JOURNAL OF BACTERIOLOGY, Dec. 2007, p. 8835–8843
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 189, No. 24
Full Activation of Enterococcus faecalis Gelatinase by a C-Terminal
Maria Florencia Del Papa,1§¶ Lynn E. Hancock,1,2§ Vinai C. Thomas,2and Marta Perego1*
Division of Cellular Biology, Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla,
California 92037,1and Division of Biology, Kansas State University, Manhattan, Kansas 66506-49012
Received 13 August 2007/Accepted 27 September 2007
Enterococci account for nearly 10% of all nosocomial infections and constitute a significant treatment
challenge due to their multidrug resistance properties. One of the well-studied virulence factors of
Enterococcus faecalis is a secreted bacterial protease, termed gelatinase, which has been shown to con-
tribute to the process of biofilm formation. Gelatinase belongs to the M4 family of bacterial zinc
metalloendopeptidases, typified by thermolysin. Gelatinase is synthesized as a preproenzyme consisting of
a signal sequence, a putative propeptide, and then the mature enzyme. We determined that the molecular
mass of the mature protein isolated from culture supernatant was 33,030 Da, which differed from the
predicted molecular mass, 34,570 Da, by over 1,500 Da. Using N-terminal sequencing, we confirmed that
the mature protein begins at the previously identified sequence VGSEV, thus suggesting that the 1,500-Da
molecular mass difference resulted from a C-terminal processing event. By using mutants with site-
directed mutations within a predicted C-terminal processing site and mutants with C-terminal deletions
fused to a hexahistidine tag, we determined that the processing site is likely to be between residues D304
and I305 and that it requires the Q306 residue. The results suggest that the E. faecalis gelatinase requires
C-terminal processing for full activation of protease activity, making it a unique enzyme among the
members of the M4 family of proteases of gram-positive bacteria.
Enterococcus faecalis is a gram-positive coccus that usually
occurs in pairs or short chains and is commonly found as part
of the resident flora in the mammalian intestinal tract (37). In
contrast to the beneficial role that enterococci play in intestinal
homeostasis, these organisms are becoming increasingly im-
portant to human health as leading causes of nosocomial in-
fections. These infections include urinary tract and abdominal
infections, bacteremia, endocarditis, and wound infections (1,
9, 29). E. faecalis is ubiquitous in the environment and can
withstand high salt concentrations and wide pH and tempera-
ture ranges. Enterococci are also resistant to desiccation and
temperatures up to 60°C. One of the most important public
health aspects of enterococci is their increasingly wide range of
antibiotic resistance (14). Multidrug resistance is common
among clinical isolates, leaving few therapeutic options for
treating enterococcal infections.
Some strains of E. faecalis can form biofilms, which may
increase their ability to colonize patients and persist at infec-
tion sites (22). Biofilms are bacterial communities growing as
surface-attached aggregates encased in an exopolymer matrix.
Biofilms are generally thought to be more resistant to antibi-
otics than the corresponding free-living bacteria, which com-
pounds the problem of antimicrobial resistance. Moreover,
recent estimates suggest that over 65% of hospital-acquired
infections stem from the ability of the infecting organism to
produce biofilms (17).
As biofilm formation is thought to be an environmentally
responsive process, we recently examined the contributions of
two-component signal transduction networks to biofilm forma-
tion by E. faecalis V583 (11). We found that of the 17 two-
component response regulator mutants tested, only the fsrA
response regulator mutant was significantly attenuated in bio-
film formation. FsrA is part of a signal transduction system that
is responsive to the accumulation of an 11-amino-acid peptide
lactone, termed gelatinase biosynthesis-activating pheromone
(24). This peptide is encoded by the fsrD gene and is thought
to be proteolytically processed from the FsrD protein and
exported by the fsrB gene product (25), and its presence in a
culture is sensed by the FsrC histidine kinase. Gelatinase bio-
synthesis-activating pheromone sensing by the FsrC sensor ki-
nase results in activation of the FsrA response regulator tran-
scription factor by a phosphoryl group. The phosphorylated
form of FsrA is believed to activate expression of two cotrans-
cribed protease genes, gelE and sprE (28). Additionally, the
FsrABC system was shown by microarray analysis to affect a
variety of other genes, some of which potentially are involved
in virulence and some of which are involved in metabolic
pathways (2). As the GelE and SprE proteases are Fsr-regu-
lated proteins, we explored the contribution of the E. faecalis
proteases gelatinase and serine protease to biofilm formation
(12). We found that the biofilm deficiency of an fsrA mutant
was due to an inability to make gelatinase. Complementation
* Corresponding author. Mailing address: Division of Cellular Biol-
ogy, Mail Code MEM-116, Department of Molecular and Experimen-
tal Medicine, The Scripps Research Institute, 10550 North Torrey
Pines Road, La Jolla, CA 92037. Phone: (858) 784-7912. Fax: (858)
784-7966. E-mail: firstname.lastname@example.org.
† Supplemental material for this article may be found at http://jb
‡ Manuscript 19055 from The Scripps Research Institute.
§ M.F.D.P. and L.E.H. contributed equally to this study.
¶ Present address: Instituto de Bioquı ´mica y Biologı ´a Molecular,
Facultad de Ciencias Exactas, Universidad Nacional de La Plata,
Calles 47 y 115 (1900), La Plata, BsAS, Argentina.
?Published ahead of print on 5 October 2007.
of the fsrA mutant by gelatinase expressed from an fsr-inde-
pendent promoter restored biofilm formation. Addition of pu-
rified gelatinase also restored biofilm formation to fsr-defective
strains, suggesting a critical, but as-yet-unknown, role for gel-
atinase in biofilm formation by E. faecalis. The involvement of
GelE in biofilm formation was also reported by Kristich et al.
GelE was originally characterized by Ma ¨kinen et al. (18) as
an extracellular Zn metalloprotease with activity against a
number of substrates, including the insulin B chain, azo dye-
impregnated collagen (azocoll), and the pheromones and in-
hibitor peptides involved in conjugative plasmid transfer in E.
faecalis. GelE was also shown to function in clearing the bac-
terial cell surface of misfold proteins and in activating an
autolysin (41). Gelatinase is synthesized as a 509-amino-acid
prepropolypeptide which is subject to cleavage of the 192
amino acids at the amino-terminal end comprising the prese-
quence or signal sequence and the prosequence.
