JOURNAL OF BACTERIOLOGY, July 2011, p. 3453–3460
Copyright © 2011, American Society for Microbiology. All Rights Reserved.
Vol. 193, No. 14
Type 3 Fimbriae and Biofilm Formation Are Regulated by the
Transcriptional Regulators MrkHI in Klebsiella pneumoniae?
Jeremiah G. Johnson, Caitlin N. Murphy, Jean Sippy, Tylor J. Johnson, and Steven Clegg*
Department of Microbiology, Roy J. and Lucille A. Carver College of Medicine, University of Iowa, Iowa City, Iowa 52242
Received 28 February 2011/Accepted 5 May 2011
Klebsiella pneumoniae is an opportunistic pathogen which frequently causes hospital-acquired urinary and
respiratory tract infections. K. pneumoniae may establish these infections in vivo following adherence, using the
type 3 fimbriae, to indwelling devices coated with extracellular matrix components. Using a colony immunoblot
screen, we identified transposon insertion mutants which were deficient for type 3 fimbrial surface production.
One of these mutants possessed a transposon insertion within a gene, designated mrkI, encoding a putative
transcriptional regulator. A site-directed mutant of this gene was constructed and shown to be deficient for
fimbrial surface expression under aerobic conditions. MrkI mutants have a significantly decreased ability to
form biofilms on both abiotic and extracellular matrix-coated surfaces. This gene was found to be cotran-
scribed with a gene predicted to encode a PilZ domain-containing protein, designated MrkH. This protein was
found to bind cyclic-di-GMP (c-di-GMP) and regulate type 3 fimbrial expression.
Klebsiella pneumoniae is an opportunistic pathogen which is
a significant cause of nosocomially acquired infections, includ-
ing catheter-associated urinary tract infections (CAUTIs) and
ventilator-associated pneumonias (5, 15, 28). K. pneumoniae
type 3 fimbriae mediate attachment to, and biofilm formation
on, extracellular matrix-coated surfaces in vitro (2, 13, 16). In
vivo, indwelling devices rapidly become coated with host ma-
terial, creating an environment that facilitates infection by type
3 fimbria-producing enterobacteria. In addition, most isolates
of K. pneumoniae causing nosocomially acquired infections are
resistant to multiple antibiotics (22, 23). The ability of K. pneu-
moniae to form biofilms and the antimicrobial resistance of the
organism are factors that make infections by these bacteria
very difficult to eradicate.
K. pneumoniae possesses several virulence factors which aid
in the ability of the organism to persist and thrive within an
animal host. One class of virulence factors, those involved in
bacterial adherence, includes the type 1 and type 3 fimbrial
adhesins. Previously, the type 1 fimbriae in K. pneumoniae have
been shown to play a role in infectivity by using a murine
bladder cystitis model in which a type 1 hyperfimbriate strain
more readily forms intracellular bacterial communities (IBCs)
within bladder umbrella cells (29). The type 3 fimbriae have
previously been characterized by our group and have been
shown to mediate bacterial adherence in vitro to human extra-
cellular matrix (HECM) components (13). In many strains of
K. pneumoniae, this fimbrial type is encoded by a chromo-
somally borne gene cluster previously shown to be comprised
of five genes (Fig. 1A). These genes include determinants
encoding the major fimbrial subunit (MrkA), a chaperone-
usher system (MrkBC, respectively), the fimbrial tip adhesin
(MrkD), and an as-yet-uncharacterized structural component
(MrkF) (6, 8).
Like many other fimbrial types in enterobacteria, the type 3
fimbriae are assembled by the chaperone/usher pathway (8).
However, the regulation of type 3 fimbrial expression in K.
pneumoniae is poorly understood in comparison to mecha-
nisms of regulation of type 1 and pap fimbrial expression in
Escherichia coli and type 1 fimbrial production in Salmonella.
Our group recently reported that type 3 fimbrial production is
affected by the intracellular concentrations of the second
messenger cyclic-di-GMP (c-di-GMP) (14). Mutants of the
K. pneumoniae phosphodiesterase MrkJ accumulated intra-
cellular c-di-GMP, which resulted in a type 3 hyperfimbriate
phenotype that more readily formed biofilms (14). Fimbrial
systems often employ complex regulatory circuits, and it is
expected that several as-yet-unidentified regulators govern the
expression of type 3 fimbriae. Here we describe a screen which
identified a mutant within a putative transcriptional regulator,
MrkI, that resulted in decreased type 3 fimbrial expression and
biofilm formation. Additionally, we have identified a determi-
nant which is cotranscribed with mrkI and encodes a PilZ-
domain containing protein, MrkH, which was shown to bind
c-di-GMP and affect type 3 fimbrial expression.
MATERIALS AND METHODS
Strains, plasmids, and DNA manipulations. The strains and plasmids used in
this study are shown in Table 1. To detect the presence of type 3 fimbriae, all
strains were grown on either glycerol-Casamino Acids (GCAA) or Luria-Bertani
(LB) medium at 37°C unless otherwise stated (6, 7, 11, 16). When necessary,
strains were cultured in medium supplemented with antibiotics at the following
concentrations: ampicillin (Amp; 100 ?g/ml), kanamycin (Kan; 25 ?g/ml), spec-
tinomycin (Spec; 100 ?g/ml), and tetracycline (Tet; 25 ?g/ml).
Plasmid and genomic DNA preparations, restriction enzyme digestions, and
PCR procedures were performed by conventional techniques using commercially
available material. All manipulations of DNA were performed according to the
Construction and screening of mini-Tn5 transposon library. Conjugation
of K. pneumoniae IApc35 with E. coli S17-1 ?pir carrying the plasmid
pUTminiTn5-Kn was performed as previously described by our group (2).
