2), which are involved in the repair of DNA damage. When cells are treated by a ... activity and normal in vitro thermostability but are completely deficient in auto-.
THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1990 by The American Society for Biochemistry
Vol. 265, No. 22, Issue of August 5. PP. 1282%X335,1990 Printed in U.S. A.
and Molecular Biology, Inc.
Reaction of LexA Repressor with Diisopropyl
Fluorophosphate
A TEST OF THE SERINE PROTEASE MODEL* (Receivedfor publication, March 5, 1990) Kenneth From
L. Roland+
the Departments
and John of $Biochemistry
W. LittleSQll and
§Molecular
and Cellular
The LexA repressor of Escherichia coli modulates the expression of the SOS regulon. In the presence of DNA damaging agents in vivo, the 202-amino acid LexA repressor is inactivated by specific RecA-mediated cleavage of the Ala-84/Gly-85 peptide bond. In vitro, LexA cleavage requires activated RecA at neutral pH, and proceeds spontaneously at high pH in an intramolecular reaction termed autodigestion. A model has been proposed for the mechanism of autodigestion in which serine 119 serves as the reactive nucleophile that attacks the Ala-84/Gly-85 peptide bond in a manner analogous to a serine protease, while uncharged lysine 156 activates the serine 119 hydroxyl group. In this work, we have tested this model by examining the effect of the serine protease inhibitor diisopropyl fluorophosphate (DFP) on autodigestion. We found that DFP inhibited autodigestion and that serine 119 was the only serine residue to react with DFP. We also examined [3H]DFP incorporation by a number of cleavage-impaired LexA mutant proteins and found that mutations in the proposed active site, but not in the cleavage site, significantly reduced the rate of [3H]DFP incorporation. Finally, we showed that the purified carboxyl-terminal domain, which contains the proposed catalytic residues, incorporated [3H]DFP at a rate indistinguishable from the intact protein. These data further support our current model for the mechanism of autodigestion and the organization of LexA.
The LexA protein of Escherichia coli regulates the expression of a set of genes, collectively termed the SOS regulon (1, 2), which are involved in the repair of DNA damage. When cells are treated by a DNA damaging agent, such as UV irradiation, the LexA repressor is inactivated by a proteolytic cleavage event at the Ala-84/Gly-85 peptide bond (l-3), resulting in derepression of the SOS genes (1,2). LexA cleavage in vivo requires functional RecA protein (4), which becomes activated by an inducing signal, probably single-stranded DNA (5). Activated RecA is also required for the cleavage of an SOS protein, UmuD (6-8), as well as a number of bacteriophage repressors, including Xc1 (5, 9). LexA cleavage is also observed in vitro at physiological pH in the presence of RecA protein and two types of cofactors, single-stranded DNA and a nucleoside triphosphate such as * This work was supported by Grant GM24178 (to J. W. L.) and Grant GM12390-03 (to K. L. R.) from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ?l To whom correspondence should be addressed: Dept. of Biochemistry, Rm. 347, Biological Sciences West, University of Arizona, Tucson, AZ 85721.
