(Received for publication, October 30, 1996, and in revised form, January 22, 1997)
From the Department of Chemistry, The Ohio State
University, Columbus, Ohio 43210 and the ¶ MRC Group in Protein
Structure and Function, Department of Biochemistry, University of
Alberta, Edmonton, Alberta T6G 2H7, Canada
Escherichia coli leader peptidase,
which catalyzes the cleavage of signal peptides from pre-proteins, is
an essential, integral membrane serine peptidase that has its active
site residing in the periplasmic space. It contains a conserved lysine
residue that has been proposed to act as the general base, abstracting the proton from the side chain hydroxyl group of the nucleophilic serine 90. To help elucidate the role of the essential lysine 145 in
the activity of E. coli leader peptidase, we have combined site-directed mutagenesis and chemical modification methods to introduce unnatural amino acid side chains at the 145-position. We show
that partial activity can be restored to an inactive K145C leader
peptidase mutant by reacting it with 2-bromoethylamine·HBr to produce
a lysine analog (-thia-lysine) at the 145-position. Modification
with the reagents 3-bromopropylamine·HBr and 2-mercaptoethylamine also allowed for partial restoration of activity showing that there is
some flexibility in the length requirements of this essential residue.
Modification with (2-bromoethyl)trimethylammonium·Br to form a
positively charged, nontitratable side chain at the 145-position failed
to restore activity to the inactive K145C leader peptidase mutant. This
result, along with an inactive K145R mutant result, supports the claim
that the lysine side chain at the 145-position is essential due to its
ability to form a hydrogen bond(s) or to act as a general base rather
than because of an ability to form a critical salt bridge. We find that
leader peptidase processes the pre-protein substrate,
pro-OmpA nuclease A, with maximum efficiency at pH 9.0, and apparent
pKa values for titratable groups at approximately
8.7 and 9.3 are revealed. We show that the lysine modifier maleic
anhydride inhibits leader peptidase by reacting with lysine 145. The
results of this study are consistent with the hypothesis that the
lysine at the 145-position of leader peptidase functions as the active
site general base. A model of the active site region of leader
peptidase is presented based on the structure of the E. coli UmuD
, and a mechanism for bacterial leader peptidase is
proposed.
Escherichia coli leader (signal) peptidase is an integral membrane serine protease that functions to cleave off the amino-terminal leader (signal) sequence from proteins that are targeted to the cell surface of bacteria. Leader peptidase has been cloned (1), sequenced (2), overexpressed (3, 4), and purified (4, 5). The use of protease inhibitors to classify leader peptidase into a specific protease class has failed (6, 7). Site-directed mutagenesis studies have demonstrated that there is an essential serine 90 (8) and lysine 145 (9, 10) but no essential histidines or cysteines (8). The most convincing evidence for serine 90 being the nucleophile in the proteolytic reaction was provided by the work of Tschantz et al. (10) which showed that when serine 90 is replaced with a cysteine, leader peptidase is still active and that this thiol leader peptidase could then be inhibited by reacting it with the cysteine-specific reagent N-ethylmaleimide (10).
The essential serine and lysine are fully conserved within the type 1 prokaryotic and mitochondrial signal peptidases (11, 12). There are 19 lysines in the E. coli leader peptidase and only lysine 145 is conserved. Interestingly, the essential lysine is replaced by a histidine in the homologous yeast, chicken, and canine endoplasmic reticulum signal peptidase subunits (11, 12).
All evidence to date points toward leader peptidase utilizing a serine/lysine dyad mechanism. With this mechanism, the lysine 145 would act as the general base to abstract the proton from the hydroxyl group of the serine 90 side chain, thereby allowing for the nucleophilic attack on the scissile peptide bond of the translocated pre-protein substrate. Leader peptidase along with LexA (13), UmuD (14), and most recently, Tsp protease (15) represent the most thoroughly characterized members of the class (Clan) of serine proteases that contain an essential lysine but no essential histidine residue (16).
For lysine to act as a general base its side chain amine must be
unprotonated. For this to be possible an enzyme must provide an
environment for the lysine in which its pKa would be
depressed. The microenvironment near the lysine would include either a
local positive charge or a hydrophobic surrounding. There are many
examples of lysine residues that have significantly lower pKa values as compared with the
pKa of 10.5 for lysine in solution (17, 18). One of
the most carefully studied active site lysines is that of acetoacetate
decarboxylase, which has a pKa of 6.0 (19, 20).
Moreover, there is crystallographic evidence that lysine is capable of
serving as a general base for a serine residue in -lactamase (21)
and most recently in the structure of the E. coli UmuD
protein (14). These two enzymes have catalytic sites that are
superimposable (14).
