APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 2011, p. 442–451
Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 77, No. 2
Diversity of the fsr-gelE Region of the Enterococcus faecalis Genome
but Conservation in Strains with Partial Deletions of the fsr Operon?†
Jessica R. Galloway-Pen ˜a,1,2,3‡ Agathe Bourgogne,1,2‡ Xiang Qin,4and Barbara E. Murray1,2,3*
Division of Infectious Diseases, Department of Medicine,1Center for the Study of Emerging and Re-Emerging Pathogens,2
and Department of Microbiology and Molecular Genetics,3University of Texas Medical School, Houston, Texas, and
Human Genome Sequencing Center, Baylor College of Medicine, Houston, Texas4
Received 25 March 2010/Accepted 14 November 2010
Most Enterococcus faecalis isolates carry gelE, but many are gelatinase nonproducers due to the lack of fsrC
(EF_1820) to EF_1841 (fsrC-EF_1841; 23.9 kb in strain V583), including most of the locus encoding Fsr, which
activates gelE expression. Analysis of 22 accessible E. faecalis genomes revealed the identity of the 53-amino-
acid propeptide of fsrD across multiple MLSTs (multilocus sequence types), although 12 distinctly different
variations were found in the EF_1814-to-EF_1902 region. Diversity was seen in fsrABC, in the region EF_1814
to EF_1902, and in a 700-kb region surrounding fsrC-EF_1841. However, analysis of five sequenced strains
carrying the fsrC-EF_1841 deletion and the putative integrative conjugative element efaB5 showed almost
identical single nucleotide polymorphisms (SNPs) in gelE and an identical junction sequence, despite their
unrelated MLSTs, in contrast to those shown by strains without the deletion. Further analysis confirmed the
conserved gelE SNPs in 6 additional strains (11 in total) with the deletion. While we were unable to detect
evidence of spontaneous deletion using OG1RF and 8 other strains, we were able to engineer a deletion of the
37-kb fsrC-EF_1841 region of OG1RF without deleterious effects, and the 37-kb mutant showed changes in
biofilm and chaining similar to those shown by fsr-gelE mutants. In conclusion, we describe the identity of fsrD
despite high plasticity within the fsrC-EF_1841 region and the surrounding sequence. However, strains lacking
the fsrC-EF_1841 region show a distinct conservation of the sequence surrounding this deletion and in gelE,
suggesting that the deletion may result from horizontal transfer and recombination.
Enterococci are Gram-positive commensal bacteria of hu-
mans, animals, and insects but can also cause a wide range of
diseases, including urinary tract infections, bloodstream infec-
tions, wound infections, and endocarditis (26). Nosocomial
enterococcal infections are frequently caused by hospital-ac-
quired strains rather than by strains from the patient’s own
community-derived indigenous flora (13, 46). These strains
often belong to specific clones which are thought to have ac-
quired factors important for virulence, colonization, and/or for
fitness in the hospital environment (46).
Gelatinase is produced by approximately 60% of clinical
Enterococcus faecalis isolates (7). The Fsr system, encoded by
the fsrABDC operon, and gelatinase production are intrinsi-
cally linked since the Fsr system is the main activator of gelE
expression (4, 36, 37). The fsrABDC operon, a homologue of
the agrABCD operon in Staphylococcus aureus, encodes the
elements of a quorum-sensing system (29, 30). While the pres-
ence of a functional Fsr system and/or gelatinase production
increases the severity of disease in animal, plant, and nema-
tode models of E. faecalis infection (10, 16, 28, 37, 44, 45),
neither are required for the organism to cause disease (7, 45).
When testing E. faecalis isolates for gelatinase activity, Qin
et al. observed that isolates that did not produce gelatinase but
had the gelE gene often lacked the fsrB gene (37). This study
was followed by a report by Nakayama et al. which found that
the majority of gelatinase nonproducer but gelE?strains
lacked a region present in strain V583 between the 5?-end
portion of fsrC and the 3?-end portion of EF_1841, including
the fsrABD genes; this missing region corresponds to a 23.9-kb
fragment in V583 (31). Strains lacking this region are gelatin-
ase nonproducers in standard assays due to the absence of a
functional Fsr system (31, 37, 40). The possible benefit of not
having the fsrC (EF_1820)-to-EF_1841 (fsrC-EF_1841) region,
a situation that occurs in approximately 28% of reported
strains, remains unclear (37, 40). Two groups have reported
the presence of gelatinase nonproducer colonies within a cul-
ture of a gelatinase producer isolate (9, 31), leading to the
hypothesis of ongoing spontaneous deletion among gelatinase
producer strains. However, in both cases, no further experi-
mentation to confirm that the gelatinase nonproducer isolates
were related to their gelatinase-positive parent isolate was
In a previous study (36), the last 150 bp of fsrB (now fsrD)
homologous to agrD were shown to be identical in five clinical
isolates, while three isolates had single silent base pair changes.
Due to the availability of multiple E. faecalis genome se-
quences, in the present study we set out to investigate the
chromosomal diversity of 22 isolates of E. faecalis from multi-
ple multilocus sequence types (MLSTs) in the region from
EF_1814 to EF_1902 (based on V583 annotation). Although
we found considerable variability in this region, we did indeed
observe identity in fsrD, and in clonally unrelated isolates lack-
ing the fsrC-EF_1841 region, we also found unique and highly
* Corresponding author. Mailing address: Division of Infectious
Disease, Center for the Study of Emerging and Re-emerging Patho-
gens, University of Texas Medical School, 6431 Fannin St., MSB 2.112,
Houston, TX 77030. Phone: (713) 500-6745. Fax: (713) 500-6766.
‡ These authors contributed equally to this study.
