173 , No. 7 Negative Regulatory Loci Coupling Flagellin Synthesis to Flagellar Assembly in Salmonella typhimurium KAREN L. GILLEN AND KELLY T. HUGHES * Department of Microbiology SC42 , University of Washington , Seattle , Washington 98195 Received 13 November 1990/Accepted 28 January 1991 The complex regulation of flagellin gene expression in Salmonella typhimurium was characterized in-vivo by using lac transcriptional-fusions to the two flagellin structural genes ( fliC [ Hl ] andfljB [ H2 ] ) .
Phase variation was measured as the rate of switching of flagellin gene expression .
Switching frequencies varied from 1/500 per cell per generation to 1/10 ,000 per cell per generation depending on the particular insertion and the direction of switching .
There is a 4-to 20-fold bias in favor of switching from thefljB ( On ) to thefljB ( Of ) orientation .
Random TnlOdTc insertions were isolated which failed to express flagellin .
While most of these insertions mapped to loci known to be required for flagellin expression , several new loci were identified .
The presence of functional copies of all of the genes responsible for complete flagellar assembly , except the hook-associated proteins ( flgK , flgL , andfliD gene products ) , were required for expression of thefliC orfljB flagellin genes .
Two novel loci involved in negative regulation offliC andfljB infla mutant backgrounds were identified .
One of these loci , designated the flgR locus , mapped to the flg operon at 23 min on the Salmonella linkage map .
An flgR insertion mutation resulted in relief of repression of thefliC andfljB genes in allfla mutant backgrounds except for mutants in the positive regulatory loci ( flhC , flhD , and fli4 genes ) .
Flagellar phase variation in Salmonella typhimurium results from the reversible inversion of a 996-bp chromosomal segment of DNA ( Fig. 1 ) ( 56-58 ) .
Contained within this segment are the coding region for Hin recombinase , which mediates the inversion reaction , and a promoter for the operon encoding one of the two flagellins located immediately adjacent to the invertible segment ( 43 ) .
Hin mediates recombination between the hixL and hixR recombination sites which flank the invertible segment One orientation ( 24 ) .
of the invertible segment allows expression of the fljB gene , encoding H2 flagellin , and the fljA gene , encoding rHl , a repressor of the fliC gene ( Hi flagellin ) .
When inversion of the fragment occurs , neither thefljB gene nor thefljA gene is expressed , and subsequently the gene for Hi flagellin , fliC , located 16 min away on the Salmonella chromosome , is expressed from the derepressed fliC promoter ( 22 , 44 ) .
This alternate expression of either the fliC or fljB gene results in the production of one of two antigenically distinct proteins ( 21 , 47 ) .
Several thousand copies of flagellin protein are produced in a normal flagellated cell and are assembled to form the filament , the external portion of the flagellum ( 35 ) .
The two other parts of the flagellum are known as the hook , which connects the filament to the basal body , and the basal body , which comprises those proteins involved in membrane attachment of the organelle and in the assembly and function of the motor , which enables rotation of the flagellum resulting in motility ( 25 , 35 ) .
Apart from the fljB and fliC genes , as many as 33 other genes appear to be directly involved in the assembly of the complete flagellar structure ( 23 , 25 ) .
These genes are clustered in four regions of the chromosome ( 33 , 40 ) .
Region I contains theflg genes in three operons , region II containsflh genes in two operons , region III contains the fli genes , including the fliC gene , in at least four operons , and region IV includes the fljAB operon and the hin gene .
These genes together constitute a regulon .
Research both in S. typhimu-rium and with the homologous genes in Escherichia coli has uncovered a complex hierarchy of regulation for these genes ( 26 , 31 , 32 , 53 ) .
In S. typhimurium one and presumably also the other of the flagellin genes are at the bottom of this regulatory scheme and are grouped in the class of late genes ( Fig. 2 ) ( 32 ) .
The members of this class are dependent on the early and middle classes any gene for their expression .
If of the genes in the early and middle classes do not express functional proteins , the genes in the late class are not expressed .
Similarly , the genes in the middle class are dependent upon the early class for their expression but are regulated independently of any defects in the late genes ( 32 ) .
The early genes include the flhD and flhC genes , which are believed to encode positive activators of flagellar expression ( 29 ) .
These early genes are required for expression of the middle genes .
The fliA gene , shown recently to encode a flagellar-specific sigma factor ( 38 ) , is included among the middle genes .
The fliA gene product is thought to be required for the expression of the late genes , including flagellin .
At the top of this regulatory scheme is the cAMP-Catab-olite gene activator protein complex , which acts at the early genes ( flhDC operon ) ( 29 , 42 ) .
In cya or crp mutant strains or in wild-type cells grown in the presence of glucose , no flagella are synthesized ( 1 , 54 ) .
These proposed classes or levels of regulation of flagellar assembly in S. typhimurium are similar to those proposed for E. coli by Komeda and Iino ( 28 ) .
The mechanism by which this regulation occurs is not known for either organism .
To better study the regulation of the flagellin structural genes , Mud-lac operon fusions ( 6 ) to these genes were constructed .
Mud-lac cassettes contain the for the genes lactose operon , without the lac promoter , and an antibiotic resistance gene ( 13 ) .
Surrounding these genes are the ends of Mu phage , which are required for transposition .
Insertion of Mud-lac into a gene can occur in the presence of Mu transposase , supplied in cis or in trans ( 17 , 18 ) .
The result is an operon fusion such that the expression of the lactose operon is dependent on the transcription of the gene into which it is inserted .
When Mud-lac is inserted into the fliC gene or the fljB gene , switching of the invertible segment can be observed as Lac ' Lac ' a change from Lac-to ( or to Lac - ) on indicator plates .
