点击显示 收起
【摘要】 The adhesion of uropathogenic Escherichia coli to the urothelial surface of the bladder is a prerequisite for the establishment of bladder infections. This adhesion process relies on E. coli adhesins and their cognate urothelial receptors, and it also is influenced by an intricate array of defense mechanisms of the urinary system. In this study, we examined the in vivo role of Tamm-Horsfall protein (THP), the most abundant urinary protein, in innate urinary defense. We genetically ablated the mouse THP gene and found that THP deficiency predisposes mice to bladder infections by type 1-fimbriated E. coli. Inoculation of too few type 1-fimbriated E. coli to colonize wild-type mice caused significant bladder colonization in THP-knockout mice. In contrast, THP deficiency did not enhance the ability of P-fimbriated E. coli to colonize the bladder. Our results provide the first in vivo evidence indicating that under physiological conditions, the mannosylated THP can serve as an effective soluble "receptor," binding to the type 1-fimbriated E. coli and competitively inhibiting them from adhering to the uroplakin Ia receptors present on the urothelial surface. These results suggest that potential THP defects, either quantitative or qualitative, could predispose the urinary bladder to bacterial infections. The generation of THP-deficient mice established the role of THP as a first line of urinary defense and should help elucidate other potential functions of this major protein in urinary tract physiology and diseases.
【关键词】 innate host defense gene knockout urinary bladder infection FimH adhesin
URINARY TRACT INFECTIONS (UTIs) are the second most common infectious disease, occurring about 150 million times annually on a global basis and costing over $6 billion in health care expense ( 48 ). One in five UTI episodes may be followed by a recurrent UTI episode, causing considerable morbidity and time lost from work. Recurring UTIs increase the risk for pyelonephritis and long-term renal damage. The emergence of drug-resistant uropathogens presents a further clinical challenge ( 4 ).
Great strides have recently been made in understanding UTI pathogenesis, the most significant of which is the revelation of the molecular players at the pathogen-host interface ( 33, 36, 58, 70 ). To cause infection, uropathogenic Escherichia coli, by far the most common cause of UTI, must first adhere to the bladder epithelial surface ( 15, 49, 65 ). The adhesion allows the E. coli to gain a foothold in the bladder and thereby prevents complete expulsion of the bacteria during micturition. Of the many adhesins produced by E. coli, the FimH lectin of type 1 fimbriae appears to be the one most critical for establishment of cystitis ( 6, 10, 27 ). In vitro and in vivo animal studies demonstrated that FimH can bind specifically to the mannosylated uroplakin Ia (UPIa), a major protein component of the urothelial plaques that cover more than 90% of the urothelial surface ( 32, 34, 68, 70, 71 ). The adhesin/receptor binding is essential to triggering the internalization of the E. coli into the superficial (umbrella) cells of the urothelium ( 29, 34 ). Once inside these cells, the E. coli propagates rapidly, culminating in the destruction of the host epithelial cells and the activation of a cascade of host inflammatory responses. Therefore, mounting experimental evidence indicates that the interaction between the E. coli adhesin and its urothelial receptors plays a critical role in the initiation of UTI.