† Supplemental material for this article may be found at http://aem
?Published ahead of print on 19 November 2010.
conserved nucleotide changes in gelE and the region close to
the fsrC-EF_1841 junction, raising the possibility that this de-
letion may have been introduced horizontally by recombina-
tion. We were also able to delete the fsrC-EF_1841 region
using allelic replacement in strain OG1RF and found that
growth, biofilm, and chaining phenotypes were similar to those
in fsr-gelE mutants of OG1RF, demonstrating that this region
is not needed for in vitro growth.
MATERIALS AND METHODS
Strains and media. The strains used in this study are listed in Tables 1 and 2.
Twenty-two E. faecalis-accessible genomes were analyzed for their genomic or-
ganization. An additional 38 isolates were chosen to represent diverse multilocus
sequence types (MLSTs), while an additional 10 isolates were selected based on
previous studies showing the lack of fsrB but the presence of gelE (40). Insertion
mutants TX5240 (insertion in fsrA), TX5241 (insertion in fsrB), TX5242 (inser-
tion in fsrC) (36), and TX5128 (insertion in gelE) (45) were used for biofilm and
chaining assays. All strains were grown routinely in brain heart infusion (BHI)
broth (Difco laboratories, Detroit, MI) at 150 to 200 rpm or on BHI agar at 37°C.
Gelatinase production was measured as previously described after overnight
incubation at 37°C (37).
Computer analyses and annotation. In addition to V583 and OG1RF, 20 E.
faecalis genomes accessible in June 2009 were analyzed for their genomic orga-
nization between EF_1814 and EF_1902. For each genome, the sequences were
recovered from the NCBI server or the Broad Institute server and analyzed with
resources described elsewhere (3). Each gene is denoted either with its charac-
terized name, by an EF_ tag based on the V583 annotation (35), or with an
OG1RF tag if it was first described as a unique sequence in OG1RF (3).
Detection of fsrC, EF_1841, or the junction sequence. To detect the presence
or lack of the fsrC-EF_1841 region, we designed a multiplex PCR based on the
following four primers. AB83 (outside the deletion) and AB84 would amplify a
575-bp fragment of EF_1841, while AB85 (outside the deletion) and AB86 would
amplify a 940-bp fragment of fsrC (EF_1820) when the genes are intact (see
Table S1 and Fig. S1 in the supplemental material). In strains lacking the
fsrC-EF_1841 region, AB83 and AB85 would amplify a 739-bp fragment of the
junction sequence. Additional primers AB74, AB75, AB244, and AB245 were
used to confirm the presence of fsrC (see Table S1).
Conjugation experiments. The possibility of transposition of efaB5 with the
fsrC-EF_1841 deletion was investigated using cross-streak and filter mating be-
tween JH2-2::Tn916 as the donor (11) and OG1RF tagged with kanamycin as the
recipient (32). Mating plates were incubated overnight at 37°C. Growth present
in the area of the cross or on the filter was resuspended in 1.0 ml of sterile saline
(0.9% NaCl), and serial 10-fold dilutions were made. Aliquots were plated onto
Todd-Hewitt agar plates containing 2 mg/ml kanamycin (to select for the recip-
ient OG1RF cells) and 10 ?g/ml tetracycline (to select for Tn916). OG1RF
transconjugants were then replica plated on 3% gelatin to detect the possible loss
of the fsrC-EF_1841 region.
Detection of the efaB5 element in relation to the fsrC-EF_1841 region deletion.
A total of 17 strains (7 from the original 38 strains plus the 10 previously
described ) from diverse sequence types (by multilocus sequence typing or
trilocus sequence typing [6, 41]) known to lack fsrB but to have gelE (40) were
tested for the 739-bp PCR product indicating the lack of the fsrC-EF_1841 region
TABLE 1. Sequenced strains analyzed in this studyb
BI 19, 39
ATCC 4200 105
aThe strains are listed by the genotype of the EF_1814-to-EF_1902 region. The three genotypes (A, B, and C) encompassing the greatest number of strains are listed
first, followed by the genotypes with one strain each. Variations of genotypes A, B, and C were detected in the strains CH188, ATCC 4200, and E1Sol and T11 and
designated A?, B?, or C?, respectively. Sequences of these isolates were analyzed for the fsrC-EF_1841 region (23.9 kb), the 14.8-kb insertion region described in
OG1RF, and the putative integrative conjugative element mobile element efaB5. Strains with genotype A (V583, Merz96, TX0104, and HH22) are within CC2.
bInformation on the MLST and the presence (?) or absence (?) of fsrB (EF_1820), gelE (EF_1818), and gelatinase production (Gel?) was either acquired from
the literature (23) and confirmed or tested by us. A “P” indicates the presence of the part of the locus/gene mentioned.
cTIGR, the Institute for Genomic Research; BI, Broad Institute; BCM, Baylor College of Medicine.
dTX5599 is the mutant constructed in this study by deleting 37 kb (the 23.9-kb region and the 14.8-kb region) in OG1RF.
VOL. 77, 2011 CONSERVED DELETION OF fsr GENES 443
(using AB83 and AB85 as described above) and for the presence of the efaB5
mobile element in conjunction with the deletion using primers EF_1877efaB5F,
EF_1877efaB5R, 5?efaB5F, 5?efaB5R, 3?efaB5F, and 3?efaB5R in a multiplex
PCR (see Table S1 and Fig. S1 in the supplemental material). The
EF_1877efaB5F and EF1877efaB5R primers should amplify a 338-bp region of
EF1877, a gene conserved within the efaB5 element, if the efaB5 element is
present. 5?efaB5F and 5?efaB5R should amplify a 616-bp region if the 5? region
of the efaB5 element is inserted before EF_1846, and the 3?efaB5F and
3?efaB5R primers should amplify a 517-bp region if the 3? region of efaB5 is
inserted before EF_1898. Therefore, in JH2-2, which contains the deletion and
efaB5 inserted between EF_1844 and EF_1898, all three regions amplified. In
OG1RF, which does not contain the deletion or the efaB5 element, 5?efaB5F
and 3?efaB5R amplified a 689-bp region between the noncoding regions before
EF_1846 and EF_1898.