Strains containing Mud-lac fusions tofljB orfliC were used to measure the frequency of switching between expression of the fljB gene and expression of the fliC gene .
These fusion strains were also used to search for genetic loci which affected switching or expression of flagellin genes .
Transposons randomly inserted into these strains yielded known as well as novel sites which affected the expression of flagellin genes .
The observation that some transposon insertions near the fliC gene turned off expression offljB led to the study offliC andfljB expression in characterized flagellar mutant strains .
The Mud-lac fusions to the fliC and fljB genes were moved into each of 39 flagellar mutant strains to observe the effect on expression of the lac operon ( and hence of thefljB orfliC gene ) .
The results obtained were in agreement with the proposed classes of regulation discussed above .
However , unexpected observations with a deletion mutant strain stimulated a search for a repressor of flagellin expression and resulted in the identification of two novel regulatory loci .
230 MATERIALS AND METHODS Bacterial strains .
All strains used in this study are listed in Table 1 .
All S. typhimurium strains were derived from S. typhimurium LT2 .
Mutants defective in flagellar assembly are collectively termedfla mutants and map to theflg , flh , fli , and flj operons .
A derivative , MudJ , of the Mu d ( lac ) phage described by Casadaban and Cohen ( 5 ) was used in this work .
MudJ refers to the transposition-defective lac operon fusion vector Mu d1734 ( Km lac ) described by Castilho et al. ( 6 ) .
This phage is deleted for transposition functions and carries kanamycin resistance in place of ampicillin resistance .
MudJ insertion mutants are isolated by providing transposition functions in cis on a single P22-transducing fragment ( 18 ) .
During this process , the MudJ insertion transposes from the transduced fragment into the recipient chromosome , while the remaining fragment , including the transposition genes , is degraded , leaving a MudJ insertion in the absence of any transposition function .
GILLEN AND HUGHES J. BACTERIOL .
P P hin P. , .
fltB flIA fIiC 0o 0 H2 flagellin repressor Site-specific recombination of HI flagellin flIB W , P flic tfiJA I ~ - b Hi flagellin FIG. 1 .
Flagellar phase variation in S. typhimurium .
This diagram shows the reversible inversion of the promoter of thefljBA operon which encodes H2 flagellin ( fljB ) and repressor ( fljA ) of the Hi flagellin structural gene ( fliC ) .
When this operon is transcribed ( top ) , H2 flagellin is produced and Hi is not .
Upon inversion , the fljBA operon is not expressed ( bottom ) ; H2 flagellin and repressor offliC are not synthesized so that only Hi flagellin is produced .
Early Genes Middle Genes Late Genes FlgR flgBA flg KL fliD flg B CD EF G HIJ flhBA fliA IfliB fliE fliFGHIJK fliLMNOPQR IfiC motA B che A W tar che R B YZ fljBA cAMP-CAP * .
flh CD > FliA FIG. 2 .
Regulatory cascade for flagellar synthesis in S. typhimurium ( 25 ) .
The cAMP-catabolite gene activator protein ( cAMP-CAP ) complex is required for expression of the early genes ( the flhDC operon ) .
The early genes are required for expression of the middle genes ; these include genes whose products are part of the hook-basal body complex .
One of the middle genes , fliA , encodes an alternative sigma factor which is presumably required for expression of the late genes ( 38 ) .
The FlgR protein acts to repress transcription of the late genes in strains defective for flagellar synthesis TABLE 1 .
The E medium of Vogel and Bonner ( 49 ) , supplemented with 0.2 % glucose , was used as minimal-medium .
Carbon-free minimal-medium ( NCE ) ( 37 ) was used in minimal-medium supplemented with carbon sources other than glucose .
Difco nutrient broth ( NB ; 8 g/liter ) , with NaCl ( 5 g/liter ) added , was used as rich-medium for growing cells .
Luria-Bertani ( LB ) medium ( 8 ) , supplemented with E salts and 0.2 % glucose , was used as rich-medium for growing P22 phage lysates .
Difco agar was added to a final concentration of 1.5 % for solid medium .
Motility plates ( 41 ) contained tryptone ( Difco ; 10 g/liter ) , NaCl ( 5 g/liter ) , and 0.35 % Bacto-Agar .
Bochner tetracycline-sensitive selection plates ( 2 ) were used as modified by Maloy and Nunn ( 36 ) .
Auxo-trophic supplements were included in media at final concentrations suggested by Davis et al. ( 8 ) .
The following additives were included in media as needed ( final concentration given ) : X-Gal ( 5-bromo-4-chloro-3-indolyl-3-D-galactoside ; 100 , ug/ml in rich-medium , 25 Fg/ml in minimal-medium ) , tetracycline hydrochloride ( 25 ig/ml in rich-medium , 10 , ug/ml in minimal-medium ) , kanamycin sulfate ( 50 , ug/ml in rich-medium , 125 , ug/ml in minimal-medium ) , ampicillin ( 100 , ug/ml in rich-medium for plasmid selection ) , and chloramphenicol ( 25 , ug/ml in rich-medium , 5 , ug/ml in minimal-medium ) .
Indicator plates with triphenyl tetrazolium chloride ( TTC ) dye were used to detect lactose utilization and flagellar phase variation with fliC : : MudJ and fljB : : MudJ fusions .
TTC-Lac indicator plates ( in which TTC was used as an acid or base indicator ) ( 34 , 51 ) , contained Bacto-Agar ( 15 g/liter ) , tryptone ( 10 g/liter ) , yeast extract ( 1 g/liter ) , NaCl ( 5 g/liter ) , and lactose ( 10 g/liter ) .
TTC ( 50 , ug/ml ) was dissolved in water , filter sterilized , and added before the plates were poured .