Despite this progress in our understanding of adhesin-receptor interactions in the urinary tract, little is known about the innate host defenses that naturally counter bacterial adhesion. In the respiratory and intestinal tracts, much of the anti-adherence activity is tied to the abundant mucus secretions that effectively shield the epithelial receptors from the invading pathogens ( 26, 31, 55, 67 ). However, several recent high-resolution electron microscopic studies suggest that such a conspicuous mucus protection is absent from most, if not all, of the urothelial surface ( 22, 34, 68 ). Indeed, the uroplakin receptors for the FimH adhesin can be seen directly exposed to the urinary space. To prevent uropathogens from binding to the urothelial surface, the urinary system must therefore rely on the physical cleansing effects imparted by micturition and, perhaps as importantly, the soluble anti-adhesion factors present in the urine. A prominent candidate for an anti-adhesive urinary factor is the Tamm-Horsfall protein (THP), or uromodulin. THP is the most abundant protein in mammalian urine, with an excretion rate of up to 100 mg daily ( 24, 25 ). Made by the tubular cells of the ascending limb of Henle's loop of the kidney, THP is anchored onto the apical membrane via glycophosphatidyl inositol linkage and is released into the urine by the activity of either proteases or phospholipases ( 5, 11, 12, 17, 40, 45 ). THP has a tendency to self-aggregate via its juxtamembrane-positioned zona pellucida domain, forming filament-like polymers ( 20 ), a characteristic potentially important for trapping of E. coli. Although the physiological function of THP is unknown, several lines of evidence suggest that THP may play a role in host defense. First, THP contains high mannose residues that are capable of interacting with the FimH adhesins ( 7, 38, 39, 42, 51 ). Both THP and its high mannose moieties are well conserved across species, suggesting an indispensable role during evolution ( 38, 53, 56 ). Second, type 1-fimbriated E. coli were found to be trapped within the renal cast primarily consisting of THP ( 37 ). Third, type 1-fimbriated E. coli can bind to immobilized THP in vitro ( 16 ), and this binding blocked the binding of E. coli to kidney epithelial cells ( 28 ). Finally, our recent in vitro adherence studies showed that soluble THP binds to type 1-fimbriated E. coli and competitively inhibits the E. coli from binding to the immobilized UPIa receptors ( 38 ). These studies, therefore, suggest a putative role for THP as a competitive inhibitor of the FimH adhesin of the type 1-fimbriated E. coli. Whether THP is important in innate urinary defense in vivo, however, is still unclear and is the subject of the present study.
MATERIALS AND METHODS
THP gene targeting. A 60-kb DNA fragment was isolated from a 129/SVJ mouse genomic (BAC) library by PCR-based mass screening using THP-specific oligonucleotide primers ( 72 ). Southern blotting with probes located at the 5'-end, middle region and 3'-end of the mouse THP cDNA confirmed that the genomic fragment contained all coding regions plus 5'- and 3'-noncoding regions. A 3.1-kb Xba I/ Spe I restriction fragment in the 5'-upstream region of the THP gene was subcloned into pBluescript, retrieved by Xba I digestion, and cloned into the Xba I site that was 5' to the Neo gene in the pLOXPNT targeting vector ( 21 ). A 2.5-kb Eco RI/ Sac I fragment (containing partial intron 4, exon 5, intron 5, exon 6, and partial intron 6) of the THP gene was subcloned into pBluescript, retrieved by Eco RI digestion, and cloned into the Eco RI site between the Neo and the TK genes in the targeting vector. The vector was linearized by Xho I and was introduced by electroporation into 129/SvEv (W4) embryonic cell line. After transfected cells were double-selected in 200 µg/ml G418 and 2 µM ganciclovir, Southern blotting was used to identify ES cell clones with the correct homologous recombination event. Two ES cell lines (designated as A2 and C12) that were heterozygous for the targeted deletion were microinjected into blastocyst-stage embryos derived from C57/BL6J mice. The injected blastocysts were transferred into the uteri of pseudopregnant C57/BL6J mice. Of the 32 chimeric mice obtained (14 from A2; 18 from C12), 22 male chimeras were bred with Black Swiss mice and all of them transmitted the mutant allele to their offspring. Male chimeras were also bred with 129/SvEv inbred mice to generate heterozygous mice, the latter of which were used to generate homozygous mice. The genotype of the offspring was determined by Southern blotting of Eco RI-digested tail DNA ( Fig. 1 ). All animal studies were conducted in conformity with American Physiological Society's Guiding Principles for the Care and Use of Animals and based on a protocol approved by investigators' institutions' Institutional Animal Care and Use Committees.