Generation and detection of a large chromosomal deletion. To detect the
possible spontaneous occurrence of the deletion of the fsrC-EF_1841 region,
primers AB83 and AB85 were used. To create a deletion mutant of OG1RF, we
used the PheS* system as described previously (18), using the primers AB152 and
AB153 (see Table S1 in the supplemental material) to amplify a 2.8-kb fragment
overlapping the junction (Jct) between fsrC and EF_1841 from JH2-2 (reference
TABLE 2. fsrC-EF_1841 region genotypes and gelatinase phenotypes of clonally diverse nonsequenced isolates tested in this study
aParentheses indicate an alternative name used in the literature, while asterisks indicate the 10 additional isolates chosen due to previous data (38) indicating the
absence of the fsrC-EF_1841 region. Isolates without asterisks are the initial 38 isolates chosen for study.
bSTs were done by MLST, except those designated with a “T,” which are trilocus sequence types (6).
cThe fsrC-EF_1841 genotypes were detected using a multiplex PCR, as described in Materials and Methods, where AB83 and AB84 amplify EF_1841, AB85 and
AB86 amplify fsrC (EF_1820), and AB83 and AB85 amplify the fsrC::EF_1841 junction sequence, indicating the deletion.
dThe efaB5 mobile element was also detected using a multiplex PCR, which would amplify three bands if present (see Materials and Methods). The absence of the
element would produce a 689-bp band between the noncoding region after EF_1845 and EF_1898. “†” indicates that none of these bands were produced.
444GALLOWAY-PEN ˜A ET AL.APPL. ENVIRON. MICROBIOL.
strain for the “23.9-kb” deletion) for the purpose of deleting the fsrC-EF_1841
region in OG1RF. This 2.8-kb fragment was cloned into pCJK47, creating
pCJK47::Jct. We transformed pCJK47::Jct into OG1RF and selected for eryth-
romycin-resistant colonies (resulting
pCJK47::Jct and OG1RF). We then selected for the excision of the plasmid on
MM9YEG agar medium containing p-Cl-Phe (catalog no. 6506; Sigma), result-
ing in colonies which carried either the parent or the deletion genotype. The
colonies with deletion of the fsrC-EF_1841 region were then scored on gelatinase
plates after incubation overnight at 37°C. One Gel?colony was selected and
named TX5599. The deletion was confirmed by pulsed-field gel electrophoresis
(PFGE) with NotI digestion, lack of gelatinase activity, and sequencing after
amplification with AB152 and AB153.
Biofilm formation assays. A biofilm density formation assay was carried out as
described previously (25). The absorbance of each well was measured at an
optical density at 570 nm (OD570). Each assay was performed in triplicate in
eight independent experiments.
RESULTS AND DISCUSSION
Genomic diversity between EF_1814 and EF_1902. Since we
had previously noted the conservation of fsrD in a small subset
(36) and had previously observed that isolates that did not
produce gelatinase but had the gelE gene often lacked the fsrB
gene (37) (which was later described to be the “23.9-kb” de-
letion ), we set out to analyze the fsr-gelE genomic region
to see how diverse or conserved this region is and if this could
be related to MLST or strains from specific clones. The 22
accessible genomes that were analyzed come from different
multilocus sequence types (MLSTs) (Table 1), sources (clinical
and commensal), and times of isolation (from 1925 to 2005)
(23). We first blasted each genome with gelE (EF_1818) and
EF_1842 to determine the location of the fsrC-EF_1841 region
on the chromosome/contig and the size of the fragment. The
analysis of this region was performed with genes EF_1814 to
EF_1902, which were present in all genomes. A total of 12
different genetic organizations, or genotypes, were detected
based on variation in the region between EF_1814 and
EF_1902 (Fig. 1), with three genotypes (A, B, and C) contain-
ing more than one isolate each. Genotype A carries the longest
DNA sequence between EF_1814 and EF_1902, due to the
presence of the previously described mobile element efaB5
(consisting of EF_1846 to EF_1897, which is ?49.5 kb in strain
V583 ), in addition to the presence of the fsrC-EF_1841
region. This genotype is present in strains within clonal com-
plex 2 (CC2), as follows: V583 (sequence type 6 [ST6]), HH22
(ST6), TX0104 (ST2), and Merz96 (ST103). Isolates within
genotype B/B? lack the fsrC-EF_1841 region but possess an
efaB5 element (the B? genotype carries an additional 14 kb
within the efaB5 element). The isolates with genotype B are
found in diverse lineages by MLST (Fig. 2) and include JH2-2
(ST8), HIP11704 (ST4), T2 (ST11), T8 (ST8), and TX1322
(ST64). Finally, genotype C isolates carry an additional 14.8-kb
fragment (encoding OG1RF_0128 to OG1RF_0140) in the
fsrC-EF_1841 region; none of the three isolates with this ge-
notype (OG1RF [ST1], T1 [ST21], and T3 [ST67]) carry an
efaB5 element or are related by ST. For the 10 remaining
EF_1814 to EF_1902 genotypes, the length of the sequence
between EF_1814 and EF_1902 was found to range from 12.1
kb (ATCC 29200) to more than 70 kb for CH188. In summary,
13 strains were found to carry a complete fsr-gelE locus, while
9 strains were found to have an incomplete locus, with 6 of
which having a conserved deletion (5 having an identical fsrC-
EF_1841 deletion junction and 1 having a single base change,
as seen in Fig. S2 in the supplemental material).