On TTC-Lac indicator plates , Lac ' colonies are white while Lac-colonies are red .
Bochner TTC-Lac indicator plates were also used in this study ( 3 ) .
They contained Bacto-Agar ( 15 g/liter ) , Bacto-Peptone ( 2.5 g/liter ) , K2HPO4 ( 9.0 g/liter ) , KH2PO4 ( 3.0 g/liter ) , and lactose ( 2 g/liter ) .
TTC ( 25 , ug/ml ) was added before being poured from a filter-sterilized solution in H20 .
On Bochner TTC-Lac indicator plates , Lac ' colonies are red while Lac-colonies are white .
The TTC-Lac indicator plates utilize TTC as an acid or base indicator , while the Bochner TTC-Lac plates utilize TTC as an indicator of the reductive capacity of the cell .
For all transductional crosses , the high-frequency generalized transducing mutant of bacteriophage P22 ( HTI0511 int-201 ) was used ( 39 ) .
Selective plates were spread directly with 2 x 108 cells and 108 to 109 phage .
Transductants were purified , and phage-free clones were isolated by streaking onto nonselective green indicator plates ( 7 ) .
P22 lysates were titered by the method of Davis et al. ( 8 ) .
S. typhimurium swim from nutrient-limiting media to nutrient-rich media , and the cells are dependent on flagellar rotation for motility .
Cells stabbed into a motility plate will swim out ( swarm ) from the initial stab as a circular wave as local nutrients are exhausted .
If cells are defective in their ability to swarm , because of a defect in either the chemotactic apparatus or the flagellar structure , they are said to be nonmotile or Mot - .
Isolation of TnlOdTc , TnlOdCm , and MudJ insertion mutants .
TnJOdTc refers to a 3-kb transposition-defective de-rivative of TnJO ( TnlO Dell6 Dell7 Tcr ) constructed by Way et al. ( 50 ) .
TnJOdCm refers to a 1-kb transposition-defective derivative of TnJO ( 10 ) .
Insertions were isolated by growing P22 transducing phage on strain TT10423 ( TnlOdTc ) or TT10604 ( TnJOdCm ) and by using this phage stock as a donor to transduce the LT2 strain ( TT10427 ) carrying the TnWO transposase-producing plasmid , pNK972 ( 50 ) , to Tcr onto NB-tetracycline plates for TnJOdTc insertion mutagenesis or to Cmr on NB-chloramphenicol plates for TnJOdCm mutagenesis .
The colonies on these plates were pooled , and a P22 transducing lysate was prepared on them and used to transduce recipient cells to Tcr or Cmr .
Techniques for the isolation of MudJ insertions in the Salmonella chromosome have been described ( 18 ) .
Insertions of MudJ into the hin and fljB genes were isolated by growing P22 transducing phage on strain TT10288 and using this phage stock as a donor to transduce a strain deleted for fliC ( H1 ) and with a TnWO insertion linked to fljB to MudJ-encoded kanamycin resistance .
all strains indicated otherwise , Tcr on NB-tetracycline plates .
The plates were replica printed to NB-kanamycin plates to screen for MudJ insertion alleles linked to the TnlO .
Putative fljB-linked MudJ insertion mutants were screened for the loss of either Hin or H2-flagellin and for their inability to swarm on motility agar .
Mot-insertion mutants were then screened for switching on TTC-lactose indicator plates .
A MudJ insertion in the fliC ( H1 ) gene was isolated in a manner similar to the isolation of fljB ( H2 ) : : MudJ insertion mutants .
P22 transducing phage grown on strain TH10288 was used to transduce a strain with a TnJOdCm insertion linked tofliC to MudJ-encoded Kmr .
The Kmr colonies were pooled ; P22 transducing phage was prepared from the pooled cells and used to transduce LT2 to both Cmr ( TnJOdCm ) and Kmr ( MudJ ) on TTC-Lac plates containing chloramphenicol and kanamycin .
Those colonies which exhibited switching on TTC-Lac indicator plates were analyzed further as potential insertions in fliC .
The putative fliC : : MudJ insertions were transduced into afljB deletion mutant and screened for loss of motility on motility agar .
One isolate was obtained which was Mot-in a AfljB strain and Mot ' in LT2 .
That isolate was also found to be > 99 % linked to a known fliC : : TnJO insertion ( obtained from M. Homma ) , and DNA sequence analysis showed that the MudJ insertion was within the fliC structural gene ( data not shown ) .
Isolation of TnlOdCm insertions in and nearfliC .
To study flagellin expression , we first sought to generate lac operon fusions to the two flagellin genes by Mud-lac ( MudJ ) insertion .
To carry out mutagenesis in the fliC region of the chromosome with the MudJ transposon , it was first necessary to isolate TnJO insertions in the fliC region .
We chose the TnWO derivative TnJOdCm for this purpose .
Strain MH111 ( fiC : : TnlO ) was transduced to Cmr with a P22 transducing lysate grown on a pool of random TnJOdCm insertion mutants .
By screening for Cmr transductants which were simultaneously transduced to Tcs , insertions of TnJOdCm linked to the fliC gene were isolated .
Of 14 Cmr Tcs mutants isolated , one was 100 % linked by P22 transduction to thefliC : : TnJO insertion and was Mot-only in afljB background , indicating that TnJOdCm was inserted in the fliC gene .
Seven of the 14 Cmr Tcs insertion mutants were Mot-in bothfljBl andfljB strains , indicating that they were probably inserted into other genes of the fli operons .
The remaining four Cmr Tcs insertion mutants did not affect flagellin synthesis .
Construction of strains with either a fljB : : MudJ or a fliC : : Mudj fusion in flagellar mutant backgrounds .