Biology,
University
of Arizona,
Tucson,
Arizona
85721
dATP (10). RecA and its cofactors form a ternary complex (11-13) which is the activated form of RecA (1). RecAindependent cleavage is also observed in vitro at alkaline pH, yielding fragments identical to those observed for the RecAmediated reaction (14). The RecA-independent cleavage reaction is termed “autodigestion” (14), and its existence suggests that the effect of RecA is indirect. In this view, the catalytic residues important in the cleavage reaction are located in the LexA protein, and RecA serves to stimulate autodigestion (14, 15). This idea is strengthened by the observations that substrates for the RecA-mediated reaction, including UmuD protein (6, 7) and Xc1 (14) also undergo alkali-stimulated autodigestion. These proteins also share a number of conserved residues in the carboxyl-terminal portions of the proteins (16, 17). Additionally, a large number of non-inducible (Ind-) LexA mutants that were isolated as defective in RecA-mediated cleavage (18) were also defective in autodigestion (18, 19). Previous genetic and biochemical studies (15, 20) have led to the formulation of a model for the mechanism of LexA autodigestion (Fig. 1, Ref. 15). This model involves two residues that are conserved among LexA, UmuD, and the bacteriophage repressors. According to this model, serine 119 (Ser119) and lysine 156 (Lys-156) lie near each other in the folded protein. The hydroxyl group of Ser-119 serves as a nucleophile to attack the Ala-84/Gly-85 peptide bond, analogous to the action of a serine protease such as trypsin. The deprotonated form of the t-amino group of Lys-156 is also required. Although its role is not well understood, it may activate the serine hydroxyl group. In support of this model, when either Ser-119 or Lys-156 are replaced by alanine, the resulting proteins have normal in vivo repressor activity and normal in vitro thermostability but are completely deficient in autodigestion and RecA-mediated cleavage (15), indicating that these two residues are absolutely required for both types of cleavage reactions. Additional support for this model comes from a recent study in which 20 independently isolated IndLexA mutations were identified by a genetic screen (18). All of the mutations were clustered in regions at or near the proposed active site or cleavage site residues (18), indicating that these regions of the protein are important for cleavage. If the mechanism of LexA autodigestion is analogous to that of a serine protease, then serine protease inhibitors such as diisopropyl fluorophosphate (DFP)’ should inhibit LexA autodigestion. DFP inhibits serine proteases by forming a covalent adduct between the nucleophilic oxygen atom of an activated serine and the diisopropyl phosphoryl (DIP) portion of DFP (21, 22). The resulting Ser-DIP residue is inactive for 1The abbreviations used are: DFP, diisopropyl fluorophosphate; DIP, diisopropyl phosphoryl; CAPS, 3-(c&lohexylamino)-i-propanesulfonic acid; Pipes, 1,4-piperazinediethanesulfonic acid; high performance liquid chromatography.
12828
HPLC,
DFP Inhibits
.LaA Autodigestion Chemicals
and
12829 Enzymes-CNBr
and
trypsin
were
obtained
from
Sigma. Trifluoroacetic acid, acetonitrile, and isopropyl alcohol were HPLC-grade and were obtained from Pierce Chemical Co. DFP was
N-Terminal
Domain
-
N-Terminal
FIG.
1. Model
for
LexA
Ala
-C
+ C-Terminal
Fragments
autodigestion.
Taken from Ref. 15.
proteolytic cleavage. Preliminary experiments in our laboratory indicated that concentrations of DFP up to 1 mM did not inhibit LexA autodigestion, as judged by visual examination of Coomassie Blue-stained gels (15). This negative result was reconciled with the model by either or both of two interpretations. First, since autodigestion is an intramolecular reaction, the active site serine may not be accessible to solvent, and therefore to the DFP. Second, since the rate of autodigestion is rather slow as compared with the catalytic rates of most serine proteases, the active site serine in LexA may not be as highly activated as the respective serines found in trypsin or chymotrypsin, and is therefore less reactive to DFP. In this study, we show that LexA autodigestion is inhibited by higher concentrations of DFP. We also show that [3H] DFP reacts selectively with the proposed catalytic nucleophile serine 119. Additionally, we examine [3H]DFP incorporation by a number of non-cleavable (Ind-) LexA mutant proteins, as well as the carboxyl-terminal fragment of LexA generated by autodigestion. MATERIALS
AND
METHODS
Construction of MTI18-The LexA mutant protein MT118’ was constructed by site-directed mutagenesis as described (15). The oligodeoxyribonucleotide used to introduce the single nucleotide substitution was 5’TCAGCGGGACGTCGATGAA3’ (the nucleotide change is underlined). Purification derivatives
of
of Proteins-LexA LexA protein;
GD80,
repressor GV80,
(14) and the VM82, AT84,
GE117, MT118, SA119, KR156, and KA156, described. described
The (23).