In this study, an apparent difference between the predicted
and actual molecular masses of the purified gelatinase revealed
that gelatinase undergoes C-terminal proteolytic processing
for full maturation. The role of this processing in gelatinase
activity was examined.
MATERIALS AND METHODS
Bacterial strains, plasmids, and growth conditions. E. faecalis strains and
plasmids used in this study are listed in Tables 1 and 2. Oligonucleotide primers
used for plasmid construction are shown in Table S1 in the supplemental mate-
rial. Plasmid pML28 was constructed from the vector pAT28 (39) by inserting a
369-bp EcoRI-BamHI fragment carrying the promoter of the aphA-3 kanamycin
resistance gene (40). The gelE coding sequence and its ribosome binding site
were then cloned in pML28 as a BamHI fragment to obtain pML29 (12). For
construction of pML33, the mature gelE coding sequence was first cloned as an
NdeI-XhoI fragment in the pET21 expression vector (Novagen) using oligonu-
cleotide primers MGelNT-NdeI and MGelCT-XhoI, yielding plasmid pML52,
which encoded a GelE protein with a six-His tag extension. A fragment of this
plasmid was generated by PCR amplification using oligonucleotide primers
MGelNT-Nde and pET20downPst (the latter primer hybridizes to the pET21
vector in a region downstream of the six histidine codons). This fragment was
digested with NcoI, which naturally occurs in the gelE sequence, and cloned in
the pML29 vector digested with NcoI and SalI (the latter enzyme blunted with
Klenow polymerase) to replace the wild-type 3? end of the gene with the histidine
tag-modified version. Plasmid pML34 was constructed by cloning the PCR am-
plification product obtained with oligonucleotide primers GelE5? and GelECT-
14BamHI, digested with BamHI, in the BamHI site of plasmid pML28. Plasmid
pML37 was derived from pML29 in which the 3? end of the wild-type gene was
replaced by the 3? end of a gelE gene missing the last 14 amino acids but fused
to six histidine codons. This construct was first generated in the pET21 expres-
sion vector (Novagen) via BamHI-XhoI cloning of a PCR fragment amplified
with oligonucleotides MGelENT-NdeI and GelECT-14XhoI, yielding plasmid
pML56. Plasmid pML56 was used as a template for PCR amplification with
TABLE 1. E. faecalis strains used in this study
StrainRelevant genotype Relevant phenotype Origin
fsrC aphA-3 promoter
fsrC aphA-3 promoter-gelE
fsrC aphA-3 promoter-gelE (His tag)
fsrC aphA-3 promoter-gelE (C-terminal 14 amino acids)
fsrC aphA-3 promoter-gelE (C-terminal 14 amino acids, His tag)
fsrC aphA-3 promoter-gelE (C-terminal 13 amino acids, His tag)
fsrC aphA-3 promoter-gelE (C-terminal 13 amino acids, His tag)
fsrC aphA-3 promoter-gelE(E303P)
fsrC aphA-3 promoter-gelE(D304P)
fsrC aphA-3 promoter-gelE(I305P)
fsrC aphA-3 promoter-gelE(Q306P)
fsrC aphA-3 promoter-gelE(E137Q)
fsrC aphA-3 promoter-gelE(E137Q) (His tag)
pML28 3 FA2-2
pML29 3 FA2-2
pML33 3 FA2-2
pML34 3 FA2-2
pML37 3 FA2-2
pML42 3 FA2-2
pML43 3 FA2-2
pML38 3 FA2-2
pML39 3 FA2-2
pML40 3 FA2-2
pML41 3 FA2-2
pML44 3 FA2-2
pML45 3 FA2-2
TABLE 2. Plasmids used in this study
VectorDescriptionPredicted gelatinase C-terminal sequence
Shuttle vector with aphA-3 promoter
GelE plus His6tag
GelE without C-terminal 14 amino acids
GelE without C-terminal 14 amino acids plus His tag
GelE without C-terminal 13 amino acids plus His tag
GelE without C-terminal 12 amino acids plus His tag
GelE with glutamic acid-to-proline substitution at position 303
GelE with aspartic acid-to-proline substitution at position 304
GelE with isoleucine-to-proline substitution at position 305
GelE with glutamine-to-proline substitution at position 306
GelE with glutamate at position 137 mutated to Gln
GelE with glutamate at position 137 mutated to Gln plus His6tag
aDerivative of pAT28 (39) carrying the aphA-3 promoter as a 369-bp EcoRI-BamHI fragment (12).
bSee reference 12.
8836DEL PAPA ET AL.J. BACTERIOL.
oligonucleotide primers NGelENT-NdeI and pET20downPst. The resulting frag-
ment was digested with NcoI and ligated to pML29 digested with XbaI (treated
with Klenow polymerase to obtain a blunt end) and NcoI. Plasmids pML42 and
pML43 were constructed by cloning PCR-amplified fragments obtained with
oligonucleotide primers MGelNT-NdeI and GelEIHis2 and with oligonucleotide
primers MGelNT-NdeI and GelEQHis2, respectively, digested with NcoI and
XbaI in similarly digested pML29. PCR products were obtained by amplification
of E. faecalis V583 genomic DNA.
Strains were cultured in Todd-Hewitt broth (THB) or M17 medium (Difco
Laboratories). Escherichia coli XL10 Gold (Stratagene) was used for plasmid
construction and propagation. Strains were cultivated in Luria-Bertani broth.
The antibiotics used for selection in E. coli and E. faecalis were spectinomycin
(150 and 750 ?g/ml, respectively) and tetracycline (15 ?g/ml). Screening for
gelatinase production was carried out on THB agar plates containing 1.5% skim
milk. Electroporation was carried out as described previously (5).
Site-directed mutagenesis. Plasmid pML29 was subjected to site-directed mu-
tagenesis using a QuickChange II site-directed mutagenesis kit (Stratagene, La
Jolla, CA) and mutagenic primers to replace residues surrounding the putative
C-terminal processing site with proline or the active site E137 with glutamine
(see Table S1 in the supplemental material). Plasmid pML33 was used as tem-
plate to generate pML45.
Gelatinase purification. For all GelE proteins, most of the purification steps
were performed at 4°C; the only exception was hydrophobic interaction chro-
matography, which was performed at room temperature. The procedure used
was the procedure described by Hancock and Perego (12).