Fig. 1. Generation of Tamm-Horsfall protein (THP)-deficient mice by targeted deletion of the THP gene. A : targeting strategy. Exons of the mouse THP gene are indicated with filled boxes ( 1 - 11 ), and 5'-upstream region and introns are indicated with thick lines ( top ). The targeting vector ( middle ) contained a 3.1-kb 5'-upstream region of the THP gene, a Neo-resistance (Neo) gene, a 2.5-kb THP fragment (spanning partial intron 4, exon 5, intron 5, exon 6, and partial intron 6), and a thymidine kinase (TK) gene. Homologous recombination would introduce a Neo gene in the targeted THP locus and delete a 650-bp proximal promoter region, exons 1-4, introns 1-3, and partial intron 4 ( bottom ). A 5'-probe (probe) was designed to detect a 7.0-kb wild-type (WT) allele and an 8.5-kb mutant [knockout (KO)] allele. Restrictions sites are E, Eco RI; X, Xba I; S, Spe I; Sa, Sac I; and N, Nco I. B : genotyping of THP-mutant mice by Southern blot analysis. +/+, WT; +/-, heterozygotes; -/-, homozygotes (KO). C : RT-PCR analysis of THP expression in mouse kidneys. Top : note the absence of THP in homozygous mice. Bottom : GAPDH was used to illustrate the integrity of mRNA. D : Northern blot analysis. Total RNAs (15 µg) resolved by agarose gel ( middle ) were hybridized with a full-length mouse THP cDNA probe ( top ) followed by a -actin probe ( bottom ). Note the reduced expression of THP in +/- mice and absence of THP in -/- mice.
RNA and protein analyses. Total RNA was extracted from mouse kidneys using an RNA extraction kit (Promega). For RT-PCR analysis, 2 µg of the total RNA were reverse-transcribed and one-tenth of the transcribed product was PCR-amplified with a pair of primers located in exon 3 of the mouse THP gene (forward: 5'-AGGGCTTTACAGGGGATGGTTG-3'; reverse: 5'-GATTGCACTCAGGGGGCTCTGT-3'). PCR conditions were as follows: 95°C for 5'; 95°C for 30'', 58°C for 30'', 72°C for 1' for 40 cycles; and 95°C for 30'', 58°C for 30'', 72°C for 8' for the last cycle. Glyceraldehyde-6-phosphate dehydrogenase was used as an internal control. For Northern blot analysis, 15 µg of the total RNA were resolved by formaldehyde-agarose gel, transferred onto a nylon membrane, and hybridized with a mouse THP cDNA probe. After autoradiography, the probe was stripped by high-stringency washing and the nylon membrane was rehybridized with a -actin probe.
For the assessment of urinary THP status, mouse urine was collected following suprapubic massage and immediately denatured in SDS-PAGE-loading buffer containing 2% SDS, 5% mercaptoethanol, and 50 mM Tris·HCl (pH 6.8). Total urinary proteins were resolved on an 8% SDS-PAGE and were either stained by Coomassie Blue to reveal the total proteins or transferred onto an Immobilon-PVDF membrane for Western blotting using a polyclonal antibody raised against human THP (Biodesign International). The reaction was visualized by a secondary antibody conjugated with horseradish peroxidase followed by development using an enhanced chemiluminescence detection system (Amersham).
Histology and immunohistochemistry. Freshly dissected kidney tissues were fixed in 10% buffered formalin and processed routinely for paraffin embedding. Three-micron tissue sections were stained with hemotoxylin/eosin and examined by light microscopy. Alternatively, the sections were microwaved in a citrate buffer (pH 6.0) for 15 min for antigen retrieval and stained with the anti-THP antibody.
Experimental bladder infection. Two groups of E. coli strains were used; the first was designed to test the bladder colonization by type 1-fimbriated E. coli. KB96 was obtained by transfecting a fimH -deficient E. coli strain AAEC191A, a K-12 derivative from MG1655 strain ( 3 ), with one plasmid containing the entire fim gene cluster, but having a stop-linker in the 5'-end of the fimH gene and another plasmid containing a fimH gene isolated from a clinical UTI strain ( 38, 57 - 59 ). This recombinant strain expressed a high monomannose-binding FimH variant. KB18 is an isogenic, negative control strain harboring a mutated, nonfunctional fimH gene and does not bind to mannosylated glycoproteins or tissue culture cells. Both strains contained ampicillin-resistant plasmids, suitable for specific bacterial selection. MJ35 is a clinical UTI isolate that expresses the type 1 fimbriae with high monomannose-binding specificity. MJ35-30, an isogenic strain of MJ35, harbors a nonfunctional fimH gene and does not bind monomannose. The E. coli strains were serially cultured three times in Luria-Bertani medium statically to facilitate fimbriation, which was assessed by aggregation assay with baker's yeast ( 70 ). The number of E. coli in the inocula was determined by calculating the colony-forming units (CFU) after plating of diluted aliquots on LB plates. The second group, designed to test bladder colonization by P-fimbriated E. coli, included a nonfimbriated E. coli (P678-54), an isogenic derivative expressing a PapG1 adhesin (HU849), and another strain expressing a PapG2 adhesin (IA2). The presence of P fimbriae and the absence of type 1 fimbriae in these strains were verified by hemagglutination and yeast aggregation, respectively ( 70 ).