Conjugants were selected on M9 minimal medium supplemented with kanamy-
* Corresponding author. Mailing address: Department of Microbi-
ology, University of Iowa, 3-403 BSB, 51 Newton Rd., Iowa City, IA
52242. Phone: (319) 335-7778. Fax: (319) 335-9006. E-mail: steven
?Published ahead of print on 13 May 2011.
cin to prevent growth of both the donor and recipient strains. Subsequently,
appropriate dilutions of bacterial suspensions in phosphate-buffered saline
(PBS) were plated on M9 minimal medium and incubated overnight at 37°C.
Bacterial colonies were screened for the production of surface-associated type 3
fimbriae using conventional immunoblotting techniques and monospecific fim-
brial antiserum at a dilution of 1:40,000 and for subsequent development with
goat anti-rabbit serum conjugated to alkaline phosphatase (4, 20). All colonies
that did not react with the fimbrial serum were isolated, retested for lack of
reactivity with fimbria-specific antiserum, and stored at ?80°C. Insertions of the
mini-Tn5 into genes encoding the structural and assembly components of the
mrkABCDF cluster were identified by standard PCR procedures and not exam-
Mapping of the mini-Tn5 insertion site. Genomic DNA was isolated from
nonfimbrial mutants, restricted with SphI, and ligated into SphI-digested
pACYC184. The nucleotide sequence of the inserted DNA fragment was deter-
mined, and K. pneumoniae-derived sequences flanking the transposon were iden-
tified. Subsequently, the location of these sequences on the K. pneumoniae
genome was identified using the genome sequence of K. pneumoniae MGH
FIG. 1. (A) Genetic organization of the mrk gene cluster. Putative promoter regions have been identified by sequence analysis and are indicated
by arrows. (B) Predicted domain architecture of MrkI. The site of the mini-Tn5 insertion is within the predicted LuxR-like DNA binding domain
(amino acids 130 to 176) in the C-terminal region of the 190-amino-acid MrkI polypeptide. (C) The location of the predicted PilZ c-di-GMP
binding site lies within the C-terminal region (amino acids 107 to 225).
TABLE 1. Plasmids, strains, and oligonucleotides used in this study
Strain/plasmid/oligonucleotide Description/sequence (5?–3?)a
Plasmid-cured variant of IA565, type 3 fimbria positive
Kanr; IApc35 mrkI insertion mutant, type 3 fimbria negative
IApc35 mrkHI deletion mutant, type 3 fimbria negative
Protein expression strain
General E. coli cloning strain
E. coli donor strain
E. coli donor strain
Invitrogen (Carlsbad, CA)
NEB (Ipswich, MA)
TetrCams; empty vector control for complementation constructs
Tetr; mrkH complementation vector
Tetr; mrkI complementation vector
Tetr; mrkHI complementation vector
Tetr; mutated mrkHI complementation vector
Ampr; 6?His-tagged Gateway expression vector
Ampr; 6?His-tagged MrkH expression construct
Ampr; 6?His-tagged MrkH(R113A) expression construct
Camr; sacB suicide vector
CamrKanr; construct used to make IApc35 mrkI::Knr
CamrSpecr; construct used to make IApc35 ?mrkHI
Ampr; subcloning vector
Ampr; general subcloning vector
Promega (Madison, WI)
Ampr; pTrc99A-based reporter construct
AmprKanr; mini-Tn5 delivery plasmid
aCam, chloramphenicol. Underlining indicates the site of substitution.
3454JOHNSON ET AL. J. BACTERIOL.
78578, available online (http://genome.wustl.edu/genomes). The nucleotide se-
quences flanking the mini-Tn5 were determined in mutants and were found not
to have resulted in large rearrangements of the DNA during transposition.
Construction of defined site-directed mutants. Approximately 1-kb regions of
DNA flanking the mrkI and the mrkHI genes were cloned into the vector
pGEM-T Easy. Fragments were ligated together, incorporating an internal XbaI
restriction site, into which a kanamycin resistance determinant was introduced
for the construction of the MrkI mutant only. K. pneumoniae-derived DNA was
excised from the pGEM-T Easy recombinant plasmids using SacI and SphI.
These fragments were ligated into either the suicide vector pDS132 (for mrkI) or
pDS132-specR(for mrkHI). The resulting plasmids, pDS132mrkI::knRand
pDS132?mrkHI, were transformed into the permissive host E. coli SM10 ?pir
and subsequently introduced into K. pneumoniae IApc35 via conjugation.
Transconjugants were selected on either LB-Kan/Amp or LB-Spec/Amp plates,
followed by counterselection on 5% sucrose plates (17, 26). Characterization of
mrkI insertion or mrkHI deletion mutants was performed using standard PCR
Detection of type 3 fimbriae. Surface production of fimbrial appendages was
detected using monospecific fimbrial antiserum as described elsewhere by our
group (11, 14). Aerobic cultures were grown at 37°C overnight on either LB agar
or as 25-ml LB cultures grown in a 125-ml flask shaken at 220 rpm. Anaerobic
and microaerophilic cultures were grown either on LB agar in anaerobic Bio-Bag
type A environmental chambers (Becton-Dickinson, Sparks, MD) or as static LB
broth cultures, respectively. When necessary, fimbriae were observed by trans-
mission electron microscopy using formaldehyde-fixed bacteria stained with
uranyl acetate as previously described (29).
