for Microbiology , Washington , D.C. 25 .
Motility and chemotaxis , p. 732-759 .
In F. C. Neidhardt , J. Ingraham , K. B. Low , B. Magasanik , M. Schaechter , and H. E. Umbarger ( ed .
) , Escherichia coli and Salmonella typhimurium : cellular and molecular biology , vol .
American Society for Microbiology , Washington , D.C. 26 .
Experiments in molecular genetics .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 27 .
Newton , A. , N. Ohta , G. Ramakrishnan , D. Mufn , and G. Raymond .
Genetic switching in the flagellar gene hierarchy of Caulobacter requires negative as well as positive regulation of transcription .
Russo , A. F. , and D. E. Koshiand , Jr. 1983 .
Separation of signal transduction and adaptation functions of the aspartate receptor in bacterial sensing .
Sanderson , K. E. , and J. R. Roth .
Linkage map of Salmonella typhimurium , edition VII .
Silverman , M. , and M. Simon .
Characterization of Esch-erichia coli flagellar mutants that are insensitive to catabolite-repression .
Silverman , M. , and M. I. Simon .
Slocum , M. K. , and J. S. Parkinson .
Genetics of methyl-accepting chemotaxis proteins in Escherichia coli : organization of the tar region .
Stocker , B. A. D. , M. W. McDonough , and R. P. Ambler .
A gene determining presence or absence of r-N-methyl-lysine in Salmonella flagellar protein .
Suzuki , H. , and T. lino .
Absence of messenger ribonucleic acid specific for flagellin in non-flagellate mutants of Salmo-nella .
Suzuki , T. , T. lino , T. Horiguchi , and S. Yamaguchi .
Suzuki , T. , and Y. Komeda .
Incomplete flagellar structures in Escherichia coli mutants .
Szekely , E. , and M. Simon .
DNA sequence adjacent to flagellar genes and evolution of flagellar-phase variation .
Xu , H. , A. Dingwall , and L. Shapiro .
Negative transcriptional regulation in the Caulobacter flagellar hierarchy .
Yamaguchi , S. , H. Fujita , T. Taira , K. Kutsukake , M. Homma , and T. lino .
Genetic analysis of three additionalfla genes in Salmonella typhimurium .
Zieg , J. , and M. Simon .
Analysis of the nucleotide sequence of an invertible controlling element .