To further investigate the role played by the essential lysine 145 in
the catalysis of leader peptidase, we have combined site-directed
mutagenesis and chemical modification. To simplify the interpretation
of the results, we have constructed and purified an active mutant in
which all native cysteine residues have been replaced with serine
residues. This cysteine-less variant of leader peptidase will be
referred to as No Cys1 leader peptidase. We
then made the K145C mutant in the No Cys leader peptidase to produce a
single cysteine within the enzyme for modification. This inactive
mutant, which will from now on be referred to as K145C, No Cys, regains
activity after being reacted with the reagent 2-bromoethylamine·HBr
to form the lysine analog -thia-lysine at the 145-position. We have
measured with the wild-type enzyme the pH-rate profile using the
pre-protein substrate pro-OmpA nuclease A. We also show that leader
peptidase is inhibited by the lysine-modifying reagent maleic
anhydride, and we provide evidence that the inactivation is due to its
modification at lysine 145. In addition, we have modeled the active
site of E. coli leader peptidase based on the structure of
E. coli UmuD
(14) and proposed a mechanism for leader
peptidase. The study reported here is consistent with leader peptidase
utilizing a serine/lysine dyad mechanism in its catalysis.
The 2-bromoethylamine·HBr (2BEA), 3-bromopropylamine·HBr (3BPA), (2-bromoethyl)trimethylammonium·Br (2BETMA), 2-mercaptoethylamine (2-MEA), and (S)-2-aminoethyl-L-cysteine·HCl were purchased from Sigma. The maleic anhydride was from Matheson Coleman & Bell. Oligonucleotides were synthesized at the Biochemical Instrument Center at The Ohio State University. The peptide substrate (NH3+-Phe-Ser-Ala-Ser-Ala-Leu-Ala-Lys-Ile-COO-) was purchased from the macromolecular Structure Analysis Facility at the University of Kentucky.
Bacterial Strains and PlasmidsThe leader peptidase
proteins were expressed in MC1061 E. coli cells harboring
the pING plasmid carrying the mutant leader peptidase gene. Six
consecutive histidine residues were engineered by
oligonucleotide-directed mutagenesis into the P1 (cytoplasmic) domain
of leader peptidase that allowed us to purify the mutants from the
chromosome-expressed wild-type leader peptidase. Briefly, amino acid
residues 35-40 were substituted with histidine residues. The sequence
of the oligonucleotide used is as follows: 5-TTC GCA CCT AAA CGG CGG
CGC GAA CGT CAT CAT CAT CAT CAT CAT GCT CGG GAC TCA CTG GAT AAA GCA-3
.
The No Cys variant of E. coli leader peptidase was produced
by using oligonucleotide-directed mutagenesis to replace the three
cysteines at positions 21, 170, and 176 with serine residues. The
oligonucleotides used to make the cysteine to serine mutations are
described in Sung and Dalbey (8). The cloning of the pro-OmpA nuclease
A gene into the
isopropyl-
-D-thiogalactopyranoside-inducible plasmid
pONF1 (22) and the overexpression of the protein were described by
Chatterjee et al. (23).
The DNA techniques were performed as described by Sambrook et al. (24). All cloning procedures used T4 kinase, T4 DNA ligase, Klenow, and restriction enzymes from Life Technologies, Inc.. Oligonucleotide-directed mutagenesis was performed as described by Zoller and Smith (25). Transformations followed the calcium chloride method of Cohen et al. (26).
Purification of the 6-His Tagged Leader Peptidase ProteinsThe 6-His tag/nickel affinity chromatography method (27)
was used to purify the overexpressed leader peptidase mutants away from
the wild-type chromosome-expressed copies of leader peptidase. All
leader peptidase proteins used in this study, except the wild-type, contained the 6-His tag. E. coli MC1061 cells containing the
pING plasmid encoding the mutant leader peptidase protein were grown in
M9 minimal media (1-8 liters) containing 100 µg/ml ampicillin until
an absorbance of 0.5 at 600 nm was reached. Expression was induced by
the addition of arabinose to a final concentration of 0.3%, and the
incubation of the cultures was continued for 4 h. The cells were
pelleted and then resuspended in an equal weight of 50 mM
Tris, pH 7.5, 10% sucrose. The cells were frozen by dropping them into
liquid nitrogen and stored at 80 °C until needed. 10 g of
frozen cell nuggets were added to 25 ml of thaw buffer (50 mM Tris, pH 7.5, 20% sucrose) and thawed. Lysozyme (6 mg)
and DNase (60 µl at 10 mg/ml) were added to the thawed cells and then
stirred for 10 min. The mixture was then freeze/thawed in a dry
ice/ethanol bath. 200 µl of 1 M magnesium acetate was added, and the solution was allowed to stir for 15 min at room temperature. The solution was then centrifuged at 18,000 rpm (4 °C, 30 min), and the pellet was resuspended in 25 ml of 10 mM
triethanolamine, 10% glycerol, pH 7.5. The centrifugation step
was then repeated once more. The pellet was resuspended by douncing in
binding buffer (5 mM imidazole, 0.5 M NaCl, 20 mM Tris, pH 8.0, 1% Triton X-100, 10 mM
-mercaptoethanol). The suspension was then centrifuged again at
18,000 rpm. The supernatant was loaded onto a 1-ml nickel column
(Novagen resin) that was equilibrated with the same buffer. The column
was then washed with 20 ml of binding buffer followed by a second wash
with wash buffer (60 mM imidazole, 0.5 M NaCl, 10 mM Tris, pH 8.0, 1% Triton X-100, 10 mM
-mercaptoethanol). The 6-His tagged leader peptidase was eluted by
using an imidazole step gradient from 100 to 500 mM
imidazole. Eluted fractions were assayed for protein by SDS-PAGE
followed by Coomassie staining. Fractions containing 6-His leader
peptidase were dialyzed against 50 mM Tris-HCl, pH 8.0, 1%
Triton X-100, 10 mM
-mercaptoethanol and then stored at
80 °C.