Transcription of mrk. Expression of mrk genes in K. pneumoniae strains grown
under aerobic or anaerobic conditions was determined by quantitative reverse
transcription-PCR (qRT-PCR) as previously described (14). Comparison of gene
expression between strains grown aerobically and anaerobically was done fol-
lowing cDNA synthesis from equal concentrations of total cellular RNA. Also,
the cloned mrk genes in the K. pneumoniae IApc35 ?mrkHI mutant were assayed
for mrkA expression under aerobic conditions using qRT-PCR.
In addition, the ability of the cloned mrkH, mrkI, and mrkHI genes and their
derivatives to affect expression of mrkA was determined using a plasmid-borne
reporter fusion, pTrc99APmrkA-lacZ, in an E. coli host. This fusion was con-
structed by cloning a XbaI/HindIII-tailed 444-bp fragment of DNA immediately
upstream of mrkA, and possessing the promoter region, into those respective
sites in pTrc99A containing a promoterless lacZ gene.
Biofilm formation assays. The ability of K. pneumoniae strains to form biofilms
on solid surfaces was determined as previously described (14, 24, 25).
Mutation of the MrkH c-di-GMP binding site. Arginine-113 of MrkH was
mutated to alanine using overlapping oligonucleotides. Using primers CNM003
and CNM004 which contained the desired mutation and pACYCmrkHI as tem-
plate, the FailSafe PCR enzyme mix (Epicentre, Madison, WI) was used with 18
cycles of the following reaction: 95°C for 30 s, 55°C for 1 min, and 68°C for 5 min.
The resulting plasmid PCR product was digested with DpnI for 1 h at 37°C and
then transformed into chemically competent E. coli DH5? (Invitrogen, Carls-
bad, CA). The appropriate mutation in the resulting plasmid and the absence
of any additional mutations in pACYCmrkHR113AmrkI were verified by DNA
Purification of MrkH and MrkH(R113A). The K. pneumoniae IApc35
mrkH and mrkH(R113A) genes were amplified from pACYCmrkHI and
pACYCmrkHR113AmrkI, respectively, by standard PCR procedures and cloned
into the Gateway vector pENTR/D-Topo (Invitrogen). These genes were sub-
sequently integrated into the Gateway-compatible destination vector pDEST17
to construct genes encoding His-tagged fusion proteins and introduced into the
expression strain BL21-AI. E. coli BL21-AI transformants carrying either
pDEST17mrkH or pDEST17mrkHR113Awere used to prepare native MrkH or
MrkH(R113A), respectively, by Ni-nitrilotriacetic acid (NTA) affinity chroma-
tography by following the manufacturer’s instructions (Qiagen, Valencia, CA).
Successful purification of both MrkH and MrkH(R113A) was assessed by
12% SDS-PAGE and Western blotting using anti-6?His antibody (Qiagen).
Additionally, MrkI was purified as a maltose binding protein (MBP)-fusion
gene product (MBP-MrkI) using a commercially available system (NEB,
Binding of c-di-GMP to MrkH. Generation of
performed as previously described (9, 10). The c-di-GMP binding assay was
based on that described by Hickman and Harwood (9). A 20-?l mixture of 0.2
mM protein and 2.0 ?M [32P]c-di-GMP in binding buffer (40 mM Tris [pH 7.8],
10 mM magnesium acetate, 50 mM KCl) was incubated on ice for 25 min. The
reaction mixtures were then brought to a 100-?l volume with binding buffer and
immediately loaded onto a slot blot apparatus (PR 600 slot blot; Hoefer Scien-
32P-labeled c-di-GMP was
tific) containing a 0.2-?M nitrocellulose membrane (0.45 mM Protran BA85;
Whatman), followed by a wash using 1.0 ml cold binding buffer. The membrane
was removed and scanned on a phosphorimager (Packard Instant Imager; Pack-
ard Instrument Company) to measure radioactive counts of membrane-bound
[32P]c-di-GMP. For the competition assay, a 10-fold excess (20 ?M) of cold
c-di-GMP was added to the reaction mixture (Biolog, Bremen, Germany). In
additional experiments, an equal amount of [?-32P]GTP was substituted for
[32P]c-di-GMP to further assess MrkH binding specificity. Additionally, MBP-
MrkI was examined for the ability to bind [32P]c-di-GMP. Reaction mixtures
containing purified LacZ? or protein buffer alone were used as controls.
Immunoblotting of a mini-Tn5 transposon bank of insertion
mutants in K. pneumoniae IApc35. More than 21,000 insertion
mutants were screened for their ability to produce surface-
associated type 3 fimbriae. Of these, 11 (0.05%) mutants con-
sistently failed to react with fimbria-specific antiserum, even
after growth on GCAA agar, which favors the phenotypic ex-
pression of these fimbriae (7). Following mapping of the in-
sertion site of the mini-Tn5, three of the mutants were shown
to have the transposon inserted into genes that are part of the
previously described mrk gene cluster. Therefore, eight muta-
tions in genes that do not encode either structural or assembly
components of the type 3 fimbrial system were isolated. The
insertion sites of these mutants are currently being identified,
and one of these is described below.
K. pneumoniae IApc35 MrkI and MrkHI mutants do not
produce surface-associated fimbriae. Of the eight mutants iso-
lated that possess the transposon in a gene distinct from the
mrkABCDF cluster, one of these was further characterized.