Protein determination, electrophoretic techniques, and Western blotting. The
protein concentration was determined with a bicinchoninic acid reagent kit
(Pierce), using bovine serum albumin as a standard. Enzyme purification was
monitored by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis
(PAGE) on a 12% polyacrylamide separating gel. Concentrated (20?) culture
supernatants from overnight cultures were run on SDS-PAGE gels and stained
with Coomassie blue or transferred to a polyvinylidene difluoride membrane.
GelE was detected with anti-His (C-terminal) monoclonal antibody (Novagen)
using an ECL Western blotting detection kit (Amersham Biosciences).
Protease activity assay. Azocoll (18) was suspended at a final concentration of
5 mg/ml in the assay buffer (50 mM Tris [pH 7.8], 1 mM CaCl2) and incubated
for 2 h at 37°C with vigorous shaking. The solution was filtered (Whatman 1MM
paper), and the precipitate was resuspended immediately in the same volume of
fresh buffer. Stirring was rapid enough to obtain what appeared to be a uniform
suspension. Tubes (diameter, 13 mm) were prewarmed at 37°C for 15 min before
the reaction was initiated by addition of enzyme. Assays were stopped by im-
mersing tubes in an ice-water bath. Chilled tubes were then centrifuged, and the
optical densities at 550 nm of supernatant fractions that were diluted 1:2 were
measured. Data were obtained for every time point in triplicate.
Molecular mass determination. The molecular masses of the native protease
and mutants of this protease were determined using matrix-assisted laser
desorption ionization—time of flight mass spectrometry. The analysis was
performed at the Mass Spectroscopy Core Facility at The Scripps Research
Biofilm formation. Biofilm formation on polystyrene was quantified as previ-
ously described (38). Briefly, E. faecalis strains were grown overnight at 37°C in
M17 medium supplemented with 0.5% lactose and spectinomycin (500 ?g/ml)
when appropriate. Each culture was diluted 1:100 in M17 medium, and 200 ?l of
the resulting cell suspension was used to inoculate sterile 96-well flat-bottom
polystyrene microtiter plates. After incubation for 24 h at 37°C, wells were gently
washed three times with 200 ?l of phosphate-buffered saline, dried in the in-
verted position, and stained with 1% crystal violet for 15 min. The wells were
rinsed again, and the crystal violet was solubilized in 200 ?l ethanol-acetone
(80:20, vol/vol). The optical density at 550 nm was determined using a microplate
reader. Each assay was performed in triplicate and repeated three times.
Autolysis assay. E. faecalis FA2-2 strains containing pML28, pML29, pML34,
and pML41 were grown overnight at 37°C in 2.5 ml THB containing 600 ?g/ml
of spectinomycin. Twenty-five microliters of each overnight culture was inocu-
lated into 2.5 ml SM17 medium containing spectinomycin and 3% glycine (the
appropriate final concentration of glycine in the medium was obtained using a
filter-sterilized 25% glycine stock solution) and grown overnight at 37°C. Then
1.5 ml of the overnight culture was centrifuged at 13,000 rpm for 3 min. The
supernatants were discarded, and the pellets were washed three times in ice-cold
sterile distilled water. After the third wash, the pellets were resuspended in 10
mM sodium phosphate buffer (pH 6.8). Two hundred microliters of the sus-
pended cells was dispensed into a 96-well plate, and the optical density at 600 nm
at 37°C was determined for 9 h at 30-min intervals. A statistical analysis (Stu-
dent’s t test) was carried out with the SAS program.
Molecular mass of mature gelatinase. The E. faecalis pro-
tease gelatinase belongs to the M4 family of bacterial zinc
metallopeptidases typified by thermolysin of Bacillus thermo-
proteolyticus (27) and aureolysin of Staphylococcus aureus (30).
An amino acid sequence alignment of the mature portions of
thermolysin, aureolysin, and gelatinase (Fig. 1) revealed that
gelatinase had an 18-amino-acid extension at the carboxy-ter-
minal end. Some M4 family members from bacteria belonging
to the family Vibrionaceae require processing at the carboxy
terminus, in addition to two processing events at the amino
terminus, to become fully active (15, 23). For this reason we
investigated the possibility that the enterococcal gelatinase un-
dergoes a carboxy-terminal processing event by determining
the molecular weight of the mature enzyme purified from the
culture supernatant. As shown in Fig. S1 in the supplemental
material, mass spectrometry analysis of purified gelatinase re-
vealed a molecular mass of 33,030 Da, which was 1,540 Da less
than the predicted molecular mass of the protein (34,570 Da)
based on its deduced amino acid sequence. N-terminal se-
quencing revealed that the purified, mature protein began with
the previously determined sequence “VGSEV” (36) (data not
shown), suggesting that the molecular mass difference had to
result from a C-terminal processing event(s) that, based on the
expected molecular masses of proteins lacking the last 12 to 15
amino acids, is consistent with cleavage of the last 14 amino
acids from the carboxy-terminal end (Fig. 2).
Gelatinase is processed at the C-terminal end. In order to
determine the site of processing of the gelatinase enzyme, we
expressed the wild-type protein and C-terminal deletion mu-
tants from an FsrA-independent promoter (aphA-3 ) on a
replicative vector in the FA2-2 strain, which is gelatinase de-
ficient due to the absence of a functional fsr system (12) (Ta-
Complementation of the gelatinase-deficient phenotype was
first tested on THB-milk agar plates (data not shown) in order
to ensure that an active gelatinase was expressed. Then 20?-
concentrated culture supernatants were analyzed for the pres-
ence of gelatinase protein by SDS-PAGE and Western blot-
ting. As shown in Fig. 3A, while strain FA2-2 carrying the
pML28 vector, which contained only the aphA-3 promoter, did
not produce gelatinase, the strain carrying plasmid pML29,
which expressed the wild-type gelE gene, secreted a protein
with a molecular mass of approximately 34 kDa. A protein of
similar size was produced by strains carrying plasmids pML33
and pML34, which expressed a wild-type GelE protein with a
six-histidine extension at the C-terminal end and a GelE pro-
tein lacking the C-terminal 14 amino acids, respectively. The
culture supernatant of the strain carrying plasmid pML37 (ex-
pressing a GelE protein lacking the C-terminal 14 amino acids
but with a six-His tag fused to the D304 residue via a Leu-Glu
linker) produced a protein band that was the same size as the
band observed in the samples described above and an addi-
tional band at a slightly higher molecular weight. A similar
band pattern was obtained with the culture supernatants of the
strains carrying plasmids pML42 and pML43, which expressed
GelE fused to the six-His tag at the I305 and Q306 residues,
respectively. Notably, the concentration of the higher-molecu-
lar-weight band relative to the lower-molecular-weight band
VOL. 189, 2007 GELATINASE C-TERMINAL PROCESSING8837
decreased in the order pML37 ? pML42 ? pML43. Analysis
of the same supernatants by Western blotting using an anti-C-
terminal His tag antibody revealed that only the truncated
GelE proteins expressed by plasmids pML37, pML42, and
pML43 maintained the histidine tag fused to the D304, I305,
and Q306 residues, respectively (Fig. 3B). The intensities of
the bands detected corresponded to the intensities of the
higher-molecular-weight bands observed in the Coomassie
blue-stained gel shown in Fig. 3A. (i.e., pML37 ? pML42 ?