The E. coli strain SB211 containing the plasmid pONF1 was used to overexpress the pro-OmpA nuclease A substrate, a hybrid of the signal peptide of the E. coli outer membrane protein A (OmpA) fused to staphylococcal nuclease A (22). The pro-OmpA nuclease A was expressed and purified as described by Chatterjee et al. (23).
Kinetic Assay Using Pro-OmpA Nuclease ATo determine the
kinetic constants (Vmax,
kcat, and Km) of the
wild-type and mutant leader peptidase proteins, we used pro-OmpA
nuclease A as a substrate. Substrate concentrations were determined by
using an E1% at 280 nm of 8.3 (23). The
cleavage reactions (75 µl) were run in TGC buffer (50 mM
Tris, 50 mM glycine, 50 mM CAPS, 10 mM CaCl2, 1% Triton X-100) at pH 9.0, unless
indicated otherwise, containing the substrate at five different
concentrations (35.2, 17.6, 13.2, 8.8, 4.4 µM). The reaction was initiated by the addition of leader peptidase (wild-type or mutant) at a final concentration of 1.37 × 104
µM, which was determined by the Pierce BCA protein assay
kit. The reaction was carried out at 37 °C, and aliquots of the
reaction were removed at various times such that less than 7%
processing of the substrate was achieved. The reaction was stopped by
the addition of 5 µl of 5 × sample buffer containing 10 mM MgCl2, and the samples were frozen
immediately in a dry ice/ethanol bath. The amount of pro-OmpA nuclease
A that was processed by leader peptidase was assayed by SDS-PAGE on a
17.2% gel, followed by staining with Coomassie Brilliant Blue. The
precursor and mature proteins were quantified by scanning the gels on a
Technology Resources, Inc. Line Tamer PCLT 300 scanning densitometer.
Percent processing was determined by dividing the area of the mature
protein band by the sum of the mature and precursor band areas. The
initial rates were determined by plotting the amount of product
versus time. The Vmax,
Km, and kcat values were
calculated from a 1/vi versus 1/[S] plot
(where vi represents initial velocity and [S]
indicates substrate concentration). We used the computer program
Microcal Origins to plot the data and for linear regression analysis of
the data. All values are from at least two different experiments.
The kinetic constants in the pH range 7 to 11 were measured using the substrate pro-OmpA nuclease A. The kinetic reactions were run as described above with the exception that the cleavage reactions were carried out in TGC buffer at the indicated pH values (7.0, 8.0, 8.5, 9.0, 9.5, 10.0, 11.0). Similar to the other kinetic assays, the amount of pro-OmpA nuclease A that was processed was quantified by SDS-PAGE and densitometry, and the kinetic parameters were extracted by using Lineweaver-Burk plots of the initial velocity values and substrate concentrations. The results plotted are the average of two separate experiments.
pH Stability StudyLeader peptidase (0.7 µg/ml) was assayed for its pH stability by dilution into TGC buffer at various pH values and incubated for 15 min (4 °C). 1 µl was then removed and added to 15 µl of pro-OmpA nuclease A (15 µM) in TGC buffer, pH 9.0, and incubated at 37 °C for 1 h. Reactions were terminated by the addition of 5 µl of 5 × sample buffer and then frozen. The amount of processing was determined by separating the precursor and the mature form by 17.2% SDS-PAGE, followed by Coomassie Brilliant Blue staining. The data were analyzed as described above using scanning densitometry. Results plotted are the average of two separate experiments.