Genetic variation of E. faecalis in a 700-kb region, including
EF_1814 to EF_1902. Since the genomes with genotype B
presented such a high degree of similarity between each other
in the EF_1814-to-EF_1902 region, we investigated the simi-
larities of a larger region to see if all the strains were conserved
or if the similarities were limited to the EF_1814-to-EF_1902
region. As mentioned previously, only the V583 and OG1RF
genomes are complete and closed, while the other genomes are
in a less finished stage, ranging from simple shotgun sequences
to large assembled contigs. OG1RF was used as a reference
since it contains a limited number of putative mobile elements.
For 17 of the genomes, we could identify a 700-kb region in
common, encompassing a sequence homologous to that of the
EF_1814-to-EF_1902 region, corresponding to position 1.02 to
1.72 Mb in OG1RF (EF_1209 to EF_2047 in V583). As shown
in Fig. 3, the 700-kb regions differ considerably among the
strains, mainly due to deletion/insertion events. The three ge-
notype B strains presented 7 major differences from each
other, indicating that although highly similar in the EF_1814-
to-EF_1902 region, those strains are not as highly conserved in
other areas of the genome.
Strains lacking the fsrC-EF_1841 region represent diverse
MLSTs and do not produce gelatinase. While Nakayama et al.
found that 56% of the 46 isolates obtained from a local hos-
pital lacked the fsrC-EF_1841 region (31), Roberts et al. re-
ported that 28% of 215 isolates lacked the fsrC-EF_1841 re-
gion in a more diverse collection of isolates (40). However, in
both studies, the clonality of those isolates was unknown. We
therefore initially studied 38 strains (Table 2) belonging to
diverse MLSTs (in addition to the sequenced strains), used a
set of multiplex primers to determine the absence or presence
of this region, and correlated this with their ability to produce
gelatinase. As shown in Table 2, only 7 of the initial 38 strains
(18%) from our collection generated a 739-bp PCR product
characteristic of the deletion of the fsrC-EF_1841 region, and
they were all gelatinase nonproducers; 6 of these strains were
unrelated by MLST, while 2 were double-locus variants of each
other (Fig. 2). A total of 22 of the 38 isolates (58%) resulted in
a PCR product, indicative of the presence of the full fsrC and
EF_1841 genes, all of which produced gelatinase. Of the re-
maining 9 strains, 4 strains were positive for only the fsrC
region, and 3 of those 4 strains did not produce gelatinase. The
other 5 strains were negative for the fsrC region but positive for
EF_1841 by PCR, and none produced gelatinase.
Through a combination of previously published data on the
sequenced strains (23), sequence analysis, and confirmation of
these results through multiplex PCR and gelatinase production
assays (Table 1), we verified that all sequenced genotype B/B?
strains (negative for the fsrC-EF_1841 region) were gelatinase
nonproducers although gelE positive, while all genotype A and
genotype C/C? strains were gelatinase producers. Of the re-
maining 7 strains, Fly1 and ARO1/DG are gelatinase produc-
ers, as to be expected since they have the fsrABDC and gelE-
sprE operons intact. However, CH188 and JH1 were gelatinase
nonproducers, despite complete fsrABDC and gelE-sprE loci.
X98, D6, and ATCC 29200 are gelatinase nonproducers since
some or all of the fsrABDC operon is absent. In sum, 14 of the
60 initial isolates (23%) examined (38 tested by PCR and the
VOL. 77, 2011 CONSERVED DELETION OF fsr GENES445
22 sequenced genomes analyzed) lacked the fsrC-EF_1841 re-
gion, and these strains do not produce gelatinase.
Limited variability in the Fsr system and none in FsrD. In a
previous study (36), the conservation of fsrD in a limited group
of isolates had been noted. Since the 13 E. faecalis strains in
this study with complete Fsr systems were more diverse by
MLST, we also compared their fsr-gelE loci for possible se-
quence diversity. It is known that among the agr loci in Staph-
ylococcus aureus, which has the best described cyclic peptide-
bacteria and is a homologue of the fsrABDC system, sequence
variation is particularly evident in the autoinducer precursor
FIG. 1. Genomic diversity of E. faecalis between EF_1814 and EF_1902. Twenty-two of the sequenced strains plus the 37-kb deletion mutant
of OG1RF are shown. Redasoft Visual Cloning was used to find open reading frames (ORFs) representing EF_1814 to EF_1902 in sequences
downloaded from either the NCBI server or the Broad Institute server. The first two ORFs (EF_1814 and EF_1815) and the last five ORFs
(EF_1898 to EF_1902) of this region, which are common between all strains, are shown in gray. The fsrC (EF_1820) and EF_1841 genes are
indicated by black arrows. When a region is missing, the junction areas are indicated with vertical squiggly lines, and the space between is filled
with a dotted line. The fsrAB (EF_1821 and EF_1822), gelE (EF_1818), and sprE (EF_1817) genes are labeled with arrows filled with diagonal lines.
The immediately adjacent loci, EF_1816, EF_1843, and EF_1844, are represented with white arrows. The area from EF_1823 to EF_1840 is
represented with a spotted background, while the one representing the 14.8-kb region contains vertical lines. The efaB5 region from EF_1847 to
EF_1896 is represented with horizontal lines. All other insertions are represented with crosshatching. Letters on the left side of the diagram
indicate the genotypes based on EF_1814 to EF_1902, corresponding to Table 1.
446 GALLOWAY-PEN ˜A ET AL.APPL. ENVIRON. MICROBIOL.
AgrD (only a cysteine conserved of the 7 to 9 amino acids),
AgrB (only 68 of the 187 amino acids are conserved), and
AgrC (only 177 of the 423 amino acids are conserved) due to
the presence of a hypervariable region which results in at least
four agr specificity groups (17, 33). Unlike what has been found
in staphylococci (30), we observed perfect conservation, not
only of the 11 amino acids that are the basis for the autoin-
ducer cyclic peptide but also of the full FsrD polypeptide (53
amino acids) of the 13 diverse strains of E. faecalis evaluated.