pML43). Since no His tag was detected in the supernatant of
the strain expressing the GelE wild-type protein with the his-
tidine tag fused to the last residue (E318; pML33), these re-
sults indicated that maturation of GelE occurred via cleavage
at the C-terminal end and that this process was progressively
delayed by deletion of the last 12 to 14 amino acids. Never-
theless, incubation of the same proteins for 12 h at 37°C re-
sulted in the presence of only one form of GelE, corresponding
to the mature form, in all preparations (Fig. 3C). This indi-
cated that maturation could eventually occur even when the
C-terminal 14 amino acids were replaced by the L-E-H6amino
Proline substitution at Q306 affects gelatinase activity. Mass
spectrometry determination of the molecular mass of mature
wild-type gelatinase from culture supernatant revealed a dis-
crepancy of approximately 1,500 Da with the predicted value
(Fig. 2). Because the predicted molecular mass of the last 14
amino acids at the carboxy-terminal end of gelatinase is 1,556
Da (Fig. 2), we investigated whether any of the residues at
positions 16 (E303), 15 (D304), 14 (I305), and 13 (Q306) from
the C-terminal end had a critical role in the processing of the
FIG. 1. Amino acid sequence alignment of members of the M4 family of bacterial metalloproteases. The amino acid sequences of aureolysin
of S. aureus, thermolysin of B. thermoproteolyticus, and gelatinase of E. faecalis were aligned using the ClustalW program. The active site residues
are indicated by the gray boxes. Symbols: asterisks, identical residues in all sequences; colons, conserved substitutions; periods, semiconserved
substitutions. The sequence of the C-terminal 18 amino acids is shown at the bottom along with the calculated molecular masses of GelE mutant
proteins lacking the last 5 (?5), 6 (?6), 12 (?12), 13 (?13), 14 (?14), and 15 (?15) amino acids. The predicted molecular masses (mw) were
obtained with the ExPASy pI-Mwtool. The predicted molecular masses of the GelE Q306P mutant proteins lacking the last five or six amino acids
are indicated in parentheses. D, daltons.
8838 DEL PAPA ET AL.J. BACTERIOL.
protease. These four residues were individually changed to
proline by site-directed mutagenesis, and the resulting proteins
were purified from culture supernatants of strain FA2-2 to-
gether with the wild-type form of GelE. SDS-PAGE analysis of
1 ?g of each protein revealed that while the GelE E303P,
D304P, and I305P mutant proteins had an electrophoretic mo-
bility equal to that of the wild-type protease, the mobility of the
Q306P mutant protein was slightly retarded (Fig. 4). A mass
spectrometry analysis carried out with the five proteins indi-
cated that the E303P and D304P proteins had molecular
masses in the range of the molecular mass obtained for the
wild-type protein, suggesting that these substitutions did not
affect the processing of the C-terminal end. The I305P mutant
protein seemed to be approximately one amino acid larger
than the wild-type GelE, while the molecular weight of the
Q306P protein indicated that five or six additional amino acids
were present compared to the wild-type protein (Fig. 4).
These results suggested that a proline at position 305 may
have a role in positioning the cleavage site but that Q306 is the
most critical residue for processing of the 14 C-terminal amino
acids. Replacement of Q306 with proline in fact resulted in an
alternative cleavage event that most likely deleted the C-ter-
minal five amino acids.
Activity of GelE mutants with synthetic substrates and in
biofilm formation. In order to compare the activities of the
four site-specific mutants of GelE (E303P, D304P, I305P, and
Q306P), the purified proteins were assayed in three indepen-
dent experiments using azocoll as a substrate. All proteinases
were still able to degrade azocoll, but the Q306P mutant pro-
tein was consistently less active than the other proteins (Fig. 5),
showing a 30% reduction in protease activity. The results sug-
gested that correct processing of the last 14 amino acids at the
C-terminal end of GelE is necessary for full activity of the
The previously described role of gelatinase in biofilm for-
mation prompted us to test whether the proline substitutions
had any effect on biofilm formation. As shown in Fig. S2 in the
supplemental material, none of the mutant proteins affected
FIG. 2. Schematic representation of gelatinase proteolytic process-
ing. The gelE gene encodes a 509-amino-acid protein with a probable
signal peptide (S.P.) that is 29 amino acids long. The propeptide
sequence extends from residue 30 to residue 191, and the mature
gelatinase starts at residue V192. The predicted masses of the mature
protein and C-terminal peptide were determined with the ExPASy
pI-Mwtool. The molecular weight of the mature protein purified from
culture supernatants was determined by mass spectrometry.
FIG. 3. Analysis of the protein profiles for the culture supernatants
of GelE derivatives. The supernatants were concentrated 20-fold and
loaded on a 12% polyacrylamide gel for Coomassie blue staining
(A) or for Western blot analysis using anti-C-terminal His tag antibody
(B). (C) Protein profile of the same supernatants after overnight in-
cubation al 37°C.
FIG. 4. Analysis of the electrophoretic mobility of GelE proline
substitution mutant proteins. Gelatinases purified from culture super-
natants of strain FA2-2 expressing the wild-type protein (GelE wt) and
the E303P, D304P, I305P, and Q306P mutant proteins were run on a
12% SDS-PAGE gel and stained with Coomassie blue. The broad
range protein marker (New England Biolabs) (lanes MW) was in-
cluded in the analysis. The molecular weights of the five proteins were
determined by matrix-assisted laser desorption ionization—time of
flight mass spectrometry. MW, molecular weight.