Chemical Modification of the K145C, No Cys Leader PeptidaseThe reactions were carried out basically as described by Smith and Hartman (28). Briefly, leader peptidase at a concentration of 1-4 mg/ml in 50 mM Tris-HCl, pH 8.5, 1% Triton X-100 was treated with a sufficient amount of freshly prepared 2 M 2BEA (18, 28-36), 2 M 3BPA (29, 30), 2 M 2BETMA (30, 31), or 2 M 2-mercaptoethylamine (2MEA) (29) such that the final concentration of reagent was 100 mM. All reagents were titrated to pH 8.5 before adding them to the enzyme. The reaction solution was then covered with nitrogen and incubated at room temperature overnight. The reaction mixture was then dialyzed against 2 volumes of 4.5 liters of 50 mM Tris-HCl, pH 8.0, 1% Triton X-100 at 4 °C. All buffers and solutions were purged with nitrogen. The extent of reaction was assayed by quantification of cysteine residues with DTNB before and after reaction with the reagents. The extent of the 2BEA reaction was also quantified directly by amino acid composition.
Quantifying Cysteines with DTNBA solution of
5,5-dithiobis(2-nitrobenzoic acid) (DTNB, 100 µl at 10 mM) in 0.1 M phosphate, pH 7.28, 1 mM EDTA was added to 275 µl of 4 mg/ml leader peptidase
and 625 µl of 6.4 M guanidine HCl. This mixture was
incubated for 15 min at room temperature, and then its absorbance was
measured at 412 nm. A molar absorbance coefficient of 13,700 M
1 cm
1 was used to calculate
the moles of cysteine present in the samples (37). All solutions were
made up fresh and purged with nitrogen. The concentration of the
protein was determined using the Pierce BCA method. The K145C, No Cys
leader peptidase was stored in reducing conditions. Before the DTNB
reaction, or other modifications, the
-mercaptoethanol was dialyzed
away in nitrogen-purged buffer (2 volumes of 4.5 liters of 50 mM Tris-HCl, pH 8.0, 1% Triton X-100) at 4 °C.
To directly quantify the
extent of aminoethylation of the cysteine 145, we measured the
appearance of a -thia-lysine residue within the leader peptidase
protein by using
(S)-2-aminoethyl-L-cysteine·HCl (Sigma) as a
standard in the amino acid compositional analysis. The amino acid
analysis was performed at the W. M. Keck Foundation Biotechnology
Research Laboratory in New Haven, CT, and the Biochemical Instrument
Center at The Ohio State University.
In the time-dependent studies, maleic anhydride dissolved in Me2SO was added to leader peptidase (0.07 mg/ml, in 10 mM Tris, pH 8.0, 5 mM EDTA, 150 mM NaCl, 2.5% Triton X-100) to a final concentration of 0.5 mM, and samples were removed at 0, 10, 30, and 60 min (38). In the concentration-dependent studies, leader peptidase (0.07 mg/ml, in 10 mM Tris, pH 8.0, 5 mM EDTA, 150 mM NaCl, 2.5% Triton X-100) was incubated at room temperature with maleic anhydride at a final concentration of 0.10, 0.25, 0.30, 0.50 mM for 60 min. After the modification, the reaction mixture was added to 12.5 µl of peptide substrate (1.5 mg/ml; NH3+-Phe-Ser-Ala-Ser-Ala-Leu-Ala-Lys-Ile-COO-) (39) and incubated for 3 h at 37 °C. Reactions were terminated by the addition of an equal volume of 0.1% trifluoroacetic acid and microcentrifuged for 5 min at room temperature to pellet any particulate matter. Samples were then analyzed by high performance liquid chromatography. The elution gradient utilized to separate the cleaved from the uncleaved peptide was as follows: 97% A, 3% B held constant for 5 min, followed by a linear gradient to 60% A, 40% B over a 10-min period. This mixture was then held constant for 5 min and then brought back down to 97% A, 3% B over a 5-min period. The solvents used were A = 0.1% trifluoroacetic acid and B = 0.1% trifluoroacetic acid in acetonitrile. Peptide products were detected spectrophotometrically at 218 nm. The column used was a 25-cm Vydac C18 column. Percent processing of the peptide was determined by quantification of the 7-mer product peptide peak and the 9-mer substrate peak: % processing = (area 7-mer peak/(area 7-mer peak + area 9-mer peak)) × 100.
Combined Maleic Anhydride and 2-Bromoethylamine ReactionA
total of 0.007 µmol of the No Cys or the K145C, No Cys mutant leader
peptidase was first reacted with 0.7 µmol of DTNB and then dialyzed
against 50 mM Tris-HCl, 0.5% Triton X-100, pH 8.5 (buffer
A). The mutants were then reacted with 2.3 µmol of maleic anhydride,
incubated at room temperature for 30 min, and then dialyzed against
buffer A. The cysteine in the K145C, No Cys mutant was then deprotected
by reaction with -mercaptoethanol followed by dialysis. These
samples (60 µl) were then reacted overnight at room temperature with
10 µl of freshly prepared 2 M 2-bromoethylamine·HBr. This product was then reacted with maleic anhydride again. All reagents
were made up fresh in nitrogen-purged buffer. After each reaction the
samples were dialyzed extensively (overnight against 9 liters of
nitrogen purged buffer A). The activity of the leader peptidase enzymes
was assayed before and after each chemical modification step.
