Type 1 fimbriae have been shown to be specifically required for Escherichia coli colonisation and pathogenesis of the urinary tract. These structural organelles mediate specific adhesion to α-D-mannosides by virtue of the FimH adhesin. FimH is a two-domain protein in which the N-terminal domain contains the receptor-binding site and the C-terminal domain is required for organelle integration. To date, FimH has only been isolated as a complex with the system-specific chaperone FimC. Here we report that a functional form of the FimH receptor-binding domain can be readily isolated and characterised by replacing the C-terminal domain with a histidine tag.
Type 1 fimbriae mediate specific adhesion to a spectrum of α-D-mannosides found on mammalian tissue surfaces. Thereby these organelles confer attachment and colonisation of Escherichia coli strains to various host surfaces including the urinary tract [1,2]. A type 1 fimbria is a 7 nm wide by about 1 μm long heteropolymer structure. Approximately 1000 copies of the major subunit, FimA, are polymerised into a right-handed helical structure which additionally contains a few percent of the minor components FimF, FimG and FimH [3,4]. It has been shown that the receptor-recognition element of type 1 fimbriae is the FimH protein . The FimH adhesin is located at the tip of the organelle as an integral part of a short fibrillum . Several lines of evidence suggest that the FimF and FimG proteins act as adapters for integration of the adhesin into the fimbrial organelle [3,6]. Furthermore, FimH has also been reported to be interspersed along the fimbrial shaft [5,7].
The FimH protein is produced as a precursor of 300 amino acids which is processed into a mature form of 279 amino acids . Two approaches have been used to investigate structure–function relationships of FimH. Linker insertion mutagenesis of the fimH gene suggested that amino acids located in the first half of FimH were involved in receptor recognition . This result was corroborated by fusion studies in which selected parts of FimH were fused to either FocH or MalE [9,10]. Recently, the 3D structure of a FimH–FimC adhesin–chaperone complex was elucidated . According to this, the FimH protein is folded into two domains, an N-terminal adhesive domain (residues 1–156) linked by a short tetrapeptide loop to a C-terminal organelle integration domain (residues 160–279). The only soluble form of FimH that has been reported on is the FimH–FimC complex. This prompted us to engineer a construct consisting of the FimH binding domain fused to a His tag. Here we show that this construct is adhesive, soluble and easy to purify.
Materials and methods
Strains, plasmids and growth conditions
The E. coli strain HB101 (F′lacI Tn5) was used in this study . Plasmid pPKL4 contains the complete type 1 fimbrial gene cluster in pBR322 . Plasmid pQE-30 is an over-expression vector (Qiagen). Plasmid pPKL241 contains the first half of the fimH gene (encoding the signal peptide and the first 156 amino acids of the mature protein) fused to a six histidine tag and ligated into the EcoRI–BamHI sites of plasmid pQE-30 (see Section 3). Cells were grown in Luria–Bertani (LB) medium supplemented with the appropriate antibiotics .
Isolation of plasmid DNA was carried out using the QIAprep Spin Miniprep Kit (Qiagen). Polymerase chain reaction (PCR) products were purified using the GFX PCR DNA and Gel Band Purification Kit (Amersham). Restriction endonucleases were used according to the manufacturer's specifications (New England Biolabs or Pharmacia). PCR for plasmid construction was performed on a Perkin–Elmer GeneAmp machine. The amplification was done for 20 cycles as follows: primary denaturation at 94°C, 15 s; primer annealing at 50°C, 30 s; and primer extension at 72°C, 30 s. Fifty picomoles of each primer and 2 U of polymerase (Expand, Boehringer-Mannheim) were used in each reaction. The sequences of the primers used are shown in Fig. 1. All constructs were sequenced to ensure PCR fidelity by the dideoxy chain termination technique  using a Sequenase version 2.0 kit from USB.
Purification of the FimH156–His protein
HB101F′ cells transformed with the plasmid coding for fimH156–His (pPKL241) were grown to an OD600 of 0.6 and FimH156–His production was induced with 1 mM IPTG for 1 h. Periplasmic protein fractions were prepared as follows: cells were harvested by centrifugation and resuspended in a buffer containing 30 mM Tris–HCl, 20% sucrose, pH 8.0. EDTA was added drop-wise to a final concentration of 1 mM and the solution incubated on ice with gentle agitation for 10 min. The cells were harvested again by centrifugation and resuspended in the same volume of ice-cold 5 mM MgSO4. Following a 10-min incubation on ice the cells were centrifuged again and the supernatant (cold osmotic shock fluid) was dialysed extensively against a buffer containing 50 mM Na-phosphate pH 7.8, 300 mM NaCl. This resulted in essentially a preparation of periplasmic proteins in phosphate-buffered solution. The FimH1–156 protein was purified from this periplasmic fraction by Ni-NTA affinity chromatography according to standard procedures (Qiagen).
Samples were subjected to SDS–PAGE (15%) gel electrophoresis and transferred to PVDF microporous membrane filters using a semi-dry blotting apparatus. Receptor blots of FimH156–His to α-D-mannosylated BSA (Sigma) were carried out essentially as previously described . In short, filter blots were first incubated with α-D-mannosylated BSA (0.5 mg l−1) followed by incubation with rabbit anti-BSA serum and finally with peroxidase-conjugated anti-rabbit serum.