Previously characterized flagellar mutant strains obtained from S. Yamaguchi were derived from a parent S. typhimurium strain in which the hin region of the chromosome originated from Salmonella abortus-equi with the genotype vh2 ( Off ) ( vh2 is the equivalent of the S. typhimurium hin gene ) , fljBe , n , X ( 30 , 52 , 53 ) .
These strains are phenotypically Hin - .
Derivatives of these strains carrying the hin + region of S. typhimurium ( fljBl `` 2 ) were made by introducing afljB : : MudJ insertion from strain TH714 by P22 transduction .
Repair of the hin region was checked by transducing the fljB : : MudJ insertion from the repaired strain to strain KS724 , which is missing the region of the chromosome from tct through hin andfljB ( 46 ) , and confirming that switching occurred in these strains .
To repair this region and leave afljB + allele , a strain containing afljB-linked TnJOdCm insertion mutation ( TH529 ) was used as a donor to transduce the Fla-fljB : : MudJ strains to fljB + ( Kms ) .
These repaired strains were then transduced to Kmr with phage grown on a fliC : : MudJ-containing strain ( TH1077 ) to observe expressio from the fliC promoter .
To move TnJOdCm insertions isolated in putative repressor genes into the fliC : : MudJ ( Cmr ) fla backgrounds , further manipulations were required .
The fla fljB : : MudJ ( Cms ) strains were transduced to Kms with a tct : : TnJO marker ( KS203 ) linked to fljB .
Tcr Kmsfla strains were transduced with wild-type strain LT2 to Tct + ( Tcs ) on NCE-isocitrate plates .
The fliC : : MudJ and then flgR : : TnJOdCm insertions were moved sequentially into these backgrounds .
Chromosomal mapping with MudP22 .
MudP22 refers to a defective P22 genome flanked by the Mu ends .
A set of 70 MudP22 insertion mutants was constructed and provided by N. Benson ( la ) .
These MudP22 insertions map throughout the Salmonella chromosome .
When a MudP22 insertion is induced with mitomycin C ( 2 , ug/ml ; Sigma Chemical Co. ) , the resulting lysate includes , on average , 20 kb of phage DNA and 100 kb of adjacent chromosomal DNA ( 45 , 55 ) .
Thus , each lysate from a specific MudP22 insertion provides a high-frequency transducing lysate for 100-kb segments of the Salmonella chromosome .
The 70 MudP22 insertions provided in the mapping kit cover the entire Salmonella chromosome and can be used to map a marker to within a 2-min segment of the chromosome .
Lysates prepared on the MudP22 library were loaded into individual microtiter dish wells and transferred onto a lawn of recipient cells .
For mapping TnJOdTc insertions , cells were plated onto modified Bochner tetracycline-sensitive plates ( 2 , 36 ) screening for lysates which could transduce the TnJOdTc insertion mutant to Tcs with the wild-type region of the chromosome .
Determination of Hin-mediated switching frequencies .
S. typhimurium strains having either fliC : : MudJ or fljB : : MudJ insertions were isolated as single colonies on TTC-Lac indicator plates .
Either a single Lac ' ( white ) colony or a single Lac - ( red ) colony was used to inoculate 30 ml of L broth .
Cultures were grown to mid-log phase ( 100 Klett units ) , diluted , and plated for single colonies on TTC-Lac indicator plates .
The number of Lac ' and Lac-colonies were counted , and only plates which contained 30 to 300 colonies were included in the results .
Colonies of mixed phenotype were present at a low frequency and were considered to result from cells which had switched after the initial plated cell had divided .
These were counted as nonswitched colonies .
The number of colonies counted in each experiment was greater than 1,000 , and each experiment was repeated at least 10 times .
Switching frequencies were determined as the number of switched colonies per total cells counted per cell divisions prior to plating .
The highest and lowest numbers from 10 trials were discarded , and the remaining numbers were averaged .
P fljB ' lacZ , Y , A kanr hixR rv hin hIxL 2 ° 1It 20 hlxR fljB ' lacZ , Y , A kanr .
I G ( U , , ,144 hixL Lac 4d FIG. 3 .
Fusion of the fljBA promoter to the lac operon by using a MudJ insertion in the fljB gene results in the reversible expression of the lac operon mediated by the Hin recombinase .
Hin recombines at the hixL and hixR recombination sites which flank the invertible segment .
The hix2 ° site ( see Discussion ) binds purified Hin in-vitro and is located upstream of the hin translation initiation codon ( 10 ) .
RESULTS Isolation of Mud-lac fusions to the hin and flagellin structural genes .
Previously , antibodies specific to either the Hi or H2 flagellin were used to distinguish between phases of flagellin expression by the use of agglutination assays ( 47 ) .
These assays are time consuming and can not be performed easily on large numbers of colonies .
To characterize flagellar phase variation in S. typhimurium , it was necessary to develop a simple in-vivo assay for the site-specific DNA inversion reaction .
Fusions of MudJ to the flagellar structural genes were used to simplify this procedure ( see Materials and Methods ) .
Four types of indicator plates were tested to assay in-vivo switching between Lac ' and Lac-phenotypes : X-Gal-containing plates , MacConkey-lactose plates , TTC-Lac plates , and Bochner TTC-Lac plates .
Starting with a Lac-orientation , each colony appeared to be Lac ' on both X-Gal-containing plates and MacConkey-lactose plates .
The most useful lactose indicator medium for detecting switching contained TTC .
On TTC-containing plates , higher levels of P-galactosidase were necessary to achieve a Lac ' phenotype .
By using TTC-Lac indicator plates , insertions of MudJ nearfljB andfliC were isolated which would switch between a Lac ' and a Lac-phenotype ( Fig. 3 ) .
Fusions of MudJ to either the fliC or fljB structural genes were obtained by using localized mutagenesis .
