tAutoregulation : A Role for a Biosynthetic Enzyme in the Control of Gene Expression ABSTRACT It was previously proposed , primarily on the basis of evidence in-vitro , that L-threonine deaminase , the ilvA gene product , is required for repression of its own synthesis and for repression of the other genes in the ilv-ADE operon .
In this communication , evidence in-vivo is presented that supports this autoregulatory model .
Further evidence is presented that suggests that L-threonine de-aminase is also required for induction of the ilvC gene product .
The autoregulatory model is presented in an expanded form to include recent evidence that L-threonine deaminase ( EC 4.2.1.16 ) is a central element for repression of the ilvADE and ilvB operons , and for induction of the ilOX operon .
The operator-relpressor model proposed by Jacob and Monod ( 1 ) in 1961 provided molecular biologists an insight into the possible mechanisms of regulation of gene expression .
After the introduction of this model , a large number of experimental facts have accumulated that have proven the validity of the model , particularly with respect to the lactose utilization system for which it was originally formulated ( 2 ) .
Because of the simplicity and aesthetic appeal of the JacobMonod model , great energy was expended after its introduction to apply it to other catabolic and biosynthetic regulatory systems .
However , the model , in its original form , has proven inadequate for the explanation of many of the facts surrounding repression and derepression of enzymes involved in biosynthetic pathways ( 3 ) .
Several authors have proposed alternative models of regulation for these systems ( 4-10 ) .
One such alternative is the autoregulatory model proposed by Hatfield and Burns ( 11 ) in 1970 , which was formulated to explain the repression pattern observed for enzymes involved in biosynthesis of the branched-chain amino-acids , visoleu-cine , r-valine , and L-leucine ( Fig. 1 ) .
Repression of these enzvmes is multivalent ; i.e. , for repression of enzyme synthesis to occur , all three branched-chain amino-acids must be present in excess ( 12 ) .
The model ( Fig. 2 ) proposed that , depending upon the chemical environment in the cell,-threonine de-aminase ( EC 4.2.1.16 ) , the ilvA gene product , can function as either a catalytically active enzyme or as a catalytically inactive , autoregulatory feedback repressor of the ilvADE operon .
The sequence of events proposed in the autoregulatory model corresponds to the molecular transitions demonstrated in-vitro ( Fig. 2 , steps 2-6 ) with purified enzyme preparations ( 13 , 14 ) .
It was demonstrated that addition of the L-threonine deaminase cofactor , pyridoxal phosphate , to purified prepara ¬ MATERIALS AND METHODS scribed by Guirard et al. ( 15 ) , except that MgCl2 was added to a final concentration of I mM for stabilization of dihydroxyacid dehydratase ( ilvD ) and the isomeroreductase ( ilv-C ) activities .
Since this buffer contains pyridoxal phosphate , it is possible to estimate the total amount of L-threonine de-aminase protein present in the cell ; i.e. , apoenzyme plus holoenzyme .
When pyridoxal phosphate is omitted from the buffer , L-threonine deaminase activity decreases to less than 1 % of the activity measurable when pyridoxal phosphate is present .
Upon addition of pyridoxal phosphate to a growing culture of mutant ilvA504 , the measurable activity ( with the buffer containing pyridoxal phosphate ) immediately increases about 2-fold .
This increase presumably reflects the more efficient formation of holoenzyme in-vivo , and is consistent with the observation that only 50-60 % of enzyme activity is recovered in-vitro by addition of pyridoxal phosphate to purified apoenzyme ( 14 ) .
The isomeroreductase was assayed as described ( 17 ) in sonicated cell-free extracts prepared in the same buffer used to assay L-threonine deaminase and dihydroxyacid dehydratase .
Enzyme activity is expressed as units per ml of culture .
One enzyme unit is the amount of enzyme required to form 1 Mumol of product per min .
Control experiments were performed , which demonstrated that addition of pyridoxal phosphate to cultures of wild-type S. typhimurium LT-2 did not alter the rates of enzyme synthesis .
RESULTS L-Threonine Deaminase , the ilvA Gene Product .
The autoregulatory model ( 11 ) predicts that cells unable to form holo-L-threonine deaminase would be unable to repress synthesis of the enzymes coded by the genes in the ilvADE operon .
This prediction was tested in S. typhimnurium strain ilvA504 .
This strain produces an L-threonine deaminase with a decreased affinity for its cofactor , pyridoxal phosphate .