When looking at the other constituents of the fsr operon, we
found that the 13 strains carrying the complete fsrABC genes
had up to 6 amino acid variations for fsrA, 7 amino acid vari-
ations for fsrB, and 9 amino acid variations for fsrC, relative to
V583. Among the 20 strains carrying gelE and sprE, we found
up to 12 amino acid variations for gelE and 15 for sprE, relative
to V583. None of the proteins of the fsr-gelE locus were trun-
cated as a consequence of a frameshift or point mutation.
A conserved junction sequence and SNPs in strains lacking
the fsrC-EF_1841 region suggest that this deletion may occur
by homologous recombination. A detailed analysis of the 6
genotype B/B? strains lacking the fsrC-EF_1841 region found
an identical sequence (?600 bp, with 300 bp of fsrC and 300 bp
of EF_1841) overlapping the junction (except for T8, which
had a single base difference). Most of the sequences from the
junction through gelE between these sequenced strains were
identical; however, some of the strains do not have complete
sequence coverage within this region. Furthermore, 11 strains
(6 sequenced strains and 5 from our collection) lacking the
fsrC-EF_1841 region also shared a nearly identical gelE gene
profile with 4 identical single nucleotide polymorphisms
(SNPs) over the 1,533 nucleotides within the gene (0.26%
variation), compared to up to 3.2% variation among the 12
other gelE sequences from strains carrying the fsrC-EF_1841
region (Fig. 4). We had speculated that there might be an even
higher level of differences in strains lacking the fsrC-EF_1841
region since the gelE gene is not expressed, at least under
standard in vitro conditions, and thus, there might be more
degeneration of the sequence.
The fact that the strains lacking the fsrC-EF_1841 region are
scattered among various MLSTs (some other members of
which have this region) (Fig. 2) suggests that the deletion of
the fsrC-EF_1841 region is not an ancient event that occurred
before those strains evolved into distinct clones. Instead, the
presence of a highly conserved and distinctive junction se-
FIG. 2. Strains lacking the fsrC-EF_1841 region are found in diverse MLST lineages. In this diagram, an eBURST of all MLSTs in the E.
faecalis MLST database (http://efaecalis.mlst.net/) are depicted. Single-locus variants are linked by a line between them. The larger the middle
circle, the more abundant this ST is in the database. The STs of 17 strains (6 sequenced and 11 from our collection) within this study lacking the
fsrC-EF_1841 region are boxed.
VOL. 77, 2011 CONSERVED DELETION OF fsr GENES447
quence suggests a common mechanism for the deletion; more-
over, the presence of a virtually identical gelE gene SNP profile
suggests that the region surrounding the deletion was intro-
duced by horizontal transfer from a common or conserved
source and recombined into the chromosome.
Rice and Carias reported that insertion of Tn5385 into a
recipient’s chromosome likely occurred by recombination
across flanking regions homologous between donors and
recipients, since identical deletions present in the donor
sequences flanking Tn5385 were found in the transconju-
gants (38). If there is frequent recombination in this region,
it is possible that the deletion could be transferred via ho-
mologous recombination through the movement of a mobile
element, such as efaB5, as postulated previously (31). EfaB5
is a Tn916-like element containing several genes related to
virulence and flanked by site-specific recombinases, indicative
of a mobile element. We reasoned that the “23.9-kb” deletion
could be occurring in a scenario similar to that described for
Tn5385 by Rice and Carias, where the deletion would be co-
transferred with efaB5 and incorporated into the genome via
homologous recombination, leading to deletion of the fsrC-
EF_1841 region. Using a multiplex PCR, we also screened for
the presence of the efaB5 element in relation to the deletion
in the 7 strains (for which gelE was sequenced and the presence
of the deletion known) plus 10 additional isolates for which
previous findings indicated the deletion (40). We found that 16
out of these 17 strains lacking the fsrC-EF_1841 region (94%)
had the efaB5 element, compared to 7 out of 22 strains which
contain the fsrC-EF_1841 region (32%) (Table 2). An addi-
tional 8 isolates did not give a PCR product indicative of the
FIG. 3. Genetic diversity of E. faecalis in an ?700-kb region, including EF_1209 to EF_2047. Seventeen of the genomes which contain ?700
kb in common corresponding to nucleotide positions 1.02 Mb to 1.72 Mb in OG1RF (EF_1209 to EF_2047 in V583) were analyzed for genomic
differences using OG1RF as a reference. The gray bars indicate the sequences corresponding to OG1RF, deletions are indicated with black bars,
and insertions of ?1 kb are indicated with white bars. The black and white bars are to scale with the size of the deletion/insertion. Smaller white
bars are insertions of ?1 kb but ?5 kb, while larger white bars designated with letters are insertions of ?5 kb or insertional elements that have
been previously described; these are positioned directly below where they appear on the diagram for reference. “a” carries a WxL locus, “b” carries
a probable ABC transporter and a two-component system, “c” is phage 3, “d” is a probable multidrug resistance transposon, “e” is a probable
38.5-kb phage, “f” is a probable 11-kb fragment that does not appear to be mobile, “g” is a probable 25-kb phage, “h” carries enlA, “i” (bar with
vertical lines) is the efaB5 element, and “k” is phage 4. The bar with diagonal lines indicates an inversion in the D6 genome compared to OG1RF.
There were 5 to 16 variations of ?1 kb between each genome and OG1RF, with an average of 9 variations per genome.