The site of insertion in this mutant was found to be in a gene
encoding a putative transcriptional regulator and annotated
KPN_03273 on the K. pneumoniae MGH 78578 genome
(GenBank accession number CP000647). The predicted size
of this gene is 573 bp, encoding a gene product of 190 amino
acids. BLAST analyses of this gene product suggested that it
belongs to a family of regulators characterized by a LuxR-like
DNA binding domain spanning amino acids 130 to 176 in its
C-terminal region (Fig. 1B). The precise site of insertion of the
mini-Tn5 was within the predicted DNA binding region encod-
ing amino acid 151. We previously named this gene mrkI,
which is located between mrkH and mrkJ, though the K. pneu-
moniae MGH 78578 genome lacks the correct annotation for
mrkH (14). The mrkHIJ genes are located adjacent to the
previously characterized mrk gene cluster and exhibit opposite
transcriptional polarity to these genes (Fig. 1A). Using inter-
genic RT-PCR, it was found that mrkH, mrkI, and mrkJ are
cotranscribed (data not shown). mrkH is predicted to encode a
protein containing a PilZ c-di-GMP binding domain at its C
terminus and an N terminus that exhibits little homology to
currently characterized domains (Fig. 1C).
MrkI and MrkHI mutants of K. pneumoniae IApc35 were
constructed by conventional techniques. Neither of the K.
pneumoniae IApc35 mrkI::Knrand IApc35 ?mrkHI mutants
produce surface-associated type 3 fimbriae following growth
under aerobic conditions (Table 2). Interestingly, the MrkI
mutant does express type 3 fimbriae when grown anaerobically
or microaerophilically as either agar or static broth cultures,
respectively, while the ?mrkHI mutant remains nonfimbriate
under either condition (Table 2). The mutants can be comple-
mented to restore fimbrial production by transformation with
VOL. 193, 2011MrkHI AND TYPE 3 FIMBRIAL EXPRESSION3455
the cloned genes (Table 2). Electron microscopy confirmed the
absence of fimbriae on the MrkI and MrkHI mutants and
many fimbriae on the surfaces of complemented strains (Fig.
2). Interestingly, in a ?mrkHI background, introduction of a
plasmid solely expressing mrkI was unable to complement fim-
brial expression, while a plasmid expressing only mrkH was
able to restore type 3 fimbriation (Table 3). Also, overexpres-
sion of mrkH in an mrkI mutant background was able to re-
store fimbrial expression despite the absence of mrkI (Table 3).
MrkI and MrkHI mutants are affected in mrkA gene tran-
scription. Using qRT-PCR analysis of RNA extracted from
aerobically grown agar cultures, it was found that both IApc35
mrkI::Knrand IApc35 ?mrkHI have significantly reduced lev-
els of mrkA transcription. Levels of the mrkA transcript in the
IApc35 mrkI::Knrstrain were approximately 20-fold lower than
those in IApc35. Similarly, using RNA extracted from IApc35
?mrkHI, a significant decrease in mrkA expression was ob-
served, with a 33-fold reduction in transcription compared to
that of the parental strain (Fig. 3A). Since MrkI mutants as-
semble surface-associated fimbriae when grown anaerobically,
mrkA expression under these conditions was determined. K.
pneumoniae IApc35 mrkI::Knrgrown anaerobically exhibited
levels of mrkA expression that were indistinguishable from
those of the parental IApc35 strain. mrkA transcription levels
in the IApc35 ?mrkHI mutant were significantly lower than
those of the parental strain and were reduced by approximately
2,000-fold (Fig. 3B). Also, we examined a possible autoregu-
latory role of MrkI on mrkHI transcription and observed no
decrease in gene expression in the MrkI mutant (data not
Expression of mrk genes is increased following anaerobic
growth. Quantitative RT-PCR analysis using RNA extracted
from K. pneumoniae IApc35 cultures grown anaerobically in-
dicated increased mrkA, mrkH, and mrkI expression compared
to that of cultures incubated aerobically. Increased expression
levels of approximately 285-, 77-, and 91-fold were observed
for mrkA, mrkH, and mrkI, respectively (Table 4).
MrkI and MrkHI mutants have a decreased ability to form
biofilms on an abiotic surface. Using crystal violet plate assays,
it was shown that K. pneumoniae IApc35 mrkI::Knrhas a de-
creased ability to form a biofilm on plastic surfaces compared
to that of the parental strain. When the cloned mrkI gene was
reintroduced into the IApc35 mrkI::Knrstrain, full restoration
of biofilm formation was observed, as shown in Fig. 4A. Also,
using these assays, the IApc35 mrkI::Knrmutant transformed
with an empty vector had a significantly decreased ability to
form biofilms, whereas the parental IApc35 strain carrying the
same plasmid is a biofilm producer. Likewise, as shown in Fig.
4B, the mrkHI deletion mutant exhibited a significantly re-
duced (approximately 7-fold) ability to form a biofilm com-
pared to that exhibited by the parental strain. Restoration of
biofilm formation was achieved by complementation with the
cloned mrkHI determinants, and such transformants also ex-
hibited an increased ability to form biofilms compared to that
exhibited by the parental strain (approximately 2-fold). Trans-
formants of the MrkHI mutant possessing an empty cloning
vector did not demonstrate biofilm formation.
Mutation of a conserved PilZ residue results in the inability
to induce type 3 fimbria production. Alignment of the PilZ
domain in MrkH with other PilZ domain-containing proteins
revealed complete conservation of five residues which have
previously been shown to be important in the ability of the PilZ
domain to bind c-di-GMP (27) (Fig. 5A). Replacement of the
TABLE 3. Phenotypic complementation analyses
Strain (plasmid) Serum titer
IApc35 (pACYC?Cmr)................................................................ 5,120
IApc35 mrkI::Knr(pACYCmrkH) ..............................................40,960
IApc35 ?mrkHI (pACYC?Cmr).................................................