Pipes-NaOH,
carboxyl-terminal Purified proteins
pH 7.0, 0.1
mM
mutant
VF115, (19) were purified as
cleavage fragment was purified were stored in buffer D (IO
EDTA, 200
mM
as mM
NaCI, and 10% glyc-
erol). The protein concentration of LexA’ was determined spectrophotometrically using .&,, = 7300 M-’ . cm-’ (24). Due to a slightly lower level of protein purity in the preparations of mutant proteins (approximately 90 to 95% purity of mutant proteins versus approximately 99% purity of LexA’), the concentrations of the mutant proteins were estimated by visual examination of SDS-polyacrylamide gels (25) stained with Coomassie Blue (26). The estimated concentrations may vary from actual levels by as much as 20%. ‘The scheme for naming the LexA mutants makes use of the single-letter designation for the amino acids. The first letter denotes the amino acid residue in the wild-type protein, and the second letter designates the amino acid substitution. The number specifies the position in the LexA amino acid sequence where the mutation has occurred. Thus, in the mutant protein MT118, a threonine residue replaces the normally occurring methionine residue at amino acid 118 of the LexA sequence.
obtained from Aldrich and was dissolved in isopropyl alcohol to a concentration of 1 M. [1,3-3H]DFP was obtained from Du Pont-New England Nuclear or Amersham Corp. and was provided dissolved in propylene glycol. The specific activity varied between lots from 3.0 to 5.8 Ci/mmol. We found no qualitative differences in our results between [1,3-3H]DFP obtained from either source. Because the [3H] DFP was provided in propylene glycol, we tested the effects of several concentrations of propylene glycol on [3H]DFP incorporation by LexA. We found that [3H]DFP incorporation was not inhibited by propylene glycol concentrations of up to 11% (data not shown). However, at a propylene glycol concentration of 20% (Figs. 8 and 9), there was a 20% decrease in the total number of counts incorporated (data not shown). Estimation of the Stability of DFP-[3H]DFP (6.6 pM, 3.0 Ci/ mmol) was incubated at 37 “C in the presence of either 50 mM CAPS, pH 10.5, or 400 mM CAPS, pH 10.5. To assess the amount of DFP inactivated by hydrolysis under these conditions, its reactivity with trypsin was measured. An aliquot was taken prior to the addition of CAPS (time 0) and at subsequent 15-min intervals, and added to 90J reaction mixtures, such that the final concentration of reactants was 0.66 pM [3H]DFP (at time 0), 21 pM trypsin, 100 mM ammonium bicarbonate, pH 7.8, and 10 mM CaC12. After incubation for 10 min at 37 “C, reactions were stopped by the addition of 1 ml of cold 5% trichloroacetic acid and the mixture kept on ice for at least 20 min. The samples were further treated and analyzed by electrophoresis as described for Fig. 2. Protein bands were cut out of the gels and incorporation of [3H]DFP into trypsin was determined by scintillation counting, as described (20). The half-life of [3H]DFP was estimated by plotting the log of counts incorporated into trypsin versus time. The half-life in 50 mM CAPS, pH 10.5, and 400 mM CAPS, pH 10.5, was 30 and 10 min, respectively. These values are shorter than those previously reported (approximately 1 h, Ref. 27), but those authors measured stability at 25 “C with a different buffer. The difference in half-lives at the two concentrations of CAPS may be due to general base catalysis (28). In experiments at 20 mM DFP, it was necessary to use 400 mM CAPS: 50 mM CAPS did not adequately buffer the reactions (data not shown), since a proton is released upon hydrolysis of DFP. Purification and Analysis of PHIDIP Peptides-For preparative digests, LexA protein was reacted with DFP (final volume 1 ml) in a solution whose final composition