448 GALLOWAY-PEN ˜A ET AL.APPL. ENVIRON. MICROBIOL.
presence of the efaB5 element nor the absence of the efaB5
element (PCR positive for the junction between EF_1846 and
EF_1898, the region flanking the efaB5 element) (see Fig. S1
in the supplemental material). One possibility is that these
isolates contain a large insertion, such as the enlA element or
the 25-kb insertion found in Fly1, in which case, although
efaB5 is not present, the junction would not amplify due to
these large insertions.
We next attempted to show transfer of the deletion. Since
the efaB5 element does not carry any marker, we used a
JH2-2::Tn916 (Tet) donor strain to mediate conjugation. After
10,000 tetracycline-resistant transconjugant colonies on gela-
tinase plates were screened, all were found to be gelatinase
producers, indicating that none had the deletion. In a recent
report, Manson et al. (21) reported that the frequency of trans-
fer of chromosomal regions mediated by conjugative plasmids
was 10?9; our assay conditions were not sufficiently sensitive to
detect transconjugants that arrive at this frequency. This mech-
anism could also mediate horizontal transfer of the deletion
and surrounding regions.
No spontaneous occurrence of a gelatinase nonproducer
phenotype. It was suggested by two groups that gelatinase
nonproducer colonies may happen by spontaneous loss of the
“23.9-kb” region (9, 31). To estimate the probability of a nat-
ural occurrence of this phenotype, we spread dilutions of over-
night cultures of OG1RF and 7 strains from CC2 (all gelE
positive) on gelatinase plates but failed to observe gelatinase
nonproducer colonies. Then, we grew OG1RF for 60 genera-
tions, extracted genomic DNA (gDNA), and used primers
AB83 to AB85 to screen for the junction if the deletion had
occurred spontaneously. With a mixture of OG1RF (which has
the fsrC-EF_1841 region) and JH2-2 (which lacks the fsrC-
EF_1841 region), our level of detection was 1 CFU of JH2-2
for 106CFU of OG1RF. We repeated this experiment 5 times
with the same outcome: we never detected a PCR product,
indicating the presence of a spontaneous deletion of the fsrC-
EF_1841 region. Since it was proposed that the efaB5 element
may be implicated in the deletion of the fsrC-EF_1841 region
(31), we also tested the 7 CC2 strains (genotype A carrying the
efaB5 element from various origins) for the natural occurrence
of the deletion by using the same conditions as the ones de-
scribed above for OG1RF but, again, failed to detect the inci-
dence of a deletion by PCR.
Generation and analysis of an isogenic mutant of OG1RF
with a deletion of the fsrC-EF_1841 region. The fsrC-EF_1841
region in OG1RF is very similar to the one present in V583,
except for an additional 14.8-kb fragment inserted between the
corresponding EF_1826 and EF_1827 loci (4), leading to a
fragment of 37 kb from fsrC to EF_1841. Since most naturally
occurring Gel?E. faecalis isolates lack this region, unlike lab-
FIG. 4. SNP analysis of gelE. SNP analysis was done on the entire gelE open reading frame of 25 strains (20 sequenced strains with intact gelE
genes and 5 strains from our collection). The position of a SNP is indicated with the numbers at the top. The letters at the top indicate the
nucleotide at each position that is shared by the most isolates (consensus). All other letters indicate the SNPs found in the corresponding strain.
Only variations from the consensus are indicated. A dash indicates that sequence was not available for that region or nucleotide due to lack of
sequence coverage. This figure includes all sequenced strains plus isolates from our collection lacking the fsrC-EF_1841 region for which we had
the gelE sequence.
VOL. 77, 2011 CONSERVED DELETION OF fsr GENES 449
oratory-generated fsr or gelE mutants, we sought to see if the
rest of this region was important for the phenotypes, i.e.,
changes observed in biofilm and chaining with laboratory-gen-
erated fsr-gelE insertion mutants. Toward this end, we replaced
the fsrC-EF_1841 region with the JH2-2 junction sequence by
allelic replacement using the PheS* system (18). One mutant
was selected and designated TX5599. Analysis of the sequence
of the 2-kb region overlapping the junction sequence in
TX5599 and JH2-2 compared to that in OG1RF showed that
one crossover event occurred between 756 and 550 bp up-
stream of the junction and the second occurred between 938
and 1,134 bp downstream of the junction. The deletion was
also confirmed by PFGE after digestion with NotI (Fig. 5).
Therefore, an isogenic mutant of OG1RF with a deletion of
the fsrC-EF_1841 region was easily obtained, indicating that
major adaptations were not needed and that this region is not
needed for growth in vitro in OG1RF. We saw very similar
biofilm levels (Fig. 6) and chaining phenotypes (data not
shown) when comparing the 37-kb deletion mutant of OG1RF
and fsr-gelE insertion mutants, also indicating that loss of the
rest of this region does not seem to compensate for the de-
crease in biofilm and chaining seen in fsr-gelE mutants. In
addition, these results also demonstrated that a large genomic
deletion can be generated in E. faecalis.
Conclusion. In summary, we analyzed 22 genomes repre-
senting 19 different MLSTs with the goal of better character-
izing the region that includes the previously designated “23.9-
kb” deletion (fsrC-EF_1841). We found 12 unique genotypes
based on the sequence surrounding this region, indicating the
high plasticity of the genome at this location. We observed
that, although the EF_1814-to-EF_1902 region was similar in
strains of different MLSTs lacking the fsrC-EF_1841 region,
strains differed by insertion and deletion events (Fig. 3) when
looking at a larger fragment (corresponding to ?700 kb in
Despite clonal variability of these genomes by MLST, we
found perfect conservation in fsrD, encoding the propeptide of
the autoinducer, within all strains and high identity among the
fsrA, fsrB, and fsrC genes, unlike what is seen in agr of S. aureus.