VOL. 189, 2007GELATINASE C-TERMINAL PROCESSING 8839
the ability of strain FA2-2 to form biofilms, but the sensitivity
of the assay may not have been high enough to detect the small
enzymatic differences between these proteins.
GelE Q306P mutation affects cell autolysis. GelE has been
shown to be involved in the maturation of the E. faecalis
muramidase-1 autolysin in vitro (33), and GelE-producing
strains have been shown to have increased autolysis (41). For
these reasons we investigated whether processing of the C-
terminal end of gelatinase had any effect on cell autolysis.
Strains derived from FA2-2 carrying plasmid pML28 (vector),
pML29 (GelE wild type), pML34 (GelE-CT14), or pML41
(GelE Q306P) were subjected to an autolysis assay as de-
scribed in Materials and Methods. The results (Fig. 6) consis-
tently and reproducibly indicated that the strain expressing the
GelE Q306P mutant protein had a slightly higher rate of au-
tolysis than the strain expressing the GelE wild-type protein
(P ? 0.05). Notably, the strain expressing the GelE protein
lacking the last 14 amino acids showed a slightly lower, al-
though not statistically significant, level of autolysis than the
wild-type protein-expressing strain (P ? 0.36). Consistent with
the results reported by Waters et al. (41), the strains expressing
GelE had a higher autocatalytic rate than the control strain not
expressing the enzyme.
These results suggest that although the Q306P mutation
reduced the protease activity of GelE, the inability to process
the C-terminal end increased the ability of the enzyme to
access or activate the target autolysin.
Active site mutation affects the GelE N- and C-terminal
processes. As a member of the M4 family of zinc metallopro-
teases, gelatinase possesses a conserved catalytic domain typ-
ified by the primary-sequence HEXXH motif responsible for
coordinating zinc in the active site (Fig. 1) (13). In order to
investigate the role of GelE activity in its maturation, the
glutamic acid residue at position 137 in the active site was
mutated to glutamine and the protein was expressed in strain
FA2-2 from the aphA-3 promoter of plasmid pML28 as de-
scribed above. The resulting construct, designated pML44, was
also modified to express an E137Q GelE mutant protein with
six His residues at the C-terminal end (plasmid pML45). The
concentrated supernatants of cultures expressing these pro-
teins were run on an SDS-PAGE gel and evaluated by Western
blotting (Fig. 7). The site-specific E137Q mutation led to com-
plete abolition of GelE activity as detected by protease activity
on THB-milk plates (data not shown). Electrophoretic analysis
of the culture supernatants revealed the disappearance of the
mature-size protein and the concomitant accumulation of an
approximately 55-kDa protein which was present at a lower
level than the wild-type enzyme (expressed by plasmid pML29)
(Fig. 7A), indicating that active GelE was necessary for post-
translational modification of this protein (i.e., cleavage of the
propeptide). Western blot analysis carried out with an anti-C-
terminal His tag antibody confirmed the presence of an ap-
proximately 55-kDa His-tagged protein in the supernatant of
the cells carrying the pML45 construct, while a His-tagged
protein that was the size of mature GelE was present, as
expected, in the supernatant of the strain carrying plasmid
pML37, as shown in Fig. 3. These results suggested that an
active GelE protein was required for the multiple maturation
processes at the N- and C-terminal ends involving an autocat-
alytic mechanism. Similar results were reported for thermoly-
sin (19) and the Streptomyces cacaoi extracellular neutral met-
alloprotease Npr (4). In contrast, mutations of the zinc-binding
metalloprotease motif of the LasA protease of Pseudomonas
aeruginosa or the BFT toxin of Bacteroides fragilis, which be-
long to different protease families, were shown to affect the
activity but not the propeptide processing (8, 10).
The availability of the active site mutant GelE proteins and
evidence of their expression and secretion in the culture su-
pernatant allowed us to test whether the E137Q substitution
affected biofilm formation. The FA2-2 strains carrying the
pML28 vector alone, the pML29 vector expressing wild-type
GelE, and the pML44 and pML45 plasmids expressing the
E137Q mutant protein with and without a six-His tag, respec-
tively, were assayed for biofilm formation as described in Ma-
terials and Methods. The results (Fig. 7C) indicated that while
expression of wild-type GelE induced biofilm growth, expres-
sion of the E137Q mutant proteins did not. This suggested that
the protease activity, rather than the GelE protein itself, is
required for biofilm formation.
The zinc metalloprotease secreted by E. faecalis, gelatinase,
is one of the few characterized virulence factors of this organ-
FIG. 6. Autolysis assay with FA2-2 strains expressing gelatinase.
The autolysis assay was carried out with FA2-2 strains carrying plas-
mids pML28 (vector alone) (F), pML29 (GelE wild type) (Œ), pML34
(GelE-CT14aa) (f), and pML41 (GelE Q306P) (?). Samples were
analyzed at 30-min intervals as described in Materials and Methods.
OD600, optical density at 550 nm; T0, time zero.
FIG. 5. Hydrolysis of azocoll by GelE and site-specific mutants as a
function of time measured by the absorbance of released azocoll dye.
The assay was carried out with FA2-2 strains carrying plasmids pML28
(vector alone) (?), pML29 (GelE wild type) (f), pML38 (GelE
E303P) (‚), pML39 (GelE D304P) (?), pML40 (GelE I305P) (?),
and pML41 (GelE Q306P) (F). The data are the averages of two
independent experiments; the error bars indicate the standard devia-
tions. Each data point is the average of triplicate measurements.
OD550, optical density at 550 nm.
8840 DEL PAPA ET AL.J. BACTERIOL.
ism. Gelatinase has been shown to contribute to disease patho-
genesis in a number of model systems, including peritonitis in
mice (34), endophthalmitis in rabbits (7), and nematode killing
(34). In addition to gelatin, only a few substrates are known for
this protease. The list of known substrates includes the entero-
coccal conjugative sex pheromones and a number of host-
derived in vitro substrates, including the insulin B chain, en-
dothelin, hemoglobin, fibrinogen, fibronectin, collagen, and
laminin (18). In addition, gelatinase has been shown to inter-
fere with innate immune defenses through inactivation of the
antibacterial peptides LL-37 and ?-defensin (31, 32). Recently,
gelatinase has been shown to be relevant for in vitro translo-
cation of E. faecalis across polarized human enterocyte-like
T84 cells (42).