IApc35 ?mrkHI (pACYCmrkH).................................................40,960
IApc35 ?mrkHI (pACYCmrkI)...................................................
IApc35 ?mrkHI (pACYCmrkHI)................................................40,960
TABLE 2. Type 3 fimbrial production of Klebsiella strains
IApc35 mrkI::Knrplus mrkI
IApc35 ?mrkHIIApc35 ?mrkHI plus mrkHI
Aerobic (shaken flask)
Aerobic (agar grown)
Microaerophilic (static tube)
Anaerobic (agar grown)
aThe serum titer represents the reciprocal of the anti-MrkA serum dilution needed to produce visible agglutination. The lowest dilution of serum used was 1:40. ND,
FIG. 2. Fimbrial production by K. pneumoniae strains. (A) K. pneu-
moniae IApc35; (B) MrkI mutant; (C) MrkI mutant transformed with
the cloned mrkI gene; (D) MrkHI mutant; (E) complemented MrkHI
mutant carrying cloned mrkHI.
3456JOHNSON ET AL.J. BACTERIOL.
conserved arginine-113 with alanine on a plasmid-borne copy
of mrkH was performed and confirmed by nucleotide sequenc-
ing. Introduction of this plasmid, pACYCmrkHR113AmrkI, into
IApc35 ?mrkHI did not restore type 3 fimbrial expression
compared to that observed using the parental pACYCmrkHI
plasmid (Fig. 5B). This significant reduction in type 3 fimbrial
production was further investigated using the reporter plasmid
pTrc99APmrkA-lacZ in an E. coli background. When the pa-
rental plasmid pACYCmrkHI was introduced into this E. coli
strain, an increase in ?-galactosidase production was observed
compared to that observed in the strain carrying the empty
vector (pACYC184?Cmr), as shown in Fig. 5C. In contrast,
when pACYCmrkHR113AmrkI transformants were assayed, a
significant decrease in ?-galactosidase activity was observed
compared to that of the parental plasmid (Fig. 5C).
MrkH binds c-di-GMP. To determine whether MrkH is
capable of binding, c-di-GMP filter binding assays were used as
previously described (9). Purification of 6?His-tagged MrkH
and MrkH(R113A) was performed, and protein purity was
determined by Western blot analysis. Equimolar amounts of
MrkH proteins immobilized on nitrocellulose were used in
the c-di-GMP binding assays. Bound [32P]c-di-GMP was de-
termined by phosphorimaging, and those results are shown in
Fig. 6. MrkH bound radiolabeled c-di-GMP, whereas the
MrkH(R113A) protein did not (Fig. 6A). Inhibition of MrkH-
mediated binding of labeled c-di-GMP was achieved by com-
petition with an unlabeled nucleotide (Fig. 6B). Additionally,
binding assays were also performed using [?-32P]GTP, but
neither MrkH nor MrkH(R113A) bound this nucleotide (data
MrkH and MrkHI activate the mrkA promoter. To examine
whether MrkH, MrkI, or MrkHI were sufficient to induce
transcription from the mrkA promoter, ?-galactosidase assays
were used. Plasmids comprised of the same vector backbone,
carrying either mrkH or mrkI alone or mrkHI together, were
introduced into E. coli NEB 5-? transformed with a PmrkA-lacZ
reporter fusion. The strain carrying both the reporter fusion
and mrkH alone was found to exhibit a significant increase
(approximately 114-fold) in transcriptional activity from the
mrkA promoter compared to that exhibited by a transformant
possessing the cloning vector alone. When mrkI alone was
introduced into the strain carrying the reporter plasmid, no
increase in ?-galactosidase production was seen compared to
that exhibited by transformants without mrkI. When a plasmid
carrying both mrkH and mrkI was transformed into the re-
porter strain, a significant increase in mrkA transcription, com-
pared to that exhibited by the strain carrying mrkH alone, was
observed (approximately 8-fold) (Fig. 7).
Similarly qRT-PCR analysis of K. pneumoniae IApc35
?mrkHI transformed with the mrkI, mrkH, and mrkHI genes
also indicated that MrkH but not MrkI could affect mrkA
expression. The MrkHI mutant transformed with mrkI alone
did not exhibit any increase in mrkA expression compared to
that exhibited by mutants transformed with the empty vector.
However, transformation with a plasmid bearing the mrkHI
FIG. 3. qRT-PCR of mrkA encoding the major fimbrial subunit in
K. pneumoniae strains. Aerobic (A) and anaerobic (B) mrkA transcrip-
tion in both IApc35 mrkI::Knrand IApc35 ?mrkHI are shown as the
relative decreases in transcription compared to that shown by the
parental strain. Statistical significance was determined using Student’s
t test (???, P value ? 0.001; ????, P value ? 0.0001).
TABLE 4. Aerobic and anaerobic transcriptions of mrk
genes in parental IApc35
Transcript (condition)Fold change (? SD)
mrkA (anaerobic).........................................................284.78 (? 47.75)a
mrkH (anaerobic)......................................................... 76.93 (? 24.45)a
mrkI (aerobic) ..............................................................
mrkI (anaerobic).......................................................... 90.65 (? 12.34)a
1.0 (? 0.1)
1.0 (? 0.15)
aP value ? 0.0001.
FIG. 4. Biofilm phenotypes of K. pneumoniae strains. (A) Biofilm
formation of parental IApc35 and the MrkI mutant carrying the empty
vector (VC) compared to that of the complemented MrkI mutant on
an abiotic surface. (B) Decreased biofilm formation of the IApc35
MrkHI mutant compared to those of parental IApc35 and the MrkHI
mutant complemented with plasmid-borne mrkHI. Statistical signifi-
cance was determined using Student’s t test (?, P value ? 0.05; ??, P
value ? 0.01; ???, P value ? 0.001).