We have aligned the amino acid sequence of
the catalytic region from the solved structure of UmuD (residues
40-139, the total length of the UmuD protein is 218 residues) with the
corresponding proposed catalytic region of leader peptidase (residues
75-202, the total length of leader peptidase is 323 amino acid
residues). The alignment protocol XALIGN (40, 41) used is a derivative of the NW_ALIGN program originally developed for SEQSEE (41). The
pairwise alignment module implemented in the comparison of UmuD
and
leader peptidase is based on the Needleman-Wunsch dynamic programming
algorithm (42), and the sequence/structure alignment algorithms are
based loosely on the protocols described by Lesk et al.
(43). Residue anchoring and residue clustering features are analogous
to the gap and extension penalties incorporated into regular dynamic
programming schemes (40). The appropriate substitutions of the UmuD
residues as well as manual manipulations were done using the program
TOM-frodo (44). The coordinates for the crystal structure of UmuD
were
kindly provided by T. Peat and W. Hendrickson (14). The rendering
of the modeled active site region and hydrophobic cleft adjacent
to the active site of E. coli leader peptidase was
created using the program Raster3D (45).
The use of the 6-His tag/nickel affinity chromatography method (27) of purification has allowed us to purify the overexpressed mutants of leader peptidase away from the wild-type leader peptidase expressed by the E. coli chromosome. Due to leader peptidase's excellent kinetic properties, background activity from the chromosomal wild-type copies of leader peptidase was always a concern when assessing the activity of overexpressed mutants. A mock purification run was performed to demonstrate that there was no detectable background wild-type leader peptidase activity from this purification procedure. Cells containing the expression vector with no leader peptidase gene insert were lysed and brought through the nickel affinity column procedure. All fractions eluted from the column showed no detectable activity (data not shown). The yield of each purified mutant was approximately the same. We obtained typically 4.6 mg of purified K145C, No Cys from 10 g of frozen cell nuggets (see "Experimental Procedures").
Activity of the Wild-type and Mutant 6-His Tagged Leader PeptidasesThe wild-type leader peptidase showed very impressive
catalytic constants (kcat = 120 s1, Km = 10.9 µM,
kcat/Km = 1.1 × 107 s
1 M
1) when
analyzed with the pre-protein substrate pro-OmpA nuclease A at its
optimal pH of 9.0 (see pH profile below and Table I). Similarly, the No Cys variant of leader peptidase with a 6-His tag
showed an almost wild-type kcat value of 110 s
1, yet its Km of 20.6 µM was approximately double the wild-type value. In
addition, the conserved arginine residue at the 146-position of leader
peptidase does not play a critical role in catalysis. The R146A mutant
showed only slightly depressed kinetic constants
(kcat = 36.8 s
1,
Km = 29.0 µM,
kcat/Km = 1.3 × 106 s
1 M
1).
Finally, all substitutions at the 145-position inactivated the purified
protein. The mutants K145A, K145R, K145H, and K145C, No Cys purified by
the 6-His tag/nickel affinity chromatography method showed no
detectable activity (Table I).
|
We were able to restore partial activity to an
inactive K145C, No Cys mutant of leader peptidase by reacting it with
the reagent 2BEA (Fig. 1A). Fig.
1A shows significant processing of the pro-OmpA nuclease A
substrate at 1- and 10-fold dilutions of the 2BEA-modified leader
peptidase. This chemically modified K145C, No Cys mutant has a
kcat value that is approximately 100-fold lower
than that of the wild-type leader peptidase (see Table I). The control experiment in which we have used the No Cys mutant, containing the
native lysine at the 145-position, showed no change in the activity
upon incubation with the 2BEA (Fig. 1B).
Modification of the K145C, No Cys mutant with the reagents 3BPA and
2MEA also restored activity to this inactive mutant, although to a
slightly lower extent (Fig. 2, A-B, and
Table I). The latter recovery in activity by 2-mercaptoethylamine is
most likely due to formation of a lysine analog at cysteine 145 as
reaction of the modified leader peptidase with a large excess of
-mercaptoethanol resulted in the disappearance of the restored
activity (data not shown). The 2MEA reaction goes nearly to completion
(see Table II). The disulfide bond formation between the
reagent and the free cysteine would most likely occur during the
dialysis step when the reagent would not be in great excess (see
"Experimental Procedures").
|
Modification of the K145C, No Cys leader peptidase mutant with the
reagent (2-bromoethyl)trimethylammonium·Br to form the non-titratable
quaternary amine lysine analog (4-thialaminine) at the 145-position
showed no detectable restoration of activity (Fig. 3).