The mutant requires either pyridoxal phosphate or L-iSo-leucine for growth .
Although normal quantities of pyridoxal phosphate are synthesized , it is not sufficient for in-vivo formation of catalytically active holo-ithreonine deaminase ( 15 ) .
Hence , the cells can not synthesize enough L-isoleucine to support normal growth .
Addition of exogenous pyridoxal phosphate to cultures of mutant ilvA504 leads to the in-vivo formation of holoenzyme , and consequently , L-isoleucine is synthesized in quantities sufficient for normal growth ( 15 ) .
In the experiments shown in Fig. 3A , samples were taken for L-threonine deaminase assay from parallel cultures of mutant ilvA504 growing with repressing concentrations of ileucine , L-valine , and L-isoleucine .
In these experiments the cell samples were centrifuged , suspended in buffer containing pyridoxal phosphate , treated with toluene , and then assayed for L-threonine deaminase activity .
This procedure converts the catalytically inactive apoenzyme to catalytically active holoenzyme and , therefore , total L-threonine deaminase protein in the cell is measured .
Addition of pyridoxal phosphate to one of these cultures at the time indicated by the arrow ( Fig. 3A ) causes an immediate repression of the differential rate of synthesis of-threonine deaminase .
The the ilvD Gene Dihydroxyacid Dehydratase , samples taken from the experiment described above were also of the the ilv-assayed for dihydroxyacid dehydratase , product D gene .
Dihydroxyacid dehydratase , like L-threonine de-aminase , is synthesized at a markedly slower ( repressed ) rate after addition of pyridoxal phosphate to the culture ( Fig. 3B ) .
2-Acetohydroxy-3-Ketoacid Isomeroreductase , the ilvC Gene Product .
The product of the ilvC operon , the isomeroreduct-ase , was initially thought to be subject to the same multivalent repression signal as the ilvADE operon ( 12 ) .
However , Arfin et al. ( 18 ) demonstrated that this enzyme is , in fact , induced by either of its substrates , a-acetolactate or a-aceto-hydroxybutyrate ( Fig. 1 ) .
The properties of a pleiotropic regulatory mutation ( see Discussion ) suggested a requiremen for L-threonine deaminase in induction of the ilvC gene product ( 16 , 19 ) .
Accordingly , we tested the role of holo-L-threonine deaminase in induction of the ilvC gene product in strain ilvA504 .
Parallel cultures were grown in medium containing repressing concentrations of-leucine , L-isoleucine , and ivaline , and inducer a-acetohydroxybutyrate .
Pyridoxal phosphate was added to one of the cultures at the time indicated by the arrow , and samples were taken during-growth of the cultures for assay of the iivC enzyme ( Fig. 3C ) .
Although enzyme synthesis occurs in the absence of added pyr-idoxal phosphate , the enzyme is synthesized at a significantly greater differential rate in the presence of the L-threonine de-aminase cofactor .
It is unlikely that the enhanced induction seen in the presence of pyridoxal phosphate reflects an increase in inducer concentration as a consequence of L-threonine deaminase activity , since addition of pyridoxal phosphate to cultures of mutant ilvA504 growing in repressing concentrations of L-leucine , iisoleucine , and L-valine in the absence of exogenous inducer does not result in an induction of the ilvC enzyme .
That some induction of the ilvC enzyme does occur in strain ilvA304 in the absence of exogenous pyridoxal phosphate is not an unexpected result , since this strain requires only L-isoleucine , and not valine or leucine , for growth .
Thus , this mutant is obviously able to synthesize quantities of the ilvC enzyme sufficient for formation of the metabolic precursors of valine and leucine .
Moreover , the mutant can also utilize a-ketobutyrate as a growth factor , indicating an adequate conversion of a-acetohydroxybutyrate to a,0-dihydroxy - , B-methyl-valerate for L-isoleucine biosynthesis ( Fig. 1 ) .
It is possible that extremely low concentrations of holoenzyme , such as may be present in the mutant , permit some induction .
DISCUSSION Whereas the Jacob-Monod model ( 1 ) was based upon the properties of certain classes of regulatory mutants , the autoregulatory model ( 11 ) was based almost entirely upon the in-vitro properties of a purified protein ( L-threonine de-aminase ) ( 13 , 14 ) .
Consequently , although at the time of its introduction it was entirely consistent with the limited genetic and biochemical data available , it lacked direct in-vivo verification .