was 90 pM LexA, 20 mM [3H]DFP (5 mCi/mmol), 400 mM CAPS, pH 10.5, 2% isopropyl alcohol, 10% propylene glycol, and 48% buffer D. After incubation at 37 “C for 30 min, 4 ml of 100 mM ammonium bicarbonate, pH 7.8 was added, and the sample was dialyzed at 4 “C against four changes, 800 ml each, of 100 mM ammonium bicarbonate, pH 7.8. The LexA-DIP formed a white precipitate during dialysis, but this did not interfere with subsequent digestions. For analytical digests, the labeling reactions were the same, except the protein concentration was 9 /LM. For digestion with trypsin, 4 ml of dialyzed material (approximately 1.5 mg) was used. Twenty ~1 of trypsin (5 ng/Fl), dissolved in 100 mM ammonium bicarbonate, pH 7.8, was added. The sample was heated to 65 “C for 10 min. Another 20 ~1 of trypsin was added and the sample was incubated for 3 h at 37 “C. Twenty ~1 of trypsin was added after the first and second hours of incubation. The sample was dried under vacuum, washed twice with 1 ml of water, and suspended in 0.1% trifluoroacetic acid. The 3H-labeled peptide was purified on a Varian (Vista Series) model 5000 HPLC with a Vydac C,s reverse phase column equilibrated with 0.1% trifluoroacetic acid (solution A). For each purification step, the column was run at a flow rate of 1 ml/min. The eluate was continuously monitored for UV-absorbing material at 220 nm using a Variable wavelength detector UV-50. Oneml fractions were collected. The amount of radioactivity in the HPLC fractions was determined by scintillation counting of 20-~1 aliquots. The first run was eluted as described in Fig. 3. Fractions containing the major peak of radioactivity (peak Zn were pooled and rerun. For the second run, the column was eluted with a linear gradient of 0 to 35% solution B (0.08% trifluoroacetic acid, 70% acetonitrile) over a lo-minute period (3.5%/min), and then the % solution B was increased at a rate of O.OS%/min for 50 min. Labeled fractions were pooled, rerun under the same conditions, and all UV-absorbing peaks were collected manually. The peak containing radioactivity was identified, dried under vacuum, and stored at -20 “C. For cyanogen bromide digestion, aliquots of protein or tryptic peptide were suspended in 70% trifluoroacetic acid containing 0.3 g/
DFP Inhibits
LexA Autodigestion
ml cyanogen hromide (29). For preparative digests, the protein concentratlon was 1 me/ml. For analvtical dieests of LexA-[‘HIDIP. MT118-[“HIDIP, and the purified tryptic peptide, the protein’con: centrations were 05, 0.5, and 0.025 mg/ml, respectively. The labeled cyanogen bromide peptide was purified hy HPLC on a Vydac C,$ column and eluted as described for the isolation of labeled tryptic peptide, except that the gradient was started at 10% solution B and increased at a rate of 0.2%/min. Peak fractions were pooled and either run again using the same conditions for further purification or dried under vacuum to a small volume for amino acid analysis. Analytical digests were run as indicated (Figs, 5 and 7). Amino Acid Analysis and SeyuenceAnalysis-Amino acid analysis of purified peptides was carried out using an Applied Biosystems 420A Derivatizer/Analyzer Amino Acid Analyzer. Amino acid sequencing of purified peptides was done on an Applied Biosystems 477A Protein Sequencer. RESULTS
A min+ L
(30,
31): Val-Ser-Gly-Met-Ser-Met-Lys 115 117
’ ,J. W. Little,
unpublished
results
119
121
3
5
8
10
30
60
1
3
5
8
10
30
60
-,4,Orr~~OD~
min+