Insertions of MudJ were isolated which were linked either to TnJO insertions mapping nearfljB or to TnJOdCm insertions linked to fliC .
As expected , the fliC : : MudJ and fljB : : MudJ insertion mutants were completely nonmotile in the absence of the other flagellin gene , while they switched between a Mot ' and Mot-phenotype when a functional flagellin gene was present .
In addition , Mot-insertions of MudJ were isolated in monophasic Hi or H2 strains .
These insertion mutants were linked to fliC orfljB but did not switch between Lac ' and Lac-phenotypes .
They were presumed to include MudJ insertions in either the hin or the fljB gene which were oriented so that lac was not transcribed .
In vivo switching rates withfljB : : Mud-lac andfiiC : : Mud-lac fusions .
Expression of the lac operon from the above fusions was used to measure in-vivo switching rates between fla-gellar phases .
The results ( Table 2 ) show that there is some variability depending on the fusion used , but the overall range in switching frequency is 1/240 to 1/5 ,900 per cell per generation .
These frequencies are about two-to fourfold higher than previously reported ( 47 ) but do show , as was observed previously ( 47 ) , that there is a bias in the direction of switching : the observed rate of switching was higher from the fliC ( Off ) to the fliC ( On ) [ fljB ( On ) tofljB ( Off ) ] orientation than the reverse .
Isolation of chromosomal insertions defective in flagellin expression .
Because of the coupling of flagellin synthesis with the assembly of flagella , strains with MudJ insertions in the flagellin genes which also carry a mutation in any of the other genes required for flagellar assembly would be expected to give a Lac-Switch-phenotype on TTC-Lac plates ( 32 ) .
Because Lac-colonies of fusion strains show switching to Lac ' , then , starting with a MudJ ( Off ) fusion , mutants defective in switching or flagellin expression could be isolated as being Switch - ( Lac - ) on TTC-Lac plates .
More than 200 TnJOdTc insertions were isolated which prevented conversion of a fljB : : MudJ ( Off ) ( Lac - ) insertion to Lac ' on TTC-Lac indicator plates .
A total of 64 insertions were isolated that did not affect lactose utilization : these did not prevent expression of an intact lac operon on an F factor .
The majority affected lactose utilization and were unrelated to flagellin synthesis .
The insertions that did not affect lactose utilization were mapped by using the MudP22 chromosomal mapping technique of Benson and Goldman ( see Materials and Methods ) ( Fig. 4 ) .
Most of the insertions mapped to loci known to be required for flagellin expression , including the flg , flh , and fli operons as well as the crp and cya loci .
In addition to these expected classes of mutants , a number of insertions were isolated which affected either flagellin expression or switching and mapped to loci not previously characterized .
These may include the genes encoding recently identified flagellar proteins which are dependent on the flagellar regulatory system for expression ( 25 ) .
The role of these novel classes of insertion mutants is currently under investigation .
Effect of flagellar mutants on flagellin expression .
The flagellin structural genes encode the last step in the hierarchy of flagellar biosynthesis .
It has been shown by Kutsukake et al. that fliC is not efficiently expressed when cells are defective in earlier steps in the flagellar biosynthetic pathway ( 32 ) .
We characterized the effect of fla alleles on lac transcription with bothfliC : : MudJ andfljB : : MudJ fusions .
A set of mutants defective in the individual genes of the flagellar assembly pathway was obtained from S. Yamagu-chi .
Either afliC : : MudJ or afljB : : MudJ allele was placed in each fla mutant background , and the cells were assayed for transcription from the flagellin promoter on lactose indicator plates .
The effect offla mutations on flagellin gene transcription is presented in Table 3 .
Mutant alleles of 32 different fla genes were tested , and four were found that did not affect fliC or fljB transcription .
These included mutations in fliC as expected and also mutations in the flgK , flgL , and fliD genes ( encoding hook-associated proteins 1 , 3 , and 2 , respectively ) .
Attachment of the hook-associated proteins immediately precedes flagellin attachment , and these genes are in the same group as the flagellin genes in the regulatory cascade .
These results agree with those described by Kutsukake et al. forfliC expression ( 32 ) .
Identification of novel regulatory loci .
In addition to testing the effects of individualfla alleles on flagellin gene transcription , we also tested strains which were deleted for several flagellar assembly genes ( Table 3 ) .
Since these deletions include single genes shown to be required for fliC and fljB expression , they were also expected to be defective in flagellin expression .
We were surprised to find one deletion Aflg ( A-J ) which did not affect the transcription of eitherfliC or fljB .
One explanation for this result is that the Aflg ( A-J ) deletion , in addition to removing theflgA throughflgJ genes , also removed a locus encoding a negative regulator produced or activated when cells are defective in flagellar assembly .
Such a negative regulator would presumably act either directly or indirectly to prevent flagellin expression .
If a negative regulator were produced when cells are defective in one of the flagellin genes other than flgK , flgL , fliD , fliC , and fljB , then loss of function of this regulator would allow expression of the flagellin genes in a fla fliC : : MudJ orfljB : : MudJ fusion strain .
Indeed , insertions of TnJOdTc and TnJOdCm were isolated which resulted in lac transcription from an fljB : : MudJ fusion in a fla mutant background .
Based on their map locations and phenotypes , these insertions were of three classes .
Class I insertions were linked to the flg region , mapping between flgA and pyrC .
This class , designatedflgR , probably includes a negative regulator which lies within the flg ( A-J ) deletion region .
Class II insertions were unlinked to theflg region , and their map location has not been determined .
Class III insertions were very tightly linked to thefljB : : MudJ fusion and do not affect the expression of afliC : : MudJ fusion .
This class may introduce a new promoter from which the lac be operon can transcribed .