As a member of the M4 family of zinc metalloproteases,
gelatinase possesses a conserved catalytic domain typified by
the primary-sequence HEXXH motif responsible for coordi-
nating zinc in the active site (13). In addition, these zinc met-
alloproteases bind three or four calcium atoms, which are
thought to stabilize the overall structure; the more thermal
stable proteases bind four calcium atoms.
Members of the M4 family are secreted bacterial enzymes
that are synthesized as preproenzymes. The presequence, or
signal peptide, targets the proenzyme to the secretion appara-
tus of either gram-negative or gram-positive cells. Maturation
of the secreted proenzyme to its active mature form is thought
to proceed by autocatalytic processing and transient associa-
tion of the propeptide or prosequence with the active form (19,
26). The primary purpose of the propeptide is to function as an
intramolecular chaperone facilitating the proper folding of the
active protease (20, 21).
Here we showed that an active gelatinase is required for
processing of the propeptide at the amino terminus and cleav-
age of the C-terminal tail. Mutation of a critical residue in the
GelE active site (E137) resulted in secretion of a protein whose
molecular weight approximated the molecular weight of the
gelatinase proenzyme and that retained the carboxy-terminal
histidine tag. This is reminiscent of the previously reported
autocatalytic processing of thermolysin (19). However, auto-
processing of GelE may not necessarily be an intramolecular
event, as reported for thermolysin (19). In supernatants of
strain FA2-2 expressing the GelE E137Q His-tagged protein,
we observed disappearance of the histidine tag upon addition
of purified wild-type GelE, as determined by Western blot
analysis with an anti-C-terminal His tag antibody (see Fig. S4
in the supplemental material). This suggested that at least the
C-terminal processing event may not be an intramolecular
In addition to the amino-terminal processing, some M4 fam-
ily members from bacteria belonging to the family Vibrionaceae
also require processing at the carboxy terminus to become fully
active (15, 23). Here we show that the E. faecalis gelatinase
also requires proteolytic processing at the carboxy terminus,
which makes it unique among members of the M4 family from
gram-positive bacteria. Determination of the precise molecular
mass of GelE purified from E. faecalis culture supernatants
resulted in a value (33,030 Da) in good agreement with the
original estimates of Ma ¨kinen et al. obtained by using SDS-
PAGE (33,000 Da), sizing column analysis (32,000 Da), or fast
protein liquid chromatography (31,500 Da) (18).
Our investigation of the reason for the approximately
1,500-Da discrepancy with the predicted molecular mass of
gelatinase indicated that a processing event takes place at the
carboxy-terminal end, most likely between the D304 and I305
residues, and this requires the Q306 residue. As mentioned
above, evidence suggested that GelE itself may be involved in
the processing. An isoleucine residue is not as common as
leucine on the imino side of the scissile bond of known sub-
strates, but such a residue was found for angiotensin and neu-
rotensin, providing further support for the autocatalytic hy-
pothesis (18). Additionally, a hydrophobic amino acid residue
has often been found at this position. Furthermore, the mo-
lecular weight of the I305P mutant protein suggests that there
is an alternative processing event, possibly between Q306 and
V307; this could also be a gelatinase processing site as glu-
tamine at the carboxylic side and leucine at the imino side of
the bond were found in the gelatinase processing site of the
insulin B chain (18).
Nevertheless, additional proteases, such as the coregulated
enzyme encoded by sprE, may contribute to the C-terminal
FIG. 7. Analysis of GelE E137Q protein secretion and biofilm formation. The culture supernatants were concentrated 20-fold and loaded on
a 12% polyacrylamide gel for Coomassie blue staining (A) or for Western blot analysis using anti-C-terminal His tag antibody (B). The molecular
weight standards were the New England Biolabs broad range protein marker (A) and the broad range prestained protein marker (B). The
arrowhead indicates the full-length GelE protein band. (C) Analysis of biofilm formation by FA2-2 strains carrying plasmids expressing wild-type
GelE (pML29), GelE E137Q (pML44), and His-tagged GelE E137Q (pML45). The control strain carried only the vector pML28. The error bars
indicate the standard deviations. OD550, optical density at 550 nm.
VOL. 189, 2007 GELATINASE C-TERMINAL PROCESSING 8841
processing, as suggested by the molecular weight of the Q306P
mutant protein, for which cleavage between the E312 and S313
or S313 and V314 residues must have occurred to explain the
molecular weight obtained by mass spectrometry. Singh et al.
recently showed that in the absence of a functional fsr system,
a basal level of GelE and SprE is produced; thus, we cannot
rule out the possibility that a basal level of SprE contributes to
GelE maturation in strain FA2-2 (35).
The role of the Q306 residue in the processing of the C-
terminal end of GelE is clearly critical for efficient cleavage, as
observed with the wild-type protein; in the purification condi-
tions used in this study, this protein is found essentially fully
processed, while the Q306P mutant protein was processed at
an alternative site, resulting in a protein with a slightly higher
molecular weight than the wild-type protein (Fig. 4). Never-
theless, the GelE-CT14 His-tagged variant seemed to be pro-
cessed to a molecular weight essentially identical to that of the
wild-type protein (Fig. 3) despite the lack of the Q306 and I305
residues, which were replaced by Glu and Leu residues (intro-
duced by the XhoI cloning site), respectively, that linked the
protease to six histidine residues. However, this processing
event was not highly efficient, and its efficiency seemed to
increase when the I305 residue alone or the I305 and Q306
residues were still present (pML42 and pML43) (Fig. 3A
and B), again suggesting a role for these residues in proper
cleavage. Further studies of the biochemical properties of
GelE are required to define the mechanism of this autocat-
The exact role of the GelE C-terminal processing event in
the overall activity of the protease appears to be marginal
because the Q306P mutant protein was only 25 to 30% less
active than the wild-type proteins in the protease assay and it
did not have a significant effect on biofilm formation. The
C-terminal processing event, however, may delay maturation
of the protein, and this may affect protease activity in some
specific environmental conditions.