VOL. 193, 2011MrkHI AND TYPE 3 FIMBRIAL EXPRESSION3457
genes resulted in a 36-fold increase in mrkA transcription com-
pared to that seen with MrkH alone (P ? 0.001).
K. pneumoniae type 3 fimbriae play an important role in the
ability of the bacteria to bind to, and subsequently form bio-
films on, HECM-coated surfaces. Both the fimbrial adhesin
(MrkD) and the polymerized fimbrial shaft protein (MrkA)
play important roles in this function (12, 13, 16, 30). We have
previously proposed that MrkD facilitates the adherence of the
organism to specific collagen molecules that form part of the
HECM. However, fimbriate bacteria that possess no functional
adhesin also form biofilms on abiotic surfaces (12, 14, 16).
Consequently, the production of type 3 fimbriae could lead to
the initiation of biofilm formation on inserted devices such as
catheters shortly after insertion and also after these devices
become coated in situ with host factors. The genetic regulation
FIG. 5. Analysis of the R113A mutation in MrkH. (A) Alignment of previously characterized PilZ domain-containing proteins with MrkH.
Conserved residues shown to affect c-di-GMP binding are indicated with stars. The alignment of the R113A mutant is also indicated ( ). (B) Effect
of the R113A mutant on type 3 fimbria production. Values are reciprocals of serum titers needed to cause visible agglutination. (C) Use of a
PmrkA-lacZ fusion to examine the ability of MrkHI to induce expression in E. coli transformants. Statistical significance was determined using
Student’s t test (????, P value ? 0.0001).
FIG. 6. Ability of MrkH to bind [32P]c-di-GMP. (A) Filter binding
assay using purified LacZ?, MrkH, and MrkH(R113A) proteins as
targets for binding. Graph represents total specific counts detected
from assays represented above the graph. (B) Filter binding assays of
binding reactions with (?) or without (?) the addition of unlabeled
FIG. 7. Ability of cloned mrkH, mrkI, and mrkHI to induce tran-
scription of a PmrkA-lacZ reporter in E. coli compared to that of a
vector control (VC). Statistical significance was determined using Stu-
dent’s t test (????, P value ? 0.0001).
3458 JOHNSON ET AL.J. BACTERIOL.
of mrk gene expression is poorly understood but, like other
enterobacterial fimbrial systems, is likely to involve complex
In order to identify regulatory elements of the type 3 fim-
brial operon, we constructed a mini-Tn5 transposon library in
K. pneumoniae IApc35. This strain is a plasmid-cured deriva-
tive of the clinical isolate K. pneumoniae IA565, produces high
levels of type 3 fimbriae, and forms robust biofilms on abiotic
surfaces (12). It possesses only one chromosomally borne copy
of the mrk gene cluster. One nonfimbriate mutant from this
library possessed a transposon insertion within a gene encod-
ing a putative transcriptional regulator, which we have previ-
ously termed mrkI (14). This gene is predicted to encode, by
comparison to families of functional proteins, a protein which
contains only one identifiable domain, a LuxR-like DNA bind-
ing domain in its C-terminal region. The N-terminal region of
MrkI exhibits little relatedness to any characterized protein
domains and therefore has no readily identifiable receiver do-
main. Interestingly, mrkI is located between two genes, as
follows: the first, which we have named mrkH, is predicted to
encode a protein which contains a C-terminal c-di-GMP bind-
ing domain (PilZ), and the second, mrkJ, is a gene which we
have previously shown to produce a functional phosphodies-
terase which modulates the intracellular levels of c-di-GMP
within K. pneumoniae (14). A defined MrkI mutant of strain
IApc35 was constructed and, like the original transposon mu-
tant, was found to be unable to assemble type 3 fimbriae. Also,
we found that mrkI is cotranscribed with mrkH. This is consis-
tent with the observation that the only promoter identified by
sequence analysis, which is likely to drive mrkI transcription,
lies upstream of mrkH. In addition to mrkHI cotranscription,
we also found that mrkJ transcription can also occur from the
mrkH promoter, though the mrkI and mrkHI mutations were
not found to significantly alter the levels of mrkJ transcription
(data not shown). Therefore, it is possible that transcription of
mrkJ can also occur from a promoter immediately upstream of
it. Deletion of mrkHI, like the single mrkI mutation, resulted in
the decreased ability of K. pneumoniae to produce surface-
associated type 3 fimbriae. Repeated attempts were made to
construct an mrkH deletion mutant but proved unsuccessful.
The precise reason for this is unclear but suggests that such a
mutation may be lethal, even though deletion of mrkHI to-
gether and reintroduction of mrkI alone is not.
Interestingly, we also demonstrated that the MrkI mutant
was nonfimbriate only when cultured under aerobic condi-
tions. When these strains were grown anaerobically on agar
or microaerophilically as deep static broth cultures, the mu-
tants exhibited fimbrial titers equivalent to or higher than
those observed for the parental strains grown aerobically.