Including the available sequenced strains, we found that strains
from various clonal backgrounds lacking the fsrC-EF_1841 re-
gion represent approximately 23% of isolates tested, and none
of these strains produced gelatinase. We also found that strains
lacking the fsrC-EF_1841 region had a very highly conserved
junction sequence (?600 bp), including an identical SNP pro-
file in gelE, suggesting that the deletion and surrounding region
may be horizontally transferred between strains. This is also
consistent with the observation that a high percentage (94%)
of strains lacking the fsrC-EF_1841 region contain the efaB5
mobile element in its flanking sequence.
Despite the frequency of the fsrC-EF_1841 deletion, we did
not detect evidence that mutants of OG1RF lacking this region
occur spontaneously or after mating with JH2-2::Tn916. This
work also demonstrated that the PheS* system can be used to
delete a large fragment (37 kb) of chromosomal DNA, and
nothing between fsrC and the homologue of EF_1841 is essen-
tial for the growth of OG1RF in BHI medium, nor does the
deletion of this region seem to compensate for the biofilm and
chaining seen in specific fsrA, fsrB, fsrC, and gelE insertion
We are grateful to K.V. Singh and S. R. Nallapareddy for helpful
This work was supported by National Institutes of Health grant R37
AI47923 from the Division of Microbiology and Infectious Diseases to
FIG. 5. PFGE of NotI digestion of OG1RF and its mutant
(TX5599), isogenic for the deletion from fsrC to EF_1841. Digestion
with NotI showed the band corresponding to the fsrC-EF_1841 region
migrating as ca. 280 kb for OG1RF and ca. 240 kb for TX5599, a
difference corresponding to the 37-kb deletion in OG1RF. The band
corresponding to the fsrC-EF_1841 region and the fragment after the
deletion area is shown by a black arrowhead.
FIG. 6. Biofilm formation does not differ between fsr-gelE mutant
strains and the 37-kb mutant fsrC-EF_1841 in OG1RF. Biofilm for-
mation in the insertion mutants of fsrA (TX5240), fsrB (TX5241), fsrC
(TX5242), and gelE (TX5128) and in the 37-kb deletion mutant
(TX5599), all in the OG1RF background, was assayed. Each circle
represents the average OD570of a triplicate of one of eight indepen-
450 GALLOWAY-PEN ˜A ET AL.APPL. ENVIRON. MICROBIOL.
B.E.M. J.R.G.-P. is supported by Molecular Basis of Infectious Dis- Download full-text
eases Training Grant T32 AI55449.
1. Ackermann, H. W., T. Caprioli, and S. S. Kasatiya. 1975. A large new
Streptococcus bacteriophage. Can. J. Microbiol. 21:571–574.
2. Banai, M., and D. J. LeBlanc. 1983. Genetic, molecular, and functional
analysis of Streptococcus faecalis R plasmid pJH1. J. Bacteriol. 155:1094–
3. Bourgogne, A., et al. 2008. Large scale variation in Enterococcus faecalis
illustrated by the genome analysis of strain OG1RF. Genome Biol. 9:R110.
4. 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.
5. Brock, T. D., B. Peacher, and D. Pierson. 1963. Survey of the bacteriocines
of enterococci. J. Bacteriol. 86:702–707.
6. Chowdhury, S. A., et al. 2009. A trilocus sequence typing scheme for hospital
epidemiology and subspecies differentiation of an important nosocomial
pathogen, Enterococcus faecalis. J. Clin. Microbiol. 47:2713–2719.
7. Coque, T. M., J. E. Patterson, J. M. Steckelberg, and B. E. Murray. 1995.
Incidence of hemolysin, gelatinase, and aggregation substance among en-
terococci isolated from patients with endocarditis and other infections and
from feces of hospitalized and community-based persons. J. Infect. Dis.
8. Dunny, G. M., B. L. Brown, and D. B. Clewell. 1978. Induced cell aggregation
and mating in Streptococcus faecalis: evidence for a bacterial sex pheromone.
Proc. Natl. Acad. Sci. U. S. A. 75:3479–3483.
9. Eaton, T. J., and M. J. Gasson. 2001. Molecular screening of Enterococcus
virulence determinants and potential for genetic exchange between food and
medical isolates. Appl. Environ. Microbiol. 67:1628–1635.
10. 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:
11. Franke, A. E., and D. B. Clewell. 1981. Evidence for a chromosome-borne
resistance transposon (Tn916) in Streptococcus faecalis that is capable of
“conjugal” transfer in the absence of a conjugative plasmid. J. Bacteriol.
12. Gardner, P., D. H. Smith, H. Beer, and R. C. Moellering, Jr. 1969. Recovery
of resistance (R) factors from a drug-free community. Lancet ii:774–776.
13. Gilmore, M. S., and J. J. Ferretti. 2003. Microbiology. The thin line between
gut commensal and pathogen. Science 299:1999–2002.
14. Gold, O. G., H. V. Jordan, and J. van Houte. 1975. The prevalence of
enterococci in the human mouth and their pathogenicity in animal models.
Arch. Oral Biol. 20:473–477.
15. Jacob, A. E., and S. J. Hobbs. 1974. Conjugal transfer of plasmid-borne
multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bac-
16. Jha, A. K., H. P. Bais, and J. M. Vivanco. 2005. Enterococcus faecalis mam-
malian virulence-related factors exhibit potent pathogenicity in the Arabi-
dopsis thaliana plant model. Infect. Immun. 73:464–475.
17. Ji, G., R. Beavis, and R. P. Novick. 1997. Bacterial interference caused by
autoinducing peptide variants. Science 276:2027–2030.
18. Kristich, C. J., J. R. Chandler, and G. M. Dunny. 2007. Development of a
host-genotype-independent counterselectable marker and a high-frequency
conjugative delivery system and their use in genetic analysis of Enterococcus
faecalis. Plasmid 57:131–144.