Results and discussion
Construction of the fimH1–156 over-expression plasmid
The region encoding the N-terminal half of the FimH adhesin (i.e. the signal peptide and residues 1–156 of the mature protein) was amplified by PCR using two specific primers (Fig. 1). The first primer contained an EcoRI cloning site and an optimised ribosome binding site. In addition, we introduced an A-T alteration in the third codon of FimH that resulted in this being changed from a rare arginine codon (CGA) to one optimally used by E. coli (CGT). Preliminary studies from this laboratory indicate that this rare arginine codon may affect translation of the fimH gene (H. Hasman, M.A. Schembri, and P. Klemm, unpublished observations). tRNAs that recognise rare arginine codons have been shown to affect gene expression when positioned among the first 25 codons of a given reading frame . In addition, a similar phenomenon has also been reported for the rare leucine codon TTG in the type 1 fimbrial FimB recombinase  and for the normal expression of type 1 fimbriae in Salmonella typhimurium. The second primer contained a BamHI cloning site and a sequence encoding a six-histidine tag fused to the end of the N-terminal linker domain. This resulted in a construct consisting of the first half of FimH (i.e. the receptor-recognition domain) fused to a histidine tag. We refer to the protein encoded by this truncated fimH gene as FimH156–His. The amplified PCR product was digested with EcoRI–BamHI and ligated into the similarly cut over-expression plasmid pQE-30 resulting in the plasmid pPKL241.
Overexpression and purification of FimH1–156
The FimH156–His protein was over-expressed in E. coli HB101F′ (pPKL241) by induction with IPTG. Periplasmic fractions were prepared from a 2-l culture and the FimH156–His protein was purified using a Ni-NTA affinity matrix (Qiagen). The results of the protein purification procedure are shown in Fig. 2. A single protein with an expected size of approximately 15 kDa was obtained after purification on the Ni-NTA affinity matrix.
We tested the functional activity of the FimH156–His protein by examining its ability to bind to α-D-mannosylated BSA (Fig. 2B). The purified protein reacted strongly in this assay in a mannose-inhibitable manner. Previous attempts by us and others to purify segments of the FimH adhesin have proved unsuccessful (; M.A. Schembri, D.L. Hasty and P. Klemm, unpublished observations), most likely as a result of the extreme instability of such proteins in the bacterial periplasm. Interestingly, we obtained higher yields of FimH156–His in the E. coli strain HB101 than the well-documented strain BL21. Previous reports have demonstrated that overexpression of the fimH gene leads to the formation of naturally occurring FimH truncates with inherent α-D-mannose recognition [6,9]. We did not observe any degradation of the FimH156–His protein (Fig. 2B, lane 2), suggesting that processing sites are not contained within the FimH156–His protein.
Fimbrial adhesins are generally thought to possess a similar two-domain protein structure to that of the FimH adhesin. Indeed, this has been inferred from the primary structure of other adhesins including PapG , FocH , SfaS  and GafD . Despite our knowledge of fimbriae accumulated over the last few decades, the highly stable nature of these structures has made the purification of the minor adhesin components extremely difficult. Furthermore, the adhesins are highly sensitive to periplasmic proteases when not complexed with their cognate chaperone proteins. The approach used here to purify the receptor-binding domain of FimH avoids this problem and could be applied to purify relevant sectors of other fimbrial adhesins for structure–function studies and facilitates the analysis of their cognate receptor targets.
Naturally occurring variants of the FimH adhesin have been identified that confer upon E. coli a higher tropism for extraintestinal sites. In particular, variants that bind strongly to terminally exposed monomannose residues have been associated with a pathogenicity-adaptive phenotype that enhances colonisation of the urinary bladder , while collagen binding is associated with meningitis-causing strains . We have also recently employed a random mutagenesis approach to identify additional FimH phenotypes with differences in binding affinity for oligosaccharide moieties rich in either terminal monomannose, oligomannose or non-mannose residues . Importantly, these changes reside within the receptor-binding domain of the protein. The work described here may be applied to the over-expression and purification of N-terminal FimH variants in order to assess their binding affinity in a tightly controlled single protein binding system. Furthermore, these proteins may help to determine the role of the fimbrial shaft in receptor recognition by FimH.
The observation that almost all uropathogenic strains of E. coli produce type 1 fimbriae suggests that FimH may be a potential vaccine candidate for the prevention of infection in the urinary tract. In mouse models, immunisation with FimH truncates and synthetic FimH peptides have been shown to prevent urogenital mucosal infection by E. coli[10,25]. In addition, passive systemic administration of immune sera directed against FimH also resulted in reduced colonisation of the urinary bladder. Our FimH1–156 histidine-tagged protein could potentially be exploited in the development of a vaccine to prevent recurrent and acute infections of the urogenital mucosa.
We thank Birthe Jul Jørgensen for expert technical assistance. This study was supported by The Danish Medical Research Council (Grant 9802358).