THP-knockout and wild-type female mice at 2-3 mo of age (weight 19-26 g) were used for ascending infection experiments. E. coli strains were resuspended in phosphate-buffered saline to various concentrations, and 50 µl were inoculated transurethrally into the bladder while mice were under anesthesia with xylazine and ketamine ( 58 ). The inoculation conditions used did not induce immediate vesicoureteral reflux, as evidenced by the lack of E. coli in the kidney. Twenty-four hours later, the mice were killed, and their bladders were aseptically removed, weighed, and homogenized. Tissue homogenates were cultured on ampicillin-LB plates and numbers of bacteria were determined by calculating the number of CFU per gram of bladder tissue. The x 2 test and Student's t -test were used for calculating the P values.
The binding of type 1-fimbriated E. coli to the urothelial surface was also visualized by reacting paraffin sections of mouse bladders with a polyclonal anti- E. coli antibody conjugated with biotin (Maine Biotechnology Services, Portland, ME) followed by fluorescein-conjugated avidin. Nuclei were counterstained with propidium iodide.
RESULTS
Targeted ablation of the mouse THP gene. Germline inactivation of the mouse THP gene was carried out using a replacement targeting vector whose homologous recombination with the endogenous THP gene would delete not only the first four exons, but also a 650-bp proximal promoter region of the THP gene, thus rendering the THP gene nonfunctional ( Fig. 1 A ). Of the 132 ES cell clones that were resistant to both G418 and ganciclovir, two clones (designated A2 and C12) had the correct homologous recombination event in one allele of the THP gene. Microinjection of the two clones into the blastocysts derived from C57/BL6J mice and implantation of these blastocysts into the uteri of C57/BL6J mice resulted in the birth of 32 chimeras. All 22 male chimeras showed germline transmission of the mutated THP allele, and they were bred with 129/SvEv females to generate heterozygotes ( Fig. 1 B ). Further breeding of heterozygotes yielded offspring with anticipated litter size and with different genotypes (THP+/+, THP+/-, and THP-/-; Fig. 1 B ) distributed in a Mendelian fashion. This indicated that THP gene deletion did not adversely affect embryonic development.
To determine whether the THP gene was indeed inactivated in the knockout mice, we assessed THP expression on both mRNA and protein levels. RT-PCR using primers located within exon 3 of the mouse THP gene detected a specific product in wild-type and heterozygous mice, but not in homozygous mice ( Fig. 1 C ). Similarly, a full-length mouse THP cDNA probe failed to detect any THP mRNA on a Northern blot using RNA from homozygous mice ( Fig. 1 D ). Immunohistochemical results using a polyclonal anti-THP antibody demonstrated strong labeling in wild-type animals in the thick ascending limb of the Henle's loop and early distal tubules ( Fig. 2 A ), renal tubules known to synthesize THP ( 9, 54 ). The labeling was confined to the luminal surface of the tubules, consistent with the fact that THP is normally anchored onto the apical leaflet of the plasma membrane via GPI linkage ( Fig. 2 A; Ref. 45 ). Heterozygous mice stained less strongly for the anti-THP antibody ( Fig. 2 B ), whereas homozygous mice were completely negative ( Fig. 2 C ). Because THP is known to be present in normal urine in large quantities ( 24, 25 ), we performed Western blot analysis using the polyclonal anti-THP antibody. The 100-kDa THP protein species was present in wild-type and heterozygous mice but was absent in homozygous mice ( Fig. 2, D and E ). Taken together, these results firmly established the complete inactivation of THP gene expression in the knockout mice.