Effect of a flgR : : TnlOdCm insertion mutant on flagellin expression in fla mutant strains .
To determine which fla mutants resulted in repression of flagellin transcription through the action of the flgR gene product , a flgR : : TnJOdCm insertion mutation was introduced by P22 transduction into fla mutants containing a MudJ fusion to either fliC orfljB .
Loss of function of the flgR gene product due to insertional inactivation resulted in relief of repression of flagellin expression in almost all of the fla mutant backgrounds ( Table 4 ) .
The three exceptions were the flhC , flhD , andfliA strains .
These genes are thought to encode positive regulators required for expression of the flagellin genes .
The flhC and flhD genes are master regulators required for expression of all the chemotactic and flagellar genes .
The fliA gene is only required for expression of the chemotactic genes and the late genes of flagellar assembly , including the hook-associated proteins ( flgK , flgL , fliD ) and the flagellin structural genes ( fliC andfljB ) ( 32 , 38 ) .
BL TA Frequency of switching Switching frequency per cell per generation ( 1o-3 ) Lac ' to Lac-Lac-to Lac ' 1.0 5.9 3.6 0.51 0.67 0.24 2.8 0.26 1.3 0.2 Strain MudJ in : TH1077 TH714 TH715 TH716 TH717 fliC fljB fljB fljB fljB ( 6 ) FIG. 4 .
Location of TnJOdTc insertions on the S. typhimurium chromosome which result in a Lac-phenotype on TTC-Lac indicator plates in a fljB : : MudJ fusion strain .
The number of insertions isolated , which map to a given location in the chromosome , is indicated in parentheses .
Expression offljB and fliC in flagellar mutant strains Mutant strain flj f fBjflliC flhB -- flhC -- flhD -- fliA ¬ fliC + fliD + fliE-fliF ¬ fliG-fliH ¬ ffiI-fliJ ¬ fliK ¬ Mutant strain flgA flgB flgC flgD flgE flgF flgG flgH flgi flgJ flgK flgL flhA i l fliBlB ¬ ¬ ¬ ¬ ¬ + ¬ f t n Muta strain flliC ffljjB ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ -- + + ¬ flC ¬ ¬ ¬ ¬ ¬ fliM fliN fliO fliP fliQ fliR AflgA-J AflgG-L AflhA-cheA Atar-flhD AfliA-D AfliE-K AfliJ-R ¬ ¬ ¬ ¬ ¬ ¬ ¬ ¬ + ¬ ¬ ¬ ¬ ¬ ¬ ¬ + ¬ ¬ ¬ ¬ + + ¬ ¬ ¬ ¬ ¬ Mstrant fliC fljB + + + + + + fljB + + + + + + + + + + + + + fliC + + + + + + + + + + + + + flgA flgB flhB flhC flhD fliA fliC fliD fliE fliF fliG fliH fliI fliJ fliK fliM fliN + + + + + + + ¬ ¬ ¬ + + + + + + + + + + ¬ ¬ flgC flgD flgE flgF flgG flgH figI flgJ flgK flgL flhA fliO fliP fliQ fliR AflgA-J AflgG-L AflhA-cheA Atar-flhD LAfliA-D AtfliE-K AfliJ-R ¬ + + + + + + + + + + + + ¬ ¬ + + + + DISCUSSION lac operon fusions to the flagellin structural genes ( fliC [ HI ] andfljB [ H2 ] ) have provided a simple phenotypic assay for in-vivo Hin-mediated flagellar phase variation in S. typhimurium .
These strains alternate between a Lac ' and Lac-phenotype readily observed on indicator media .
The frequency of switching was found to be about two-to fourfold higher than that determined in previous studies by using antibody agglutination assays ( 47 ) .
It may be that lac fusions provide a more sensitive assay in measuring phase variation in-vivo .
In addition , we found a bias in the switching rates as had been previously found .
By using the lac fusions , this bias was able to be accurately determined and ranged between 4-and 20-fold , depending on the individual fusion , in favor of switching from the fljB ( On ) to the fljB ( Off ) orientation .
We would like to propose a model to explain the observed switching bias .
The bias may be due to the presence of a pseudo-hix recombination site called the hix20 site ( 12 ) .
The hix2 ° site is located just upstream of the hin translation initiation codon ( Fig. 3 ) , was shown to bind Hin in DNase footprinting assays ( 12 ) , and may be a site for autoregulation by Hin .
It has been shown that Hin binds hix sites as a dimer ( 12 , 19 ) and that Hin is the limiting factor in the rate of phase variation in-vivo ( 4 ) .
It has also been shown that Hin dimers bound at the hix recombination sites will then form tetramers and bring the sites in close proximity ( 16 ) .
This results in looping of the DNA between the hix sites and is necessary for recombination to occur .
We propose that this looping of the DNA resulting from Hin tetramer formation plays a role in the switching bias .
Long-range protein-protein interactions have been shown to be involved in regulation of the ara , deo , gal , and lac operons of E. coli ( 9 , 11 , 14 , 15 , 20 , 48 ) .
In vivo binding of Hin to a defective hix site is enhanced by the presence of another functional hix site even at a distance of 1 kb ( 16a ) , suggesting a role of tetramer formation in binding .
Hin binds the hixL recombination site with greater affinity both in-vitro and in-vivo than it does the hixR site ( 12 , 19 ) .
In the fljB ( On ) orientation , the hix2 ° site is proximal to hixL ; in thefljB ( Off ) orientation , the hix20 site is proximal to hixR .
We propose that the hix2 ° site competes with hixL and hixR for binding limiting concentrations of Hin dimers during tetramer formation .
Since hixR is a poorer binding site than hixL , the hix2 ° site , it can compete more effectively with hixR in the fljB ( Off ) orientation than it can with hixL in the fljB ( On ) orientation .