The observation that the Q306P mutant protein had a higher
rate of autolysis (Fig. 6) raised the possibility that this GelE
variant may remain more tightly associated with the cell mem-
brane or cell wall, perhaps due to the slightly hydrophobic
nature of the 14-amino-acid C-terminal tail. This could favor
processing and maturation of the autolysin substrate of GelE,
muramidase-1 (33), giving rise to the observed phenotype.
Thus, processing of the C-terminal tail of gelatinase may have
a regulatory role in cell division through the regulation of the
It is possible that the C-terminal 14 amino acids have a
secondary signaling function. The role of the conjugative pep-
tide pheromones in E. faecalis, which arise from processing of
a bacterial signal peptide sequence of surface lipoproteins, is
well characterized and provides a precedent for secondary
functions for processed peptide sequences. However, so far
there is no evidence for such a role for the GelE C-terminal
peptide (3, 6). There is also no evidence that the C-terminal 14
amino acids of GelE have a distinct function in the protein
other than slightly inhibiting gelatinase activity, as reported for
the Vibrio cholerae non-O1 hemagglutinin/protease, whose
processing of the C-terminal 2 kDa results in increased pro-
tease activity but decreased hemagglutinin activity (23). Thus,
the precise role of this domain in the biology of Enterococcus
This research was supported in part by Public Health Service grant
AI052289 from the National Institute of Allergy and Infectious Dis-
eases. The Stein Beneficial Trust partially supported oligonucleotide
synthesis and DNA sequencing.
We acknowledge Paula Oliviera for her technical contributions.
1. Bonten, M. J., M. K. Hayden, C. Nathan, J. van Voorhis, M. Matushek, S.
Slaughter, T. Rice, and R. A. Weinstein. 1996. Epidemiology of colonisation
of patients and environment with vancomycin-resistant enterococci. Lancet
2. Bourgogne, A., S. G. Hilsenbeck, G. M. Dunny, and B. E. Murray. 2006.
Comparison of OG1RF and an isogenic fsrB deletion mutant by transcrip-
tional analysis: the Fsr system of Enterococcus faecalis is more than the
activator of gelatinase and serine protease. J. Bacteriol. 188:2875–2884.
3. Chandler, J. R., and G. M. Dunny. 2004. Enterococcal peptide sex phero-
mones: synthesis and control of biological activity. Peptides 25:1377–1388.
4. Chang, P. C., and Y. H. Lee. 1992. Extracellular autoprocessing of a metal-
loprotease from Streptomyces cacaoi. J. Biol. Chem. 267:3952–3958.
5. Cruz-Rodz, A. L., and M. S. Gilmore. 1990. High efficiency introduction of
plasmid DNA into glycine treated Enterococcus faecalis by electroporation.
Mol. Gen. Genet. 224:152–154.
6. Dunny, G. M. 2007. The peptide pheromone-inducible conjugation system of
Enterococcus faecalis plasmid pCF10: cell-cell signalling, gene transfer, com-
plexity and evolution. Philos. Trans. R. Soc. Lond. B 362:1185–1193.
7. Engelbert, M., E. Mylonakis, F. M. Ausubel, S. B. Calderwood, and M. S.
Gilmore. 2004. Contribution of gelatinase, serine protease, and fsr to the
pathogenesis of Enterococcus faecalis endophthalmitis. Infect. Immun. 72:
8. Franco, A. A., S. L. Buckwold, J. W. Shin, M. Ascon, and C. L. Sears. 2005.
Mutation of the zinc-binding metalloprotease motif affects Bacteroides fra-
gilis toxin activity but does not affect propeptide processing. Infect. Immun.
9. Gilmore, M. S., D. B. Clewell, P. Courvalin, G. M. Dunny, B. E. Murray, and
L. B. Rice (ed.). 2002. The enterococci: pathogenesis, molecular biology, and
antibiotic resistance. ASM Press, Washington, DC.
10. Gustin, J. K., E. Kessler, and D. E. Ohman. 1996. A substitution at His-120
in the LasA protease of Pseudomonas aeruginosa blocks enzymatic activity
without affecting propeptide processing or extracellular secretion. J. Bacte-
11. Hancock, L. E., and M. Perego. 2004. Systematic inactivation and phenotypic
characterization of two-component signal transduction systems of Enterococ-
cus faecalis V583. J. Bacteriol. 186:7951–7958.
12. Hancock, L. E., and M. Perego. 2004. The Enterococcus faecalis fsr
two-component system controls biofilm development through production of
gelatinase. J. Bacteriol. 186:5629–5639.
13. Ha ¨se, C. C., and R. A. Finkelstein. 1993. Bacterial extracellular zinc-con-
taining metalloproteases. Microbiol. Rev. 57:823–837.
14. Huycke, M. M., D. F. Sahm, and M. S. Gilmore. 1998. Multiple-drug
resistant enterococci: the nature of the problem and an agenda for the
future. Emerg. Infect. Dis. 4:239–249.
15. Kothary, M. H., and A. S. Kreger. 1987. Purification and characterization of
an elastolytic protease of Vibrio vulnificus. J. Gen. Microbiol. 133:1783–1791.
16. Kristich, C. J., Y. H. Li, D. G. Cvitkovitch, and G. M. Dunny. 2004. Esp-
independent biofilm formation by Enterococcus faecalis. J. Bacteriol. 186:
17. Lindsay, D., and A. von Holy. 2006. Bacterial biofilms within the clinical
setting: what healthcare professionals should know. J. Hosp. Infect. 64:313–
18. Ma ¨kinen, P.-L., D. B. Clewell, F. An, and K. K. Ma ¨kinen. 1989. Purification
and substrate specificity of a strongly hydrophobic extracellular metallo-
endopeptidase (“gelatinase”) from Streptococcus faecalis (strain 0G1-10).
J. Biol. Chem. 264:3325–3334.
19. Marie-Claire, C., B. P. Roques, and A. Beaumont. 1998. Intramolecular
processing of prothermolysin. J. Biol. Chem. 273:5697–5701.
20. Marie-Claire, C., E. Ruffet, A. Beaumont, and B. P. Roques. 1999. The
prosequence of thermolysin acts as an intramolecular chaperone when ex-
pressed in trans with the mature sequence in Escherichia coli. J. Mol. Biol.
21. McIver, K. S., E. Kessler, J. C. Olson, and D. E. Ohman. 1995. The elastase
propeptide functions as an intramolecular chaperone required for elastase
activity and secretion in Pseudomonas aeruginosa. Mol. Microbiol. 18:877–
22. Mohamed, J. A., and B. E. Murray. 2005. Lack of correlation of gelatinase
production and biofilm formation in a large collection of Enterococcus fae-
calis isolates. J. Clin. Microbiol. 43:5405–5407.
8842 DEL PAPA ET AL.J. BACTERIOL.
23. Naka, A., K. Yamamoto, T. Miwatani, and T. Honda. 1992. Characterization Download full-text
of two forms of hemagglutinin/protease produced by Vibrio cholerae non-O1.
FEMS Microbiol. Lett. 77:197–200.
24. Nakayama, J., Y. Cao, T. Horii, S. Sakuda, A. D. L. Akkermans, W. de Vos,
and H. Nagasawa. 2001. Gelatinase biosynthesis-activating pheromone: a
peptide lactone that mediates a quorum sensing in Enterococcus faecalis.
Mol. Microbiol. 41:145–154.
25. Nakayama, J., S. Chen, N. Oyama, K. Nishiguchi, E. A. Azab, E. Tanaka, R.
Kariyama, and K. Sonomoto. 2006. Revised model for Enterococcus faecalis
fsr quorum-sensing system: the small open reading frame fsrD encodes the
gelatinase biosynthesis-activating pheromone propeptide corresponding to
staphylococcal AgrD. J. Bacteriol. 188:8321–8326.
26. O’Donohue, M. J., and A. Beaumont. 1996. The roles of the prosequence of
thermolysin in enzyme inhibition and folding in vitro. J. Biol. Chem. 271:
27. O’Donohue, M. J., B. P. Roques, and A. Beaumont. 1994. Cloning and
expression in Bacillus subtilis of the npr gene from Bacillus thermoproteolyti-
cus Rokko coding for the thermostable metalloprotease thermolysin. Bio-
chem. J. 300:599–603.
28. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2001. Character-
ization of fsr, a regulator controlling expression of gelatinase and serine
protease in Enterococcus faecalis OG1RF. J. Bacteriol. 183:3372–3382.
29. Richards, M. J., J. R. Edwards, D. H. Culver, and R. P. Gaynes. 2000.
Nosocomial infections in combined medical-surgical intensive care units in
the United States. Infect. Control Hosp. Epidemiol. 21:510–515.
30. Sabat, A., K. Kosowska, K. Poulsen, A. Kasprowicz, A. Sekowska, B. B. van
Den, J. Travis, and J. Potempa. 2000. Two allelic forms of the aureolysin
gene (aur) within Staphylococcus aureus. Infect. Immun. 68:973–976.
31. Schmidtchen, A., I. M. Frick, E. Andersson, H. Tapper, and L. Bjorck. 2002.
Proteinases of common pathogenic bacteria degrade and inactivate the an-
tibacterial peptide LL-37. Mol. Microbiol. 46:157–168.
32. Schmidtchen, A., I. M. Frick, and L. Bjorck. 2001. Dermatan sulphate is
released by proteinases of common pathogenic bacteria and inactivates an-
tibacterial alpha-defensin. Mol. Microbiol. 39:708–713.
33. Shockman, G. D., and M. C. Cheney. 2007. Autolytic enzyme system of
Streptococcus faecalis. V. Nature of the autolysin-cell wall complex and its
relationship to properties of the autolytic enzyme of Streptococcus faecalis. J.
34. Sifri, C. D., E. Mylonakis, K. V. Singh, X. Qin, D. A. Garsin, B. E. Murray,
F. M. Ausubel, and S. B. Calderwood. 2002. Virulence effect of Enterococcus
faecalis protease genes and the quorum-sensing locus fsr in Caenorhabditis
elegans and mice. Infect. Immun. 70:5647–5650.
35. Singh, K. V., S. R. Nallapareddy, E. C. Nannini, and B. E. Murray. 2005.
Fsr-independent production of protease(s) may explain the lack of attenu-
ation of an Enterococcus faecalis fsr mutant versus a gelE-sprE mutant in
induction of endocarditis. Infect. Immun. 73:4888–4894.
36. Su, Y. A., M. C. Sulavik, P. He, K. K. Ma ¨kinen, P.-L. Ma ¨kinen, S. Fiedler, R.
Wirth, and D. B. Clewell. 1991. Nucleotide sequence of the gelatinase (gelE)
from Enterococcus faecalis subsp. liquefaciens. Infect. Immun. 59:415–420.
37. Tannock, G. W., and G. Cook. 2002. Enterococci as members of the intestinal
microflora of humans, p. 101–132. In M. S. Gilmore, D. B. Clewell, P.
Courvalin, G. M. Dunny, B. E. Murray, and L. B. Rice (ed.), The entero-
cocci: pathogenesis, molecular biology, and antibiotic resistance. ASM Press,
38. Toledo-Arana, A., J. Valle, C. Solano, M. J. Arrizubieta, C. Cucarella, M.
Lamata, B. Amorena, J. Leiva, J. R. Penades, and I. Lasa. 2001. The en-
terococcal surface protein, Esp, is involved in Enterococcus faecalis biofilm
formation. Appl. Environ. Microbiol. 67:4538–4545.
39. Trieu-Cuot, P., C. Carlier, C. Poyart-Salmeron, and P. Courvalin. 1990. A
pair of mobilizable shuttle vectors conferring resistance to spectinomycin for
molecular cloning in Escherichia coli and in Gram-positive bacteria. Nucleic
Acids Res. 18:4296.
40. Trieu-Cuot, P., and P. Courvalin. 1983. Nucleotide sequence of the Strepto-
coccus faecalis plasmid gene encoding the 3?5?-aminoglycoside phospho-
transferase type III. Gene 23:331–341.
41. Waters, C. M., M. H. Antiporta, B. E. Murray, and G. M. Dunny. 2003. Role
of the Enterococcus faecalis GelE protease in determination of cellular chain
length, supernatant pheromone levels, and degradation of fibrin and mis-
folded surface proteins. J. Bacteriol. 185:3613–3623.
42. Zeng, J., F. Teng, and B. E. Murray. 2005. Gelatinase is important for
translocation of Enterococcus faecalis across polarized human enterocyte-
like T84 cells. Infect. Immun. 73:1606–1612.
VOL. 189, 2007 GELATINASE C-TERMINAL PROCESSING8843