Nevertheless , because of the explicit predictions of the model it is amenable to rigorous testing at both the genetic and biochemical levels .
The experiments reported here were designed to test the postulated role for holo-L-threonine deaminase in multivalent repression of the ilvADE operon .
When mutant ilvA-304 is grown under repressing conditions addition of pyridoxal phosphate is gratuitous for growth but obligatory for repression .
These experiments provide in-vivo evidence in support of the autoregulatory model .
They further suggest that , as originally sl ) ecified , holoenzyme , but not apoenzyme , l ) articipates in the mrultivalent repression mechanism .
Although in its original form the autoregulatory model l ) roposed a role for-threonine deaminase in regulation of only the ilvADE operon , there is now evidence that extends the role of this enzyme to regulation of expression of the other operons of the isoleucine-valine biosynthetic system , ilvW and ilvB .
The first evidence that implicated L-threonine deaminase in induction of the ilvC operon was based upon the properties of a lpleiotropic regulatory mutant of Escherichia coli K12 , which was isolated and described by Pledger and Umbarger ( 19 ) .
The single mutation in this strain affected the properties of-threonine deaminase in crude extracts , the inducibility of the ilvC operon , and the repressibility of the ilvADE operon .
Although the site of this mutation , ilvY , is separated from the structural gene for ithreonine deaminase by an operator gene ( ilvO ) ( Fig. 4 ) , Calhoun et al. ( 16 ) demonstrated that it nevertheless affected the structure and properties of the purified mutant enzyme .
The experiments reported here in S. typhimurium suggest that L-threonine deaminase is involved in induction of the ilvC operon and they further implicate a holo-form of the enzyme in this induction lprocess .
Several lines of evidence suggest that , for both E. coli and S. typhimurium , ithreonine deaminase is a central control element in regulation of expression of the ilvADE , ilvB , and ilvC operons .
There is indirect evidence that multivalent repression in both S. typhimurium and E. coli involves amino-acylated tRNA molecules for all three branched-chain amino-acids ( 17 , 21-25 ) .
Also , recent results indicate that leucyl - , isoleucyl - , and valyl-tRNA bind to immature L-threonine deaminase from E. coli .
Data obtained from the Amicon ultrafiltration membrane assay indicates binding of leucyl - , iso-leucyl - , and valvl-tRNA to immature L-threonine deaminase from E. coli K-12 ( Calhoun , D. H. , unpublished observations ) .
It is , therefore , al ) l ) rol ) riate at this time to expand the original autoregulatory model to include these recent observations ( Fig. 4 ) .
The central feature of the autoregulatory model l ) rol ) osed in 1970 related to the participation of L-threonine deaminiase in the control of its own synthesis and the synthesis of the other structural genes in the ilv.ADE oleron .
The conceptual validity of this autoregulatory scheme with respect to the ilv-ADE operon is suggested by : ( a ) The pleiotropic mutation in strain CU18 ( 19 ) , which results in synthesis of an altered L-threonine deaminase protein and altered regulation of the ilvADE operon ; ( b ) the experiments reported here , which implicate holo-L-threoiiine deaminase in repression of the ilv-ADE operon ; and ( c ) the experiments of Wasmuth et al. ( 20 ) , which also suggest that holo-L-threonine deaminase parti-cipates in the multivalent repression of the ilvADE operon .
The conclusion that L-threonine deaminase is involved in repression of the ilvB operon and substrate induction of the ilv-C operon is based upon the experiments reported here , the data of Wasmuth et al. ( 20 ) , and the additional l ) leiotrol ) ic effect of the mutation carried by strain CU18 ( 19 ) .
It has not been possible to obtain direct in-vivo evidence that distinguishes between immature and mature enzyme involvement in regulation of gene expression in this pathway .
However , involvement of the immature enzyme in relpression of the ilvADE and ilvB operons is suggested by the finding that the immature enzyme preferentially binds aminoacylated tRNA in-vitro ( 11 ) .
In addition , results from in-vivo exl ) eriments of Wasmuth et al. ( 20 ) were interpreted in favor of role in repression for immature rather than mature enzyme .