The result is that cells spend more time in the fljB ( Off ) orientation than in the fljB ( On ) orientation , creating the switching bias observed in-vivo .
The isolation of MudJ insertion mutants in the flagellin structural genes provided not only a simple method of screening for mutants unable to transcribe the fliC or fljB genes , but also a means to distinguish these mutants from mutants defective in hin and unable to carry out the site specific DNA inversion reaction .
Both classes of mutants were isolated .
Most of the insertion mutants mapped to genes known to be required forfliC andfljBl expression , but a small number of novel insertion mutants , mapping to the 3 - , 15 - , 33 - , and 50-mmn regions of the chromosome , were + It was unexpected that new genes ( classes I and II ) involved in negative regulation of the flagellin structural genes would be identified at the onset of these studies .
There are at least two possible models ( Fig. 5 ) which could account for the action of the flgR gene product .
One model suggests that FIgR is activated by a signal produced from a defective flagellar structure due to a mutation affecting a step in the assembly pathway .
The activated FlgR protein would then prevent flagellin expression .
This could occur through direct transcriptional repression of the fliA gene or by the inactivation of the fliA gene product .
A second model is that FIgR is assembled into the final flagellar structure .
If any of the steps in fiagellar assembly were defective ( except the late steps involving flgK , flgL , fliC , fliD , and fllB ) , FlgR would no longer be associated with the flagellar structure and would then act either directly or indirectly , presumably throughfliA repression or inactivation , to prevent flagellin expression .
The study of the flgR gene may lead to an understanding of the regulatory circuit which senses the inability to synthesize complete flagellar structures .
Involvement of the figA gene in repression of the late genes has been reported for E. coli ( 27 ) .
An insertion of transposon TnS into the figA gene was found to relieve repression of late-gene synthesis in flagellar mutant backgrounds in a manner similar to the results reported here for flgR .
We find that expression of flagellin is repressed in a figA mutant , and a flgR : : TnJOdCm insertion mutant is not defective in flagellar formation .
These results suggest that , in S. typhimurium , flgA and flgR are separate genes but do not rule out the possibility that the FIgA and FlgR phenotypes result from mutations in different domains of one protein .
It may be that the figA : : TnS insertion mutant of E. coli , in which repression of late-gene expression is relieved in fla mutant backgrounds , is simply polar on the E. coli flgR locus .
These differences should be resolved by further characterization of the flgR locus .
( late genes ) ( early genes ) ( middle genes ) 10 -7 -0 p flhC , flhD tP f e , P , M .
B ( early genes ) ( middle genes ) ( late genes ) P f ` la 01-2P fliA p flhC , flhD P fla FlgR 0 defective flagella signal FlgR * .
C ( early genes ) ( middle genes ) ( late genes ) p flhC.flhD .2 .
P flIiA P fla We would like to thank Shigeru Yamaguchi , Nick Benson , John Roth , and Bill Kay for their generous gifts of the numerous strains used in the course of this work and Colin Manoil and Jon Visick for critically reviewing the manuscript .
This work was supported by Public Health Service grant GM43149 from the National Institutes of Health .
K.T.H. is a recipient of a Junior Faculty Research Award from the American Cancer Society .
Adler , J. , and B. J. Templeton .
The effect of environmental conditions on the of Escherichia J. motility coli .
la.Benson , N. , and B. Goldman .
Bochner , B. R. , H. Huang , G. L. Schieven , and B. N. Ames .
Positive selection for loss of tetracycline resistance .
Bochner , B. R. , and M. A. Savageau .
Generalized indicator plate for genetic , metabolic and taxonomic studies with microorganisms .
Bruist , M. F. , and M. I. Simon .
Phase variation and the Hin protein : in-vivo activity measurements , protein overproduction , and purification .
Casadaban , M. J. , and S. N. Cohen .
Lactose genes fused to exogenous promoters in one step using Mu-lac bacteriophage : in-vivo probe for transcriptional control sequences .
Castilho , B. A. , P. Olfson , and M. J. Casadaban .
Plasmid insertion mutagenesis and lac gene fusion with mini-Mu bacteriophage transposons .
Chan , R. K. , D. Botstein , T. Watanabe , and Y. Ogata .
Specialized transduction of tetracycline resistance by phage P22 in Salmonella typhimurium .
Properties of a high frequency transducing lysate .
Davis , R. W. , D. Botstein , and J. R. Roth .
Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. 9 .
Eisman , E. , B. von Wilcken-Bergmann , and B. Muiller-Hill .
Specific destruction of the second lac operator decreases expression of the lac operon in Escherichia coli fivefold .
Elliott , T. , and J. R. Roth .
Characterization of TnlOd-Cm : a transposition-defective TnJO specifying chloramphenicol re 11 .
Flashner , Y. , and J. D. Gralla .
Dual mechanism of repression at a distance in the lac operon .
Glasgow , A. C. , M. F. Bruist , and M. I. Simon .
DNA-binding properties of the Hin recombinase .
Methods Enzymol. , in press .
Haber , R. , and S. Adhya .
Interaction of spatially separated protein-DNA complexes for control of gene expression : operator conversions .
Hamilton , E. P. , and N. Lee .
Three binding sites for AraC protein are required for autoregulation of araC in Escherichia coli .
Hleichman , K. A. , and R. C. Johnson .
The Hin invertasome : protein-mediated joining of distant recombination sites at the enhancer .
Hughes , K. T. Unpublished results .
Hughes , K. T. , and J. R. Roth .
Conditionally transposi-tion-defective derivative of Mu dl ( Ap Lac ) .
Hughes , K. T. , and J. R. Roth .
Transitory cis complementation : a method for providing transposition functions to defective transposons .