Fig. 2. Analysis of the THP product in the kidneys and urine of control and mutant mice. A polyclonal antibody against THP was used to immunohistochemically stain kidneys from WT ( A ), heterozygous ( B ), and homozygous ( C ) mice. The strongest labeling of THP was in the thick ascending limb of Henle's loop and early distal tubules in the +/+ mouse ( A ); the labeling in the +/- mouse was much less ( B ); and there was no labeling in the -/- mouse ( C ). Magnifications are all x 200. SDS-PAGE and Coomassie blue staining of the total urinary proteins (30 µl urine/lane; D ) and Western blotting using a polyclonal antibody against human THP ( E ). The 100-kDa THP (arrow) was absent from urine samples of -/- mice ( lanes 7 - 9 ). Molecular mass standards are, from top to bottom (short bars on the left ), 215, 132, 95, and 42.5 kDa.
THP deficiency predisposes mice to bladder infection. To examine the potential role of THP in protecting against urinary bladder infections, we inoculated the bladders of THP-knockout and wild-type mice with isogenic E. coli strains that varied in their expression of two adhesins that have been shown to be important in establishing UTIs, type 1 and P fimbriae. Type 1 fimbriae are hairlike structures composed primarily of the structural subunit FimA and exhibiting a mannose-sensitive lectin subunit FimH at the fimbrial tips. The two recombinant type 1-fimbriated strains expressed fimbriae bearing either a uropathogenic fimH allele or a mutated, nonfunctional fimH ( 38, 58 ). Twenty-four hours after transurethral inoculation, the FimH-negative mutant (KB18) was not present in the bladders of either mouse group, even at an inoculum of 10 8 E. coli ( Fig. 3 A ), whereas the strain expressing a functional FimH adhesin (KB96) effectively colonized the bladders of both wild-type and knockout mice. This result indicated that E. coli colonization was strictly FimH dependent and that high numbers of bacteria could effectively overcome host defenses ( 38, 58 ). However, significant differences were observed when lower numbers of E. coli were inoculated. For example, at an inoculum of 10 5 E. coli, 65% (11/17) of the bladders of the knockout mice were colonized, compared with 24% of the wild-type mice ( P < 0.05; Fig. 3 B ), and even with an inoculum of 10 4 E. coli, 55% of the bladders of knockout mice were colonized, compared with only 5% of the wild-type mice ( P < 0.01). Furthermore, larger numbers of E. coli could usually be recovered from the bladders of knockout mice than wild-type mice ( Fig. 3 C ). Similar results were obtained with a clinical UTI strain (MJ35) that expressed the type 1 fimbriae ( Fig. 3 D ). The greater level of colonization in knockout mice could also be seen when tissue sections were examined for E. coli by immunofluorescence ( Fig. 4, B and C ). Thus, under normal circumstances, in which the urinary tract would most likely be exposed to relatively small numbers of pathogenic E. coli, THP plays a critical role in the first line of defense, presumably by saturating the FimH adhesins, inhibiting their binding to urothelial cells and, thereby, facilitating the pathogen's elimination.
Fig. 3. Colonization of urinary bladders of WT and KO mice by type 1-fimbriated Escherichia coli. A and B : percentage of bladders colonized. C : colony-forming units (CFU) of E. coli recovered from bladders. WT and homozygous KO mice were challenged via transurethral inoculation with various numbers of E. coli. A : with an inoculum of 10 8 bacteria, an E. coli strain defective for the FimH adhesin (KB18, fimH ) exhibited no bladder colonization, whereas an isogenic strain (KB96) expressing a functional FimH adhesin colonized all bladders of both WT and homozygous mice. B : with the lower E. coli inocula (e.g., at 10 4 and 10 5 ), the frequency of colonization of bladders of KO mice was much higher than with WT controls. P values ( x 2 test) indicate the significance of differences in bladder colonization between WT and mutant groups. C : CFU of E. coli in each colonized bladder. The data were taken from the same experiment shown in B and were plotted to indicate the numbers of E. coli cfu in each of the colonized bladders. The number of animals in each group is indicated above the x -axis, and bladders without any colonization were not plotted. Note that at each inoculum density except the very highest, E. coli colonized KO mice in greater numbers than the WT mice. D : similar results were obtained with clinical strains, as inoculation with a FimH-null control strain (MJ35-30, fimH ) failed to colonize either KO ( n = 6) or WT mice ( n = 6), whereas a strain expressing a functional FimH adhesin (MJ35; 10 5 bacteria) caused bladder colonization in all KO mice ( n = 6), but only 1 in 5 in WT mice.