The MrkHI mutant, in contrast, was consistently nonfimbriate
under both aerobic and anaerobic conditions. It is possible that
the MrkI mutant is fimbriate when grown anaerobically due to
increased expression of mrkH under these conditions, which
facilitates fimbria production independently of MrkI. Conse-
quently, during anaerobic growth, MrkH and MrkI are likely to
facilitate increased mrkA expression, resulting in a strongly
fimbriate phenotype. However, in the absence of MrkI, the
increased MrkH production, compared to that exhibited by
bacteria grown aerobically, may enable MrkH to interact with
an orphan activator to facilitate mrkA transcription. Also, it is
possible that K. pneumoniae, in response to different environ-
mental conditions, produces different regulators that can in-
teract with MrkH to modulate surface expression of type 3
fimbriae. Currently, we are investigating the interaction be-
tween MrkH and MrkI in the presence and absence of c-di-
GMP. However, MrkH may act as a protein with a more
general function of sensing the intracellular concentrations
of c-di-GMP. This is supported by the observation that un-
like many organisms in which c-di-GMP serves a regulatory
role, all sequenced K. pneumoniae genomes (K. pneumoniae
MGH78578, K. pneumoniae NTUH-K2044, and K. pneu-
moniae 342) possess only one PilZ domain-containing protein
(MrkH) that is not predicted to act as a cellulose synthase. This
is not unique within the Enterobacteriaceae, as it appears that
many members of this family contain only the PilZ domain
carrying protein YcgR. However, the previously described N-
terminal YcgR domain of these proteins exhibits no related-
ness to that of MrkH. Currently, no other c-di-GMP binding
proteins have been characterized in K. pneumoniae, so it is
possible that c-di-GMP sensing is a major function of MrkH.
The ability of MrkH and MrkI to facilitate transcription of
mrkA in an E. coli transformant that possesses no mrk genes of
its own and the DNA binding domain present in MrkI led us to
speculate that MrkI binds the promoter region of mrkA. How-
ever, we were not able to demonstrate binding in vitro using gel
mobility shift assays (data not shown). The fact that MrkA can
be made in the absence of MrkI under specific growth condi-
tions makes it less surprising that it was not possible to dem-
onstrate this interaction. However, it is possible that MrkI
binds to this region, and we were unable to replicate in vitro the
conditions for binding in vivo. Additional bacterial factors may
be required for this activity. The results shown by E. coli trans-
formants are consistent with the observation that MrkH and
MrkI affect transcription, as detected by qRT-PCR in K. pneu-
moniae. The presence of MrkH alone facilitates detectable
levels of mrkA transcription in both E. coli and K. pneumoniae
transformants, but this is significantly lower than that observed
when both MrkH and MrkI are present. The MrkI mutant of
K. pneumoniae is able to produce MrkH, but this mutant is
phenotypically nonfimbriate and exhibits no mrkA transcrip-
tion. The level of mrkA transcription in E. coli transformants
possessing only MrkH could simply be due to the relatively
high concentrations of MrkH produced by the cloned gene.
Since analysis of the K. pneumoniae genome indicates that the
only PilZ-possessing protein in these bacteria that is not in-
volved in cellulose metabolism is MrkH, it is possible that this
protein acts as a c-di-GMP signaling adaptor for many systems
that are regulated by the intracellular concentrations of the
molecule. Consequently, its effect on gene transcription may
depend on the intracellular concentrations of MrkH. In the
absence of MrkI, high levels of MrkH may more weakly inter-
act, directly or indirectly, with additional regulators. Further
studies will be required to investigate this hypothesis.
It is becoming increasingly clear that the genetic regulation
of fimbrial genes encoding appendages assembled by the chap-
erone-usher pathway is subject to a complex regulatory circuit
involving different families of DNA binding proteins (1, 18,
21). The type 3 fimbrial system, a fimbrial type commonly
observed to be produced by enterobacteria associated with
nosocomially acquired infections, is also likely to be regulated
VOL. 193, 2011MrkHI AND TYPE 3 FIMBRIAL EXPRESSION3459
by multiple gene products. The identification of these regula- Download full-text
tory factors will facilitate an understanding of type 3 fimbria
production and its role in host cell interaction. Due to the
multitiered regulatory networks that govern other fimbrial sys-
tems, it is possible that MrkH and MrkI regulate type 3 fim-
brial expression by acting upstream of a primary regulator.
However, the location of mrkHIJ immediately adjacent to
mrkABCDF may indicate an evolutionary selection for these
two gene clusters. Currently, efforts are under way to further
investigate the role of both MrkH and MrkI as fimbrial
regulators and also to determine whether MrkI regulates
expression of nonfimbrial genes.
We thank Carrie Harwood (University of Washington, Seattle, WA)
for kindly providing the WspR protein to synthesize labeled c-di-GMP.
This work was supported by a grant from the NIH to S.C. (R01
AI050011). J.G.J. was supported by the NIH Mechanisms of Parasit-
ism Training Grant (T32 AI007511).
1. Blomfield, I. C. 2001. The regulation of pap and type 1 fimbriation in
Escherichia coli. Adv. Microb. Physiol. 45:1–49.
2. Boddicker, J. D., R. A. Anderson, J. Jagnow, and S. Clegg. 2006. Signature-
tagged mutagenesis of Klebsiella pneumoniae to identify genes that influ-
ence biofilm formation on extracellular matrix material. Infect. Immun.
3. de Lorenzo, V., and K. N. Timmis. 1994. Analysis and construction of stable
phenotypes in gram-negative bacteria with Tn5- and Tn10-derived minitrans-
posons. Methods Enzymol. 235:386–405.
4. Dickson, C. 2008. Protein techniques: immunoprecipitation, in vitro kinase
assays, and Western blotting. Methods Mol. Biol. 461:735–744.
