The interaction of L-arabinose and D-fucose with AraC protein.

Abstract The interaction of the inducer l-arabinose and an anti-inducer d-fucose with araC protein is studied using spectrofluorimetric methods. AraC protein has a fluorescence spectrum typical of tryptophan-containing proteins. The fluorescence is quenched by l-arabinose suggesting that l-arabinose causes at least one conformational change in araC protein. The quenching of fluorescence does not occur in the presence of d-fucose. These observations are consistent with the model for positive control of the l-arabinose operon BAD by araC protein and suggest that the conformational change induced by l-arabinose reflects the transition to araC activator. The interaction of d-fucose with araC protein prevents the conformational change which yields activator, thus, accounting for the anti-inducer properties of this analogue of l-arabinose.

L-arabinose isomerase (uraA), and Lribulose-5-phosphate-4-epimerase (araD) and two controlling sites, the initiator (araZ) and the operator (a~u0) (1). Genetic and physiological studies by Englesberg and co-workerg (2, 3) suggest a model in which the expression of the L-arabinose operon is both positively and negatively controlled by the protein product of the araC gene. According to the model the uraC protein exists in two active conformational states, PI (the repressor) and P2 (the activator).
L-Arabinose induces t'he operon by removing P1 from ara0 and shifting the equilibrium to I'2 which acts at uraZ to stimulate expression of -the operon. In viuo studies measuring ara mRNA by hybridization to ara DNA demonstrate that ara mRNA is both activated and repressed by Pz and P1, respectively (4-6).
P1 binds specifically to Xh8Odara DNAs which carry ara0 providing strong evidence that repressor functions at the level of transcription (7). Proof that Pz activates transcription of araBAD mRNA has recently been obtained using a purified transcription system (8). The ability of araC protein to act as either an activator or repressor of gene expression is unique among the regulatory proteins that have been studied. * This work was supported by National Science Foundation Grant GB 24093 to E. Englesberg.
$ Present address, Department of Bacteriology, University of California, Los Angeles, California 90024.
The purification of uraC protein (7, 9) makes it possible to study its proposed interaction with L-arabinose. We find that purified arc& protein has an intrinsic fluorescence typical of tryptophan-containing proteins.
The fluorescence is partially quenched by L-arabinose suggesting that L-arabinose causes at least one conformational change in araC protein.
This conformational change is prevented by n-fucose, an analogue of Larabinose and an anti-inducer of the L-arabinose operon. Binding constants obtained from the spectrofluorometric studies are in good agreement with those obtained by other methods.

Bacterial
and Phage Strains-Bacterial strains used in this study containing lesions in the aru region of the chromosome were originally isolated in Escherichia coli B/r in the laboratory of Ellis Englesberg.
SB7515, the source of araC protein for these experiments, was constructed by P. Cleary and is an E. coli B/r strain which is singly lysogenic for a heat-inducible, lysis-inhibited defective ara transducing phage, Xh80dara+ cI&F,~s (7). SB5500, a strain unable to acti;eiy transport L-arabinose, was isolated by the method of Issacson and Englesberg on inhibitory media (mineral L-arabinose glycerol media containing reduced -_ amounts of K+ ions) (10). A series of L-broth cultures is inoculated with approximat,ely 200 cells from an overnight L-broth culture of an F-araAdthr+leu-strain and incubated for 24 hours at 37". A loopful from each culture is streaked once across a plate of inhibitory media and the plates are incubated for 48 hours at 37'. Five to ten resistant mutants are picked from each streak and purified by restreaking on homologous media. A mutant (SB5500) wa+~ found which on the basis of complementation tests was araD+A-B+C+ L-arabinose permease activity as measured by the procedure of Novotny and Englesberg (11) is completely absent in SB55OO.
Purified lac repressor was generously provided by Arthur Riggs, City of Hope National Medical Center, Duarte, California.
The cultures are incubated at 37" with shaking for 15 hours and then chilled in an ice bath. The bacteria are harvested by centrifugation, resuspended in 1.5 ml of a solution containing immole ofEDTA and immole of glutathione, pH 7.6, and disruuted with the microtiD of a Branson Sonifier. Cell debris is iemoved by one 15-rnii centrifugation at 27,000 X g. L-Ribulokinase is assayed as described by Englesberg et al. (2). Fluorescence Methodology-The spectrofluorimeter used is an uncorrected Hitachi MPFSA equipped with a cell holder thermostatically controlled at 25". Spectra are measured at an emission slit width of 6 nm.
C, C*, and CT are native, quenched, and total UT& protein, respectively; A is L-arabinose, and F is o-fucose.
The experiment is done at saturating concentrations of L-arabinose so that CT = C*A + CF. The above equation is derived from the equilibrium expression C + A s C*:A and C + Fe CF.