No recovery in activity of the K145C, No Cys mutant is seen even after
treating the protein for 24 h with 2BETMA. In contrast, recovery
is seen after 2 h of treatment with 2BEA. This is consistent with
lysine 145 being involved in a critical hydrogen bond or serving as a
general base in the catalysis.
To assess the extent of reaction between the K145C, No Cys mutant of
leader peptidase with each of the above reagents, we used DTNB to
quantitate the number of cysteines with and without the modification
reaction. We have found that all of these reactions went nearly to
completion (Table II). The extent of modification by 2BEA was also
directly measured by amino acid compositional analysis using the
standard (S)-2-aminoethyl-L-cysteine·HCl
(Sigma). We saw the appearance of 1.1 residues of aminoethylated
cysteine (-thia-lysine) upon reaction with 2BEA.
The pH-rate profile of leader peptidase
using the pre-protein substrate pro-OmpA nuclease A measured by
plotting its kcat/Km versus pH, reveals a classic bell-shaped curve having a
maximum activity at approximately pH 9.0 (Fig.
4A). The dependence of kcat/Km on pH is consistent
with two ionizable catalytic residues within the free enzyme with
apparent pKa values of 8.7 and 9.3. Leader peptidase
appears to be stable up until pH 11 (Fig. 4B).
Inhibition of Leader Peptidase with Maleic Anhydride
We have
found that maleic anhydride inhibits leader peptidase in a
concentration-dependent (Fig. 5A)
and time-dependent (Fig. 5B) manner. The
addition of 0.5 mM maleic anhydride to leader peptidase
rapidly inactivated the enzyme with 80% loss in activity in 10 min
(Fig. 5B). To investigate whether this inhibition with maleic anhydride was due to the modification of lysine 145, we reacted
the K145C, No Cys mutant with maleic anhydride to modify all accessible
lysines. We then attempted to react the cysteine at the 145-position
with 2BEA to restore activity and also produce a single accessible
amine that then could be modified by maleic anhydride with subsequent
inhibition of activity. Our preliminary experiments had shown that
after treatment with maleic anhydride the activity of K145C, No Cys
leader peptidase was no longer recoverable (data not shown). Maleic
anhydride is known to react with thiol groups as well as amino groups
(46). Therefore we protected the cysteine by reacting it with
DTNB to form the 2-nitro-5-thiobenzoic acid-protected cysteine
before reacting it with maleic anhydride. We were then able,
after deprotection with -mercaptoethanol, to recover the activity of
the K145C, No Cys leader peptidase by aminoethylation and then finally
inhibit this recovered activity with maleic anhydride (Fig.
5C). The recovered activity from the DTNB/maleic
anhydride/
-mercatoethanol/2BEA reaction was approximately 33% of
that seen from the 2BEA reaction alone (data not shown). Maleic
anhydride is so far the only lysine-specific reagent we have found that
inhibits leader peptidase to a significant extent.
Sequence Alignment of Leader Peptidase with UmuD
As a first step toward
obtaining an idea of what the active site of leader peptidase may look
like, we have modeled the active site region of E. coli
leader peptidase based on the x-ray crystal structure of E. coli UmuD (14). Previously, van Dijl and colleagues (47) have
shown that these proteins are structurally and functionally related.
The residue anchoring feature of the alignment program XALIGN proved to
be particularly important in the alignment of UmuD
and leader
peptidase. High scoring pairwise alignments were only achieved when the
two catalytic residues of the Ser/Lys dyad in each of the proteins were
constrained to match. There is a 23.4% sequence identity and 37.2%
sequence similarity (conservative substitutions) between the UmuD
(residues 40-139) and leader peptidase (residues 75-202) (Fig.
6A). The regions of highest homology surround
the putative catalytic residues (Ser-90/Lys-145 in leader peptidase and
Ser-60/Lys-97 in UmuD
). There is a 31.1% identity and 53.3% sequence
similarity (conservative substitutions) between UmuD
and leader
peptidase when comparing these aligned sequences between Pro-48 and
Lys-98 (in UmuD
). Prior studies comparing the sequences of UmuD and
Bacillus subtilis signal peptidase (SipS) revealed a 25%
identity and 42% similarity between these proteins (47). A model of
the leader peptidase active site was built by replacing the residues
within the UmuD
structure with the corresponding aligned residue from
leader peptidase. The model of the leader peptidase active site reveals
a shallow hydrophobic cleft just adjacent to the catalytic site. This
hydrophobic cleft involves Ile-144, Leu-95, Phe-192, Ile-101, Ile-86,
Val-103, Val-131, and Phe-84 (Fig. 6C). Similar to UmuD
,
the proposed catalytic serine and lysine residues appear to be buried
within a hydrophobic environment (Fig. 6, B and
C).