5. Frank, D. N., S. S. Wilson, A. L. St. Amand, and N. R. Pace. 2009. Culture-
independent microbiological analysis of foley urinary catheter biofilms. PLoS
6. Gerlach, G. F., B. L. Allen, and S. Clegg. 1988. Molecular characterization of
the type 3 (MR/K) fimbriae of Klebsiella pneumoniae. J. Bacteriol. 170:
7. Gerlach, G. F., B. L. Allen, and S. Clegg. 1989. Type 3 fimbriae among
enterobacteria and the ability of spermidine to inhibit MR/K hemagglutina-
tion. Infect. Immun. 57:219–224.
8. Gerlach, G. F., S. Clegg, and B. L. Allen. 1989. Identification and charac-
terization of the genes encoding the type 3 and type 1 fimbrial adhesins of
Klebsiella pneumoniae. J. Bacteriol. 171:1262–1270.
9. Hickman, J. W., and C. S. Harwood. 2008. Identification of FleQ from
Pseudomonas aeruginosa as a c-di-GMP-responsive transcription factor.
Mol. Microbiol. 69:376–389.
10. Hickman, J. W., D. F. Tifrea, and C. S. Harwood. 2005. A chemosensory
system that regulates biofilm formation through modulation of cyclic
diguanylate levels. Proc. Natl. Acad. Sci. U. S. A. 102:14422–14427.
11. Hornick, D. B., B. L. Allen, M. A. Horn, and S. Clegg. 1992. Adherence to
respiratory epithelia by recombinant Escherichia coli expressing Klebsiella
pneumoniae type 3 fimbrial gene products. Infect. Immun. 60:1577–1588.
12. Hornick, D. B., J. Thommandru, W. Smits, and S. Clegg. 1995. Adherence
properties of an mrkD-negative mutant of Klebsiella pneumoniae. Infect.
13. Jagnow, J., and S. Clegg. 2003. Klebsiella pneumoniae MrkD-mediated
biofilm formation on extracellular matrix- and collagen-coated surfaces. Mi-
14. Johnson, J. G., and S. Clegg. 2010. Role of MrkJ, a phosphodiesterase, in
type 3 fimbrial expression and biofilm formation in Klebsiella pneumoniae.
J. Bacteriol. 192:3944–3950.
15. Jones, R. N. 2010. Microbial etiologies of hospital-acquired bacterial pneu-
monia and ventilator-associated bacterial pneumonia. Clin. Infect. Dis.
16. Langstraat, J., M. Bohse, and S. Clegg. 2001. Type 3 fimbrial shaft (MrkA)
of Klebsiella pneumoniae, but not the fimbrial adhesin (MrkD), facilitates
biofilm formation. Infect. Immun. 69:5805–5812.
17. Link, A. J., D. Phillips, and G. M. Church. 1997. Methods for generating
precise deletions and insertions in the genome of wild-type Escherichia coli:
application to open reading frame characterization. J. Bacteriol. 179:6228–
18. Mikkelsen, H., G. Ball, C. Giraud, and A. Filloux. 2009. Expression of
Pseudomonas aeruginosa CupD fimbrial genes is antagonistically controlled
by RcsB and the EAL-containing PvrR response regulators. PLoS One
19. Miller, V. L., and J. J. Mekalanos. 1988. A novel suicide vector and its use
in construction of insertion mutations: osmoregulation of outer membrane
proteins and virulence determinants in Vibrio cholerae requires toxR. J.
20. Moravek, M., et al. 2004. Colony immunoblot assay for the detection of
hemolysin BL enterotoxin producing Bacillus cereus. FEMS Microbiol. Lett.
21. Nicastro, G. G., A. L. Boechat, C. M. Abe, G. H. Kaihami, and R. L. Baldini.
2009. Pseudomonas aeruginosa PA14 cupD transcription is activated by the
RcsB response regulator, but repressed by its putative cognate sensor RcsC.
FEMS Microbiol. Lett. 301:115–123.
22. Nordmann, P., G. Cuzon, and T. Naas. 2009. The real threat of Klebsiella
pneumoniae carbapenemase-producing bacteria. Lancet Infect. Dis. 9:228–
23. Ohana, S., et al. 2005. Spread of a Klebsiella pneumoniae strain producing
a plasmid-mediated ACC-1 AmpC beta-lactamase in a teaching hospital
admitting disabled patients. Antimicrob. Agents Chemother. 49:2095–2097.
24. O’Toole, G. A., and R. Kolter. 1998. Flagellar and twitching motility are
necessary for Pseudomonas aeruginosa biofilm development. Mol. Micro-
25. O’Toole, G. A., et al. 1999. Genetic approaches to study of biofilms. Methods
26. Philippe, N., J. P. Alcaraz, E. Coursange, J. Geiselmann, and D. Schneider.
2004. Improvement of pCVD442, a suicide plasmid for gene allele exchange
in bacteria. Plasmid 51:246–255.
27. Pratt, J. T., R. Tamayo, A. D. Tischler, and A. Camilli. 2007. PilZ domain
proteins bind cyclic diguanylate and regulate diverse processes in Vibrio
cholerae. J. Biol. Chem. 282:12860–12870.
28. Ronald, A. 2002. The etiology of urinary tract infection: traditional and
emerging pathogens. Am. J. Med. 113(Suppl. 1A):14S–19S.
29. Rosen, D. A., et al. 2008. Utilization of an intracellular bacterial community
pathway in Klebsiella pneumoniae urinary tract infection and the effects of
FimK on type 1 pilus expression. Infect. Immun. 76:3337–3345.
30. Schurtz, T. A., D. B. Hornick, T. K. Korhonen, and S. Clegg. 1994. The type
3 fimbrial adhesin gene (mrkD) of Klebsiella species is not conserved among
all fimbriate strains. Infect. Immun. 62:4186–4191.
3460JOHNSON ET AL.J. BACTERIOL.