RESULTS
L-Arabinose Induces Conformational Change in araC Protein-The fluorescence emission spectrum of araC protein has a maximum at 338 nm (Fig. 1). This is characteristic of proteins which contain L-tryptophan residues (13). The addition of Larabinose causes a 45% decrease in the fluorescence emission of araC protein with no apparent than .ge in the shape of the spectrum.
The fluorescence emission spectrum of L-tryptophan or lac repressor in Huffer A is not quenched by the addition of Larabinose.
Furthermore, the quenching of araC protein fluorescence by L-arabinose is very specific since u-arabinose, u-galactose, and u-glucose do not quench and D-fucose only slightly quenches (2 to 6%) the fluorescence.
The quenching of araC protein fluorescence could occur as the result of a conformational change or as the result of L-arabinose being bound on or very close to a tryptophanyl residue(s) of aruC protein.
D-Fucose and L-arabinose bind to the same site on araC protein since u-fucose is a competitive inhibitor of Larabinose binding both in tivo (16,19) and in vitro (17). The structural similarities of the two compounds are also consistent with binding to the same site. Thus, the inability of n-fucose to elicit a significant amount of quenching of fluorescence is strong evidence against a local of a conformational change.
effect and supports the alternative 6893 allowing a calculation of the apparent K, of the interaction. The data shown in Fig. 2 are plotted as suggested by Scatchard (14) where fi is the fractional quenching of Auorescence and 2; is the free L-arabinose concentration.
The value for the maximum quenching is obtained from a plot of l/L versus l/quenching. The concentration of free L-arabinose can be assumed to be equal to the total concentration of L-arabinose since the concentration of aruC protein monomers in the cuvette is approximately 4 X 10m7 M and that of L-arabinose at least lo'-" M where binding begins to take place. If one assumes that the amount of binding of L-arabinose to aruC protein is directly proportional to the quenching observed, then the data shown in Fig. 3 suggest that L-arabinose binds to independent sites with an apparent K, of 3 x 10-" M.
Fucose Inhibits Interaction with J;-Arubinose-As mentioned above D-fucose has only a slight effect on the fluorescence emission spectrum of uruC protein.
The greatest decrease in Auorescence observed upon addition of n-fucose with three different preparations of uruC protein was 6%. This is in great contrast to the 45?$ quenching observed in the presence of saturating amounts of L-arabinose.
If one adds D-fucose before the addition of L-arabinose, the quenching of fluorescence normally seen upon the addition of L-arabinose is not observed.
In addition, if uraC protein fluorescence has been quenched by the addition of L-arabinose, the quenching can be reversed by the addition of u-fucose. The data presented in Fig. 3 are obtained in the following way: varying amounts of D-fucose are added to uruC protein in Buffer A containing 0.1 M L-arabinose causing an increase in fluorescence intensity proportional to the amount of n-fucose added. The data are plotted by the same method used in the experiment shown in Fig. 2. fi is the fractional increase in fluorescence and L is the free n-fucose concentration which may be equated to the total concentration because again l/K is much greater than the ligand concentration.
If we assume that the quenching observed reflects the binding of L-arabinose, then the D-fucose must bind to independent sites with an apparent Ki of 6 X lo-" M.
Other Estimates of Binding Constants-The strength of binding of L-arabinose and D-fucose to aruC protein has been estimated by indirect methods both in tivo (15, 16) and in vitro (17, 18) and the values obtained are on the order of lo-* to 10m2 M. Assuming that induction is directly proportional to the binding of Larabinose to araC protein, an apparent K, of 5 X 10m3 M is calculated (Fig, 4). This is in excellent agreement with the value of 3 x lW3 M obtained above by the spectrofluorimetric method.

DISCUSSION
The interconversion of activator and repressor mediated by Larabinose is the key feature of the model for regulation of the L-arabinose operon proposed by Englesberg et al. (2,3). L-Arabinose is considered to be both the inducer and effector because L-ribulokinase is inducible in amA-mutants and no evidence has been found for any alteration in the accumulated L-arabinose in these strains (1). Further support for a direct interaction between inducer and aruC protein has been obtained from studies using D-fucose, a non-metabolixable analogue of L-arabinose which prevents growth of Eschervichia coli B/r on a mineral salts L-arabinose medium by inhibiting induction of the operon (2). Mutations conferring D-fucose resistance map in uraC (the uruCc allele) and result in constitutive expression of the L-arabinose operon (2). Many arac" alleles also permit Dfucose to serve as a gratuitous inducer of the araBAD operon and suggest that the fucose-resistant UTUC~ allele produces a product with modified inducer specificity (19). Similar aruCc alleles have also been found in Escherichiu coli K12 (18). These mutants provide strong evidence for a direct interaction between the inducer and uruC protein in wild-type cells. In tivo studies with merodiploids suggest that while D-fucose does not significantly alter repressor function it may block the L-arabinosemediated conversion of repressor to activator (19). In vitro studies with cell-free systems (17, 18) further support this proposed interaction of D-fucose with uruC protein and rule out the possibility that the only effect is at the level of transport.
The results presented above demonstrate that L-arabinose interacts directly with uruC protein and may induce a conformational change. Furthermore, the anti-inducer D-fucose can prevent the conformational change induced by L-arabinose. These observations are consistent with the model for control of the L-arabinose operon by uruC protein and suggest that the observed conformational change reflects the transition from uruC repressor (PI) to aruC activator (PJ. The interaction of D-fucose with uruC protein does not result in the conformational change which yields uruC activator, thus accounting for the anti-inducer properties of this analogue of L-arabinose. Finally, the interaction of L-arabinose with uruC protein occurs only at high L-arabinose concentrations (> 10m3 M).