In prior work, we have shown that substituting lysine at the 145-position of leader peptidase with alanine, asparagine, or histidine resulted in an inactive enzyme, demonstrating that this residue was important for activity (10). To further investigate the role of lysine 145 we have purified, using the 6-His tag/nickel affinity chromatography method (27), a variety of mutants with substitutions at the 145-position and measured the activity using an assay that has a sensitivity of 100,000-fold over background. It is interesting that the mutant K145R, which maintains the positive charge at this residue, was found to show no detectable activity even when 100 µg of enzyme was added to the reaction (data not shown). Replacing the lysine 145 with a histidine also resulted in an inactive enzyme. This is consistent with our earlier results (10) and implies that if the intriguing exchange of lysine for histidine is significant, in going from prokaryotic to eukaryotic signal peptidases (11, 12), there are certainly other important structural differences as well.
It is striking that in another inactive mutant, K145C, No Cys, there is
significant restoration of activity upon reaction with
2-bromoethylamine·HBr. This is a well characterized chemical modification reaction and, when combined with site-directed
mutagenesis, is a convenient way to introduce an unnatural amino acid
into a specific position within an enzyme (18, 28-36). Converting cysteines to aminoethylcysteine (-thia-lysine) residues has also been used in the past to create new trypsin-susceptible sites in
proteins (35, 36). The fact that we did not observe total restoration
of activity (approximately 1% of the wild-type activity) is somewhat
surprising since the side chain of the
-thia-lysine is only subtly
different from that of lysine. The only difference in the proteins
would be the substitution of one thioether group for a methylene group.
The bond length of C-C is 1.54 Å, whereas that of C-S is 1.82 Å, and
the bond angle of C-C-C is 109°, whereas that of C-S-C is 105°
(18). Model building studies indicate that the primary amine group in
-thia-lysine and lysine can be superimposed to within 0.1 Å (30).
To date, there has not been a reported crystal structure of a protein
containing a
-thia-lysine. Although there has been a report of a
small molecular structure of the amino acid itself, the authors (48)
state that due to unexpected bond length and torsion angles in this
structure it would be difficult to predict the C2 to C5 separation
distance of a
-thia-lysine in a macromolecule from this
structure.
Other studies that have utilized aminoethylation to restore activity to an inactive lysine to cysteine mutant include ribonuclease A (30), aspartate aminotransferase (18), and ribulosebisphosphate carboxylase/oxygenase (28). These studies have also reported incomplete restoration of activity by aminoethylation (8, 7, and 60%, respectively). The fact that the functionally restored leader peptidase K145C-EA, No Cys (lysine 145 changed to aminoethyl cysteine) shows a reduced kcat value, and yet a normal Km value (as compared with the No Cys mutant, Table I) is consistent with the lysine 145 being directly involved in catalysis. In these studies we have removed the other cysteines in leader peptidase to simplify the interpretation of the aminoethylation results. The cysteine-less (No Cys) leader peptidase had almost wild-type activity (Table I).
As a control, we confirmed that the 2BEA reagent had no effect on the activity of the No Cys mutant (Fig. 1B.) showing that the restoration of the activity upon the reaction with 2BEA is from the aminoethylation of the cysteine 145 residue. If the restoration of the activity in the K145C, No Cys mutant were due to modification of some other residue other than the cysteine 145, the addition of 2BEA to the No Cys mutant would have increased the activity in this enzyme as well. The inactive K145H mutant also showed no recovery of activity when treated with 2BEA (data not shown).
To assess length requirements of the side chain at the the
145-position, we reacted the K145C, No Cys mutant with
3-bromopropylamine·HBr and 2-mercaptoethylamine that generates a side
chain amine that is 5 atoms from the C instead of 4 atoms like lysine. These modifications also restored activity to the
K145C, No Cys mutant, although slightly less than that seen with
2-bromoethylamine (Fig. 2, A and B, and Table I),
revealing that there may be significant flexibility in the active site.
It is not clear why the conservative substitution of
-thia-lysine
for lysine only gives approximately 1% the activity of the wild type,
yet the longer residues
-thia-homo-lysine and
-dithio-homo-lysine
give almost the same activity as
-thia-lysine. One possible reason
why leader peptidase with
-thia-lysine is not fully active is that
-thia-lysine prefers a gauche (C-S-C) side chain torsion angle,
whereas lysine prefers an anti-(C-C-C) side chain torsion angle (30,
49).
Reacting the K145C, No Cys mutant with the reagent 2BETMA, which puts a
positively charged, nontitratable lysine analog (4-thialaminine) at the
145-position, resulted in no restoration of activity (Fig. 3). This is
consistent with the K145R result and also with the hypothesis that
lysine 145 is essential due to a critical hydrogen bond or its role as
a general base and not because of its charge. It is possible that the
additional bulkiness of the methyl groups on the -amino group within
the lysine analog (4-thialaminine) is responsible for the lack of
activity.