previous attempts to inhibit LexA autodigestion with DFP were unsuccessful (see Introduction), a more sensitive assay showed an interaction between LexA and DFP. When 15 ELM LexA protein was incubated with various concentrations of [ ‘H]DFP at pH 10.5, we found that a constant proportion of DFP, roughly 0.04%, reacted with LexA after a lo-min incubation (data not shown). Therefore, we reasoned that, at a concentration of 20 mM DFP, about 50% of the LexA should react, resulting in an easily detectable fraction of the molecules resistant to autodigestion. When the DFP concentration was increased to 20 mM, we found that the extent of autodigestion was significantly diminished (Fig. 2), although the initial rate of autodigestion did not appear greatly reduced (compare the l-, 3-, and 5-min time points between panels A and B). As a control for this experiment, DFP was inactivated by preincubation in 400 mM CAPS, pH 10.5, at 37 “C for 120 min before adding the LexA (the half-life of DFP under these conditions was 10 min; see “Materials and Methods”). This material had no inhibitory effect on autodigestion (data not shown), indicating that no breakdown product of DFP, such as fluoride ions, nor any alkali-stable impurity of the DFP preparation was responsible for the observed inhibition of autodigestion. On the basis of the requirement for a high ccncentration of DFP, we can explain why inhibition had not been seen previously. At 1 mM DFP, only about 3% of the LexA would have been resistant to autodigestion. This minor resistant fraction would not have been obvious by visual inspection of Coomassie Blue-stained gels, since those experiments were designed to examine the effect of DFP on initial rates, and therefore LexA autodigestion was not followed to completion.” DFP Reacts Selectively with Serine 119-To determine which amino acid residue interacts with DFP, LexA protein was reacted with 20 mM [‘HIDFP (5 mCi/mmol). Unreacted [‘H]DFP was removed by dialysis, and the labeled LexA protein was digested with trypsin. The LexA-[“HIDIP digest was fractionated on a Cl8 reverse-phase HPLC column. A major peak, representing 63% of the total recovered radioactivity, was observed (Fig. 3, peah II). The peak was further purified. Amino acid analysis of the purified peak yielded the following composition: Ala, Gly, Lys, 2 Ser, and 2 Met. This composition is consistent with that expected from the predicted sequence of the tryptic peptide (7-mer) that contains
1
B
LexA Autodigestion Is Inhibited by DFP-Although
Ser-119
0
0
L+-woor,-
NI-,
-(IEponm
FIG. 2. Kinetics of inhibition of LexA autodigestion by DFP. These SDS-polyacrylamide gels show LexA autodigestion in the presence (panelA) or absence (panel R) of 20 mM DFP. The reactlon volume was 400 ~1, and the final concentration of components was 15 @M LexA, 400 mM CAPS, pH 10.5, 20 mM DFP. 5% isopropyl alcohol, and 50% buffer D. All components except CAPS were combined and incubated at 37 “C for 3 min. An aliquot was taken (24 pl), and then prewarmed 1 M CAPS, pH 10.5, was added with rapid mixing to a final concentration of 400 mM. Incubation was continued at 37 “C. Aliquots (40 ~1) were taken at the indicated times and precipitated by addition to 1 ml of cold 5% trlchloroacetic acid. kept on ice for ~20 min and then centrifuged for 15 min in an Eppendorf microcentrifuge at 4 “C. Samples were washed once with 1 ml of ether, dried under vacuum. and susnended in 5 x loading buffer (19). The aliquots were then analyzed hy electrophoresis on‘a 15% SDS-polyacrylamide gel (25) and stained with Coomassle Blue R-250 (26).
II 2,000 -
- 100
1,500 r B
** 1,000 .-
.*
.'
.*
**
.-
**
.-
x! g z 3
-* a- .-
-50
500 2.
10
,- .*
.'
**
m
III
20
Fractton
30
40
50
Number
FIG. 3. Radioactivity profile of the first C,, HPLC column of a tryptic digest of LexA-[“HIDIP. A tryptic digest of LexA[‘HIDIP was loaded onto a C,” reverse-phase HPLC in 0.1% trifluoroacetlc acid. The column was brought from 0 to 10% solution B (0.08% trifluoroacetic acid, 70% acetonitrile) over a 5.min period. The %B was then increased at a rate of 1.8%/mm for 50 min.