19. LeBlanc, D. J., and L. N. Lee. 1982. Characterization of two tetracycline
resistance determinants in Streptococcus faecalis JH1. J. Bacteriol. 150:835–
20. Maekawa, S., M. Yoshioka, and Y. Kumamoto. 1992. Proposal of a new
scheme for the serological typing of Enterococcus faecalis strains. Microbiol.
21. Manson, J. M., L. E. Hancock, and M. S. Gilmore. 2010. Mechanism of
chromosomal transfer of Enterococcus faecalis pathogenicity island, capsule,
antimicrobial resistance, and other traits. Proc. Natl. Acad. Sci. U. S. A.
22. Manson, J. M., S. Keis, J. M. Smith, and G. M. Cook. 2003. Characterization
of a vancomycin-resistant Enterococcus faecalis (VREF) isolate from a dog
with mastitis: further evidence of a clonal lineage of VREF in New Zealand.
J. Clin. Microbiol. 41:3331–3333.
23. McBride, S. M., V. A. Fischetti, D. J. Leblanc, R. C. Moellering, Jr., and
M. S. Gilmore. 2007. Genetic diversity among Enterococcus faecalis. PLoS
24. Moellering, R. C., Jr., and A. N. Weinberg. 1971. Studies on antibiotic
syngerism against enterococci. II. Effect of various antibiotics on the uptake
of14C-labeled streptomycin by enterococci. J. Clin. Invest. 50:2580–2584.
25. Mohamed, J. A., W. Huang, S. R. Nallapareddy, F. Teng, and B. E. Murray.
2004. Influence of origin of isolates, especially endocarditis isolates, and
various genes on biofilm formation by Enterococcus faecalis. Infect. Immun.
26. Murray, B. E. 1990. The life and times of the Enterococcus. Clin. Microbiol.
27. Murray, B. E., and B. Mederski-Samaroj. 1983. Transferable beta-lac-
tamase. A new mechanism for in vitro penicillin resistance in Streptococcus
faecalis. J. Clin. Invest. 72:1168–1171.
28. Mylonakis, E., et al. 2002. The Enterococcus faecalis fsrB gene, a key com-
ponent of the fsr quorum-sensing system, is associated with virulence in the
rabbit endophthalmitis model. Infect. Immun. 70:4678–4681.
29. Nakayama, J., et al. 2001. Gelatinase biosynthesis-activating pheromone: a
peptide lactone that mediates a quorum sensing in Enterococcus faecalis.
Mol. Microbiol. 41:145–154.
30. Nakayama, J., et al. 2006. Revised model for Enterococcus faecalis fsr quo-
rum-sensing system: the small open reading frame fsrD encodes the gelatin-
ase biosynthesis-activating pheromone propeptide corresponding to staphy-
lococcal AgrD. J. Bacteriol. 188:8321–8326.
31. Nakayama, J., R. Kariyama, and H. Kumon. 2002. Description of a 23.9-
kilobase chromosomal deletion containing a region encoding fsr genes which
mainly determines the gelatinase-negative phenotype of clinical isolates of
Enterococcus faecalis in urine. Appl. Environ. Microbiol. 68:3152–3155.
32. Nallapareddy, S. R., et al. 2006. Endocarditis and biofilm-associated pili of
Enterococcus faecalis. J. Clin. Invest. 116:2799–2807.
33. Novick, R. P. 2003. Autoinduction and signal transduction in the regulation
of staphylococcal virulence. Mol. Microbiol. 48:1429–1449.
34. Oprea, S. F., et al. 2004. Molecular and clinical epidemiology of vancomycin-
resistant Enterococcus faecalis. J. Antimicrob. Chemother. 53:626–630.
35. Paulsen, I. T., et al. 2003. Role of mobile DNA in the evolution of vanco-
mycin-resistant Enterococcus faecalis. Science 299:2071–2074.
36. 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.
37. Qin, X., K. V. Singh, G. M. Weinstock, and B. E. Murray. 2000. Effects of
Enterococcus faecalis fsr genes on production of gelatinase and a serine
protease and virulence. Infect. Immun. 68:2579–2586.
38. Rice, L. B., and L. L. Carias. 1998. Transfer of Tn5385, a composite, mul-
tiresistance chromosomal element from Enterococcus faecalis. J. Bacteriol.
39. Rice, L. B., et al. 1991. Chromosomally mediated beta-lactamase production
and gentamicin resistance in Enterococcus faecalis. Antimicrob. Agents Che-
40. Roberts, J. C., K. V. Singh, P. C. Okhuysen, and B. E. Murray. 2004.
Molecular epidemiology of the fsr locus and of gelatinase production among
different subsets of Enterococcus faecalis isolates. J. Clin. Microbiol. 42:2317–
41. Ruiz-Garbajosa, P., et al. 2006. Multilocus sequence typing scheme for
Enterococcus faecalis reveals hospital-adapted genetic complexes in a back-
ground of high rates of recombination. J. Clin. Microbiol. 44:2220–2228.
42. Sahm, D. F., et al. 1989. In vitro susceptibility studies of vancomycin-resistant
Enterococcus faecalis. Antimicrob. Agents Chemother. 33:1588–1591.
43. Sharpe, M. E. 1964. Serological types of Streptococcus faecalis and its vari-
eties and their cell wall type antigen. J. Gen. Microbiol. 36:151–160.
44. Sifri, C. D., et al. 2002. Virulence effect of Enterococcus faecalis protease
genes and the quorum-sensing locus fsr in Caenorhabditis elegans and mice.
Infect. Immun. 70:5647–5650.
45. 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.
46. Willems, R. J., et al. 2001. Variant esp gene as a marker of a distinct genetic
lineage of vancomycin-resistant Enterococcus faecium spreading in hospitals.
VOL. 77, 2011CONSERVED DELETION OF fsr GENES451