Hughes , K. T. , P. Youderian , and M. I. Simon .
Phase variation in Salmonella : analysis of Hin recombinase and hix recombination site interaction in-vivo .
Huo , L. , K. J. Martin , and R. Schleif .
Alternative DNA loops regulate the arabinose operon in Escherichia coli .
Genetics ' and chemistry of bacterial flagella .
Genetics of structure and function of bacterial flagella .
Eno , T. , Y. Komeda , K. Kutsukake , R. M. Macnab , P. Matsu-mura , J. S. Parkinson , M. I. Simon , and S. Yamaguch .
New unified nomenclature for the flagellar genes of Escherichia coli and Salmonella typhimurium .
Johnson , R. C. , and M. I. Simon .
Hin-mediated sitespecific recombination requires two 26 bp recombination sites and a 60 bp recombinational enhancer .
Jones , C. J. , and R. M. MacNab .
Flagellar assembly in Salmonella typhimurium : analysis with temperature-sensitive mutants .
Fusions of flagellar operons to lactose genes on a Mu lac bacteriophage .
Transcriptional control of flagellar genes in Escherichia coli K-12 .
Komeda , Y. , and T. Ilno .
Regulation of expression of the flagellin gene ( hag ) in Escherichia coli K-12 : analysis of hag-lac gene fusions .
Komeda , Y. , H. Suzuki , J. Ishidsu , and T. Iino .
The role of cAMP in flagellation of Salmonella typhimurium .
Kutsukake , K. , and T. IDno .
Refined genetic analysis of the region II che mutants in Salmonella typhimurium .
Kutsukake , K. , T. Dino , Y. Komeda , and S. Yamaguchi .
Functional homology offla genes between Salmonella typhimu-rium and Escherichia coli .
Kutsukake , K. , Y. Ohya , and T. Dino .
Transcriptional analysis of the flagellar regulon of Salmonella typhimurium .
Kutsukake , K. , Y. Ohya , S. Yamaguchi , and T. Dino .
Operon structure of flagellar genes in Salmonella typhimurium .
Detection of fermentation variants with tetrazolium .
In F. C. Neidhardt , J. L. Ingraham , K. B. Low , B. Magasanik , M. Schaechter , and H. E. Umbarger ( ed .
) , Escherichia coli and Salmonella typhi ¬ murium : cellular and molecular biology .
American Society for Microbiology , Washington , D.C. 36 .
Maloy , S. R. , and W. D. Nunn .
Selection for loss of tetracycline resistance by Escherichia coli .
Maloy , S. R. , and J. R. Roth .
Regulation of proline utilization in Salmonella typhimurium : characterization of put : : Mu d ( Ap , lac ) operon fusions .
Ohnishi , K. , K. Kutsukake , H. Suzuki , and T. Dino .
Gene piA encodes an alternate sigma factor specific for flagellar operons in Salmonella typhimurium .
Sanderson , K. E. , and J. R. Roth .
Linkage map of Salmonella typhimurium , edition VI .
Sanderson , K. E. , and J. R. Roth .
Linkage map of Salmonella typhimurium , edition VII .
Silverman , M. , and M. Simon .
Genetic analysis of flagellar mutants in Eschernchia coli .
Silverman , M. , and M. Simon .
Characterization of Esch-erichia coli flagellar mutants that are insensitive to catabolite-repression .
Silverman , M. , and M. Simon .
Phase variation : genetic analysis of switching mutants .
Silverman , M. , J. Zieg , and M. I. Simon .
Flagellar phase variation : isolation of the rHI gene .
Defective phage formation by lysogens of integration deficient phage P22 mutants .
Somers , J. M. , and W. W. Kay .
Genetic fine structure of the tricarboxylate transport ( tct ) locus of Salmonella typhimu-rium .
Stocker , B. A. D. 1949 .
Measurements of rate of mutation of flagellar antigenic phase in Salmonella typhimurium .
Valentin-Hansen , P. , B. Albrechtsen , and J. E. L0ve Larsen .
DNA-protein recognition : demonstration of three genetically separated operator elements that are required for repression of the Escherichia coli deoCABD promoters by the DeoR repressor .
Vogel , H. J. , and D. M. Bonner .
Acetylornithase of Escherichia coli : partial purification and some properties .
Way , J. C. , M. A. Davis , D. Morisato , D. E. Roberts , and N. Kleckner .
New TnlO derivatives for transposon mutagenesis and for construction of lacZ operon fusions by transposition .
Winkelman , J. W. , and D. P. Clark .
Proton suicide : general method for direct selection of sugar transport-and fermentation-defective mutants .
Yamaguchi , S. , H. Fujita , A. Ishihara , S.-I .
Aizawa , and R. M. MacNab .
Subdivisions of flagellar genes of Salmonella typhimurium into regions responsible for assembly , rotation , and switching .
Yamaguchi , S. , H. Fujita , T. Taira , K. Kutsukake , M. Homma , and T. Iino .
Genetic analysis of three additionalfla genes in Salmonella typhtmurium .
Yokota , T. , and J. S. Gots .
Requirements of adenosone 3 ' ,5 ' - cyclic phosphate for flagella formation in Escherichia coli and Salmonella typhimurium .
Youderian , P. , P. Sugino , K. L. Brewer , N. P. Higgins , and T. Elliott .
Packaging specific segments of the Salmonella chromosome with locked-in Mud-P22 prophages .
Zieg , J. , M. Hilmen , and M. Simon .
Regulation of gene expression by site-specific inversion .
Zieg , J. , M. Silverman , M. Hilmen , and M. Simon .
Recombinational switch for gene expression .
Zieg , J. , and M. Simon .
Analysis of the nucleotide sequence of an invertible segment of-an invertible controlling element .