The asterisk denotes the proposed nucleophile, Ser-119. This tryptic peptide is the only one in LexA with the composition indicated (30, 31). The amino acid sequence of the peptide was then determined by Edman degradation to be the sequence shown above. Only 25% of the radioactivity applied to the sequencer was recovered in the Edman degradation fractions, presumably
DFP Inhibits
LexA Autodigestion
due to the instability of Ser-[3H]DIP under the conditions used for sequencing. However, 72% of the recovered radioactivity eluted during the fifth round of Edman degradation (data not shown), the same round in which the second serine in the peptide (i.e. Ser-119) was identified, consistent with the hypothesis that DFP reacts selectively with Ser-119. No other round of Edman degradation yielded more than 6% of the total recovered counts (data not shown). Two other peaks of radioactivity eluted from the preparative HPLC column (peaks Z and ZZZ,Fig. 3). Peak I comigrates with a breakdown product of [3H]DFP (data not shown). Based on the position of its elution, and the fact that DFP can react slowly with solvent-exposed tyrosine residues, particularly at high pH (32, 33), peak III probably represents a 33-amino acid peptide that contains Tyr-98 (see below), the only tyrosine in LexA (30,31). Although the above data are strongly suggestive that DFP reacts selectively with Ser-119, the fact that there is an additional serine residue (Ser-116) in the tryptic peptide could potentially complicate our interpretation of the results. Therefore, we performed two additional experiments to provide independent evidence that Ser-119 is the reactive residue. Ser-119 is the only serine in LexA that is flanked by methionines (30, 31) and it would be expected to be found in a cyanogen bromide digestion product derived from only serine and methionine (see 7-mer sequence above and Fig. 4, panel A), since cyanogen bromide cuts on the carboxyl side of methionine residues. It is also the only serine which, if labeled with [3H]DFP, would yield the same 2-residue radioactive product from cyanogen bromide digestion of LexA-[3H]DIP or the purified 7-mer-[3H]DIP (Fig. 4, panel A). By contrast, if Ser-116 were labeled, the radioactive products for LexA[3H]DIP and 7-mer-[3H]DIP would be a 94-amino acid peptide and a tetra-peptide, respectively (30 and 31, Fig. 4, panel B). LexA-[3H]DIP and the purified tryptic peptide (7-mer[3H]DIP) were separately digested with cyanogen bromide. The digests were analyzed by HPLC. The results (Fig. 5) A. M SW-7 19 is labeled: tntact protf?,n
Ml-
II SGMSM -
M( 94 aa
22 aa
Tryptic peptide
1 M-
7aa t 2 aa
76 aa
“SG,‘,,? t 2 aa
6.
If SW1 76 is /abeled:
/ntact protein I M-M
I
II SGMSM (I
L
I M-
I 94 aa Tryptic peptide
I I VSGMSMK w 4 aa
FIG. 4. Predicted peptides from CNBr digestion of LexAr3H]DIP and 7-mer-r3H]DIP in which either Ser-119 or Ser116 are labeled. The numbers in boldface show the size of the predicted ‘H-labeled peptides if Ser-119 (panel A) or Ser-116 t’panel El is labeled. The arrows indicate points of cleavaze by CNBr. Kev residues are indicated in capital letters, using the single-letter symbol for the amino acids.
12831
25
30
Fraction
35
40
45
Number
FIG. 5. Radioactivity profiles from Cl8 reverse-phase HPLC analvses of CNBr digests of LexA?HlDIP and 7-merTH1 DIP:The dpm in 204 kquots from l-ml fractions of CNBr digests of LexA-[3H]DIP (m) and 7-mer-[3H]DIP (A) are plotted. The CNBr digestions were carried out and analyzed as described under “Materials and Methods,” except the %B was increased at a rate of 0.3%/ min. Four times as many dpm of the LexA-[3H]DIP were loaded as of the 7-mer-[3H]DIP.
show that both digests yield corn&rating labeled peptides. Under the conditions used for elution, a 94-amino acid peptide and a tetra-peptide would be separated by at least 20 fractions (data not shown). To verify the identity of this peptide, the labeled peak from the LexA-[3H]DIP digestion was further purified, and the amino acid composition was determined. The composition was Ser, Lys, Asp, Gly, 2 homoserine and 2 Ile. Homoserine is derived from the reaction of cyanogen bromide with methionine (29). This composition is consistent with the sum of two predicted CNBr peptides, Ser-Met and Lys-Asp-Ile-GlyIle-Met. Further attempts to separate these two peptides using different gradient conditions and different solvents systems were unsuccessful (data not shown). When purified 7-mer[3H]DIP was digested with CNBr and the labeled peak purified, the amino acid composition was only Ser and homoserine, as expected for the purified Ser-Met peptide. Although the above results were consistent with the idea that Ser-119 is labeled with [3H]DFP, the possibility remained that, due to the harsh conditions used for cyanogen bromide digestion, the peak of radioactivity that was purified may have been a breakdown product of the LexA-[3H]DIP that copurified with the Ser-Met peptide. To test this possibility, we took advantage of a mutant LexA protein, MT118, which autodigests at a rate similar to the rate of LexA autodigestion and reacts well with DFP (data not shown). Since MT118 does not have one of the methionines flanking Ser-119 (see 7-mer sequence above and Fig. 6, panel B), the expected radioactive cyanogen bromide digestion product would be a 96-residue peptide, instead of the dipeptide expected from cyanogen bromide digestion of LexA-[3H]DIP (30 and 31, Fig. 6, panel A). We separately digested LexA-[3H]DIP and the LexA mutant protein MT118-[3H]DIP with cyanogen bromide and analyzed the digests by reverse-phase HPLC. The results (Fig. 7) show that the radiolabeled products of the MT118-[3H] DIP digest do not comigrate with the radiolabeled products of the LexA-[3H]DIP digest. This shows that the conditions used for cyanogen bromide digestion do not cause breakdown of the serine-[3H]DIP moiety and, because the migration of the radiolabeled product from digestion of MT118-[3H]DIP shifted in a manner consistent with its being a larger peptide (Fig. 7, panel B, peak ZZZ),this result also provides further evidence that Ser-119 is the reactive residue. The data in Fig. 7 also show that there is a small peak in
12832
DFP A.