Fig. 4. Morphological identification of E. coli bound to the urothelial surface. Twenty-four hours after mice were transurethrally challenged with 10 4 E. coli, the bladders were excised, fixed, and subjected to immunofluorescent staining with an anti- E. coli antibody (green) followed by propidium iodide counterstaining of the bladder nuclei (red). Fluorescent microscopy showed little to no E. coli adhesion in WT mice ( A ) and a large number of E. coli (green; arrows) adhering to the urothelial surface in homozygous mice ( B ). Both panels are x 400. C : hematoxylin and eosin staining showing internalized E. coli in a superficial umbrella cell (arrows; x 1,000).
Parallel experiments in which the bladders were inoculated with E. coli strains expressing P-fimbriae failed to show any effect of THP absence on E. coli colonization ( Fig. 5 ). This observation is consistent with the fact that P-fimbriae appear to be dispensable for bladder colonization ( 69 ), and it clearly indicates that the protective role afforded by THP against bacterial infection is type 1 fimbria specific. Whether THP deficiency could predispose bladder colonization by other strains of P-fimbriated E. coli remains to be determined.
Fig. 5. THP absence did not increase bladder colonization by P-fimbriated E. coli. WT and mutant mice (7 mice/group) were challenged transurethrally with nonfimbriated negative controls [KB18 ( fimH ) and P678-54], a type 1-fimbriated-positive control (KB96) or P-fimbriated strains (HU849 expressing PapG1 adhesin and IA2 expressing PapG2 adhesin). Although larger numbers of KB96 colonized the bladders of THP-KO mice than the WT mice ( P < 0.01; Student's t -test), neither P-fimbriated E. coli strain showed any enhanced colonization in the mutant mice, suggesting that urinary defense provided by THP is type 1 fimbria specific.
DISCUSSION
Unlike the intestinal tract, which is well populated by E. coli and many other species of bacteria, the urinary tract is normally sterile ( 18, 23 ). This sterility is usually maintained despite the fact that the urinary tract is steadily exposed to numerous intestinal pathogens by virtue of the anatomic proximity of the urinary and intestinal orifices and the fact that the uroplakin receptors are directly exposed to the urinary space and readily accessible to E. coli type 1 fimbriae ( 22, 34 ). The ability of the urinary tract to ward off the invading pathogens has long been attributed to the voiding action of the bladder ( 46 ). In addition, it was recently discovered in an experimental mouse model that the urothelium could defend against the E. coli by shedding off luminal cells that had become apoptotic following E. coli attachment and invasion ( 34 ). By ablating the THP gene, we provide the first in vivo evidence indicating that the urinary tract possesses yet another crucial defense mechanism involving THP. We demonstrated that intravesicular inoculation of type 1-fimbriated E. coli caused significantly greater bladder colonization in mice lacking THP than the normal controls (Figs. 3 and 4 ). This result, combined with our recent observation that purified THP can bind to type 1-fimbriated E. coli and prevent them from binding to the uroplakins ( 38 ), strongly suggests that THP can serve as an innate defense, preventing E. coli binding to the urothelial surface. It should be noted that intravesicular inoculation of large numbers of E. coli (ranging from 10 7 to 10 10 ) could cause bladder colonization/infection even in wild-type mice as shown in our present study and many other previous studies ( 6, 14, 35, 58, 60 ). However, these large inocula may not be physiological because they likely exceed by several orders of magnitude the number of E. coli that would initially invade the urinary tract under most circumstances. Therefore, under physiological conditions where a small number of E. coli invades the urinary tract, the amount of urinary THP would likely be sufficient to overwhelm the invading pathogens.