The pH-rate profile that shows a bell-shaped curve when the kcat/Km is plotted versus pH is consistent with two ionizable groups in the free enzyme with apparent pKa values of approximately 8.7 and 9.3 (Fig. 4A). It is likely that the apparent pKa value of 8.7 from the ascending limb of the plot corresponds to the pKa of lysine 145 that would be required to be deprotonated to function as a general base. Lysine 145 would have a pKa value 1.8 units lower than a free lysine in solution (pKa = 10.5). Leader peptidase is not irreversibly affected by pH changes until pH 11 (Fig. 4B); therefore, denaturation of the protein at pH values higher than about pH 9 would not be an explanation for the descending arm of the pH profile (pKa = 9.3). It could be that a protonated residue with a pKa of 9.3 is essential for leader peptidase activity, such as a residue involved in an oxyanion hole might be deprotonated at this pH. It is also possible that this pKa reflects a deprotonation of a residue involved in a salt bridge important for active site geometry. The descending arm of the pH-rate profile might correspond to the deprotonation of a tyrosine residue functioning as a critical hydrogen bond donor for binding or for catalysis. There is a conserved tyrosine 143 (Fig. 6, B and C), just two residues from the critical lysine, but this residue has been changed to a phenylalanine by Black et al. (9) and found to be nonessential using a very sensitive in vivo assay. van Dijl et al. (47) also found this tyrosine to be nonessential in B. subtilis SipS.
There is a strictly conserved arginine directly carboxyl-terminal to the proposed catalytic lysine. It is well known that spatially proximal positive charges to a lysine can decrease its pKa. Westheimer's hypothesis (50) concerning such an effect has recently been confirmed by Highbarger et al. (19). To investigate the possible importance of the immediately proximal positively charged arginine at position 146, this residue was mutated to an alanine. The less than dramatic decrease in activity resulting from the R146A mutation indicates that it is not the positive charge from Arg-146 that is responsible for the decreased pKa of Lys-145 (see Table I). Moreover, when we look at the modeled active site of leader peptidase (Fig. 6B), we see that the charged guanidinium group of Arg-146 is pointed completely away from the active site region and therefore would be predicted to have a very small electrostatic effect on Lys-145. Therefore, it is more likely that the decreased pKa of Lys-145 comes from its hydrophobic environment (Fig. 6, B and C). It is important that in future studies the apparent pKa values acquired from the pH-rate profile be assigned to specific residues by other methods such as NMR spectroscopy (51).
We have found that leader peptidase is inhibited by the lysine modifying reagent maleic anhydride (Fig. 5). Kim et al. (52) found only low level inhibition of leader peptidase upon the addition of the lysine modifiers succinic anhydride and trinitrobenzene sulfonic acid. It is not yet clear why only maleic anhydride was successful in modifying lysine 145. Other lysine-modifying reagents we have tried without any effect on activity include potassium cyanate, trinitrobenzene sulfonic acid, succinic anhydride, acetic anhydride, and pyridoxyl 5-phosphate. It may be that these other reagents were not able to gain access to the active site region due to the active site's proposed hydrophobic environment. Another possibility is that maleic anhydride is the only reagent tried so far that can bind and induce a conformation within leader peptidase that allows lysine 145 to be reactive (lowered lysine pKa). Finally, lysine 145 may be less reactive to modification by lysine-specific reagents because of its possible hydrogen bond to serine 90 and the potentially buried nature of Lys-145.
Taken together, the results from this study as well as others (9, 47)
are consistent with the hypothesis that leader peptidase utilizes a
lysine as its general base. We have proposed a mechanism for leader
peptidase whereby the deprotonated -amine of lysine 145 abstracts
the proton from the O
of serine 90 making it nucleophilic enough to
attack the scissile bond of the pre-protein substrate (Fig.
7A). The tetrahedral intermediate I could be
stabilized by a yet unidentified oxyanion hole, and the breakdown of
the tetrahedral intermediate I to form the acyl-enzyme intermediate would be accelerated by the protonation of the leaving amine (mature protein) via lysine 145 (Fig. 7B). It is possible that the
lysine 145 could also act as the general base in the formation of
tetrahedral intermediate II, whereby lysine 145 would activate a water
that would attack the ester carbonyl of the acyl-enzyme intermediate (Fig. 7C).
There are many unanswered questions regarding the mechanism of leader (signal) peptidase. From site-directed mutagenesis and chemical modification studies, it is clear that serine 90 and lysine 145 are critical residues, but it is not yet clear whether there exists an oxyanion hole similar to the classical serine proteases. Another unanswered question is whether lysine 145 serves as the general base in both activation steps of the reaction. The forthcoming x-ray crystal structure of the soluble fragment of E. coli leader peptidase (53) may help to answer these questions regarding the mechanism of this very unique serine protease.
We gratefully thank Michael Dunne of Chemical Abstract Services for advice on nomenclature regarding the modified cysteine residues. We thank Dr. Michael N. G. James of the University of Alberta for the use of the computer facilities that made the modeling studies possible.