Inhibits
LexA
LexA+
Intact t I
M-M-AG 22 aa
-W-M-
76 aa
7aa
94 aa
t
Autodigestlon fragments
M-M-A{G----MSM-M 60 aa
34aa
l Y
2 88 6.
MT118 Intact
M-M-AG
--T$M
- M
I 96 88 AutodigestIon fragments t M-M-A+-
T$M,-
II
t M
1 I 36‘aa
FIG. 6. Predicted peptides from CNBr digestion of LexA and MT1 18. The predicted 3H-labeled fragments from CNBr digestion of LexA-[3H]DIP (panel A) or MT118-[3H]DIP (panel B) proteins and CNBr digestion of the autodigestion fragments of both proteins that reacted with [3H]DFP after cleavage. The numbers in boldface show the size of the predicted 3H-labeled peptides. The arrows indicate points of cleavage by CNBr. Key residues are indicated in capital letters, using the single-letter symbol for the amino acids. Ser119 is denoted by an asterisk. aa, amino acid.
500
- MTllII-DIP
10
20
30
40
50
HPLC Fraction FIG. 7. Radioactivity profiles from Cl8 reverse-phase HPLC of CNBr digests of LexA-[3H]DIP and MT118-[3H]DIP. [3H] DIP-labeled proteins were prepared as described under “Materials and Methods,” except that the protein concentration during labeling was 9 gM. 100 pg of LexA-[3H]DIP (panel A), 100 pg of MT118-[3H] DIP (panel B), or 100 pg of both proteins (panel C) were digested with CNBr as described under “Materials and Methods.” Peptides were eluted as described in Fig. 3. The peaks labeled I, II, and I11 are described under “Results.” Two times as many dpm of MT118-[3H] DIP as LexA-[3H]DIP were applied to the HPLC column. the LexA-[3H]DIP digest that comigrates with the major peak in the MT118-[3H]DIP digest (peak III). This is probably a partial digestion product, since Met-Ser bonds are often poor substrates for cyanogen bromide cleavage (29). If the Met-Ser bond is not cleaved, then the radiolabeled product would be a
Autodigestion 96-residue peptide and would therefore comigrate with the expected MT118-[3H]DIP product under these conditions. To control for the possibility that the differences in the radioactivity profiles were due to a difference in digestion conditions, the two proteins, LexA-[3H]DIP and MT118-[3H] DIP, were combined and digested together. The resulting HPLC profile (Fig. 7, panel C) is essentially identical to the combined profiles obtained from the two independent cyanogen bromide digestions (Fig. 7, panels A and B). A second labeled peak for the MT118-[3H]DIP digest (Fig. 7, panel B, peak II) was observed. Although this peak has not been rigorously identified, it probably represents a 36-residue peptide by the following reasoning. The 96-residue peptide resulting from cyanogen bromide digestion of MT118 contains both Ser-119 and the Ala-84/Gly-85 cleavage site (Fig. 6, panel B). Under the conditions used for reacting LexA and MT118 with [3H]DFP, the proteins are undergoing autodigestion. The amount of DFP used prevents cleavage of about 50% of the molecules, while the rest of the molecules are cleaved (Fig. 2 and data not shown). As will be shown below, the carboxyl-terminal fragment (C-fragment) of autodigestion, which contains the proposed catalytic site (including Ser-119), also reacts with DFP. If the [3H]DFP reacted with Ser-119 after autodigestion of MT118 had occurred, then the resulting 3H-labeled cyanogen bromide fragment would be 36 amino acids in size (Fig. 6, panel B). However, CNBr digestion of the C-fragment-[3H]DIP derived from wild-type LexA would yield the dipeptide Ser-Met (Fig. 6, panel A). Therefore, one would