That mature enzyme participates in induction of the ilvC is suggested by the following observations : ( a ) Was-operon muth et al. ( 20 ) , using the pyridoxal phosphate analogue , 4-deoxypyridoxine , found that the holoenzyme participating in repression is different from that involved in induction of the gene ; ( b ) the inducers of the enzyme do , in fact , bind ilvC ilvC to mature L-threonine deaminase ( Kuska , J. & Hatfield , G. W. , unpublished observations ) ; and ( c ) Arfin et al. ( 18 ) found that induction of the ilvC enzyme initially proceeds 2-to 3-fold faster in valine-starved cells ( conditions expected to promote conversion of immature to mature L-threonine deaminase ) than in repressing medium ( conditions expected to impede this maturation ) ( Fig. 4 ) ( 14 ) .
Arrows in Fig. 4 are used only to indicate the location of the structural genes controlled by L-threonine deaminase and are not necessarily intended to imply binding to DNA by the protein itself .
Experiments in progress , and results from in-vitro protein-synthesizing systems ( 27 ) should make it possible to determine whether L-threonine deaminase exerts its control at the level of transcrip-on gene expression tion or translation .
We thank Dr. H. E. Umbarger for many helpful discussions during this work and for providing the a-acetohydroxybutyrate and a-acetolactate required for the enzyme assays .
We also thank Twyla A. Miner for her expert technical assistance and suggestions .
Data reported in this communication were obtained from studies supported by research grants from the National Science Foundation ( GB-27519 ) and the American Cancer Society ( no. 524 ) .
D.H.C. is the recipient of U.S. Public Health Service Postdoctoral Fellowship AM-36517 .
G.W.H. is the recipient of U.S. a Public Health Service Research Career Development Award ( GM-70530 ) .
Jacob , F. & Monod , J. ( 1961 ) J. Mol .
Beckwith , J. R. & Zipser , D. ( 1970 ) The Lactose Operon ( Cold Spring Harbor Laboratory , Cold Spring Harbor , N.Y. ) .
Calvo , J. M. & Fink , G. R. ( 1971 ) Annu .
Cline , A. L. & Bock , R. M. ( 1966 ) Cold Spring Harbor Symp .
Gruber , M. & Campagne , R. N. ( 1965 ) Kon .
Hennig , U. , Dietrich , J. , Murray , K. N. & Deppe , G. ( 1967 ) Wiss .
Kovach , J. S. , Phang , J. M. , Ference , M. & Goldberger , R. F. ( 1969 ) Proc .
Kovach , J. S. , Phang , J. M. , Blasi , F. , Barton , R. W. , Ballesteros-Olmo , A. & Goldberger , R. F. J. Bacteriol .
Koshland , D. E. & Kirtley , M. E. ( 1966 ) in Major Problems in Developmental Biology , ed .
Locke , M. ( Academic Press , Inc. , New York ) , pp. 217-249 .
Bollon , A. P. & Magee , P. T. ( 1973 ) J. Bacteriol .
Hatfield , G. W. & Burns , R. O. ( 1970 ) Proc .
Freundlich , M. , Burns , R. 0 .
& Umbarger , H. E. ( 1962 ) Proc .
Hatfield , G. W. & Burns , R. 0 .
Hatfield , G. W. & Bums , R. 0 .
Guirard , B. M. , Ames , B. N. & Snell , E. E. ( 1971 ) J. Bacteriol .
Calhoun , D. H. , Rimerman , R. A. & Hatfield , G. W. ( 1973 ) J. Biol .
Szentirmai , A. , Szentirmai , M. & Umbarger , H. E. ( 1968 ) J. Bacteriol .
Arfin , S. M. , Ratzkin , B. & Umbarger , H. E. ( 1969 ) Bio-chem .
Pledger , W. J. & Umbarger , H. E. ( 1973 ) J. Bacteriol .
Wasmuth , J. , Umbarger , H. E. & Dempsey , W. B. ( 1973 ) Biochem .
Blatt , J. M. & Umbarger , H. E. ( 1972 ) Biochem .
Low , B. , Gates , F. , Goldstein , T. & Soll , D. ( 1971 ) J. Bacteriol .
Alexander , R. R. , Calvo , J. M. & Freundlich , M. ( 1971 ) J. Bacteriol .
Neidhardt , F. C. ( 1966 ) Bacteriol .
Iaccarino , M. & Berg , P. ( 1971 ) J. Bacteriol .
O'Neill , J. P. & Freundlich ( 1972 ) Biochem .
Haar , R. A. & Umbarger , H. E. ( 1973 ) Abstracts Annu .