Based on our recent understanding of UTI pathogenesis, it can be proposed that THP exerts its defensive role by acting as a competitive inhibitor of the interaction of the FimH lectin adhesin with urothelial receptors. Using an in vitro adhesion system, we previously showed that type 1-fimbriated E. coli can bind to highly purified urothelial plaques that cover over 90% of the bladder surface ( 70 ). This binding can be abolished by D -mannose suggesting that it is mediated by the mannosylated glycoproteins of the urothelial plaques. A gel-overlay assay established that E. coli specifically recognized UPIa and UPIb but not their deglycosylated forms lacking the high mannoses, strongly indicating that the urothelial binding by the type 1-fimbriated E. coli was mediated by the FimH adhesin of the E. coli and the mannose receptors of the uroplakins ( 70 ). The fact that the binding can be reproduced with urothelial plaques purified from human, monkey, cattle, and mice suggests that the high mannose modification of the uroplakin I's is highly species conserved ( 70 ). Our in vitro results were recently corroborated structurally by an in vivo infection model in which the tip of the type 1 fimbriae was seen to interact directly with the center of the uroplakin particles where UPIa and UPIb are likely to be situated ( 32, 34 ). Finally, both in vitro and in vivo experiments showed that allelic variants of type 1-fimbriated E. coli prevalent in UTIs, rather than those prevalent in feces, bind to uroplakins and cause more bladder colonization ( 58 ). This suggests that FimH-uroplakin interaction provides a selective advantage for certain intestinal E. coli to survive in the urinary tract. These studies established the role of uroplakins as the "immobile" urothelial receptors for the type 1-fimbriated E. coli.
As a competitive inhibitor preventing FimH adhesin from binding to the uroplakin receptors, THP possesses several important properties. THP is present in all placental mammals ( 24, 25 ) and is synthesized by the renal tubular cells and excreted exclusively in the urine with a concentration reaching 0.1 mg/ml. It is conserved across species not only in its primary structure but also its high mannose moieties. Thus the high mannose glycosylation has been found in all species examined including murine, bovine, porcine, and human ( 38, 53, 56 ). The glycosylation of THP is so persistent that, even when transfected and expressed in cultured cells originally not synthesizing the THP, THP is still modified by the high mannoses ( 52 ). It has therefore been suggested that THP's high mannose glycoylation is determined by the fundamental protein structure and unique folding of the peptides surrounding the glycosylation site. Van Rooijen and co-workers ( 66 ) recently mapped the glycosylation site to Asn 251 in human THP. Using thin-layer chromatography and later 1 H NMR spectroscopy, Serafini-Cessi and colleagues ( 7, 51 ) established that native human THP bears predominantly the Man 6 -GlcNAc 2 high mannose type (73 vs. 19% Man 7 -GlcNAc 2 and 8% Man 5 -GlcNAc 2 ). Consistent with the structural data proving the existence of high mannoses in THP, type 1-fimbriated E. coli were found to bind in a mannose-specific fashion to highly purified THP ( 38 ). Given the large amounts of THP in the urine, the binding between THP and type 1 fimbriae can potentially saturate the adhesins of E. coli, thereby preventing the latter from binding to the uroplakin receptors. The major advantage of this type of defense mechanism is that it prevents E. coli from adhering to urothelial cells, thereby averting the serious consequences that follow adhesion.
It should be noted that the bladder colonization experiments described here are of a short-term nature. Long-term experiments should tell us whether THP deficiency also plays a role in the pathogenesis of chronic and recurrent bladder infections.
THP abnormalities, both quantitative and qualitative, have been associated with several pathological conditions in humans. A notable example is type I diabetes, in which patients have a profound reduction in urinary THP ( 1, 2, 62, 63 ). Diabetes is also frequently associated with an abnormal glycosylation pattern of the urinary proteins including the THP ( 8, 41, 61 ). It is therefore of particular interest to note that diabetic patients are extremely prone to UTIs, perhaps due to altered and reduced THP ( 13, 19, 47 ). Although diabetic patients also have granulocyte