expect two peaks of radioactivity from the MT118L3H]DIP digest, but only one from the LexA-13H]DIP digest (Fig. 6). From the above data, we conclude that DFP reacts selectively with Ser-119, the proposed nucleophile in LexA autodigestion. We use the term selective to refer to this reaction because the hydroxyl group of serine does not normally react with DFP (22, 34) unless the serine is involved in a specific three-dimensional structure, such as that found in serine proteases (21,22). Kinetics of PHIDFP Incorporation by LexA and the Noncleavable LexA Mutant Proteins SAll9 and AT84-Our working model for LexA autodigestion suggests that there are two distinct components involved in cleavage, an active site which includes Ser-119 and Lys-156 and a cleavage site which includes Ala-84 and Gly-85. Additional residues near each of these sites may also be involved. Therefore, based on this idea and the above results, one can predict that cleavage-impaired LexA mutant proteins fall into at least two classes, based on their ability to react with DFP. Proteins with mutations at or near the active site would not be expected to react with DFP, while those with mutations at or near the cleavage site may react with DFP with initial rates of incorporation similar to LexA+. To test this prediction, we studied the kinetics of [3H] DFP incorporation by LexA+ and two LexA mutant proteins, SA119 (15) and AT84 (18, 19), which contain amino acid substitutions in the predicted active site and cleavage site, respectively. LexA
protein
incorporated
[3H]DFP
at
an
initial
rate
of
approximately 30 dpm/min/pg (Fig. 8). The rate of incorporation into intact LexA decreases at later times because, under these conditions (57 NM DFP), most of the LexA molecules autodigest with a half-life of approximately 10 min, so that the number of LexA molecules available to react with the DFP decreases significantly over the time course of the experiment. By 1 h, greater than 95% of the LexA molecules are cleaved, as judged by examination of Coomassie-stained gels (data not shown). Thus, there is a competition between
DFP
Inhibits
LexA
Autodigestion
12833 TABLE
I
3H-DFP
5-c)OO
0
w
0
10
20
30
40
Time
(minutes)
incorporation into LexA mutant proteins during autodigestion at pH 10.5 Results are the average of at least two experiments. These experiments were carried out as described for Fig. 8, except that the reaction constituents were 14 FM protein, 33 PM [3H]DFP (3.0 Ci/mmol), 50 mM CAPS, pH 10.5, 10% propylene glycol, and 60% buffer D. Reactions were incubated for 10 min at 37 “C to measure relative initial rates (see Fie. 8).
LexA
SA119 50
60
8. Kinetics of [3H]DFP incorporation by LexA+, AT84, and SAllQ. Reactions were carried out at 37 “C. Reaction constituents were 30 pM protein, 57 pM [3H]DFP (5.8 Ci/mmol), 50 mM CAPS, pH 10.5, 20% propylene glycol, and 52% buffer D. All components except CAPS were combined and incubated at 37 “C for 3 min. The reactions were initiated by the addition of CAPS. Aliquots were taken at indicated times, precipitated with trichloroacetic acid, and analyzed by electrophoresis as described for Fig. 2. Bands of intact protein were cut out of the gels and [3H]DFP incorporation was determined by scintillation counting as described (20). FIG.
% Wild-type dpm
% Wild-type Cleavage”
Protein
dpm (*SD)
LexA+
2535
(84)
100
100
GD80 GV80 VM82 AT84
2572 2464 3571 2430
(15) (450) (316) (265)
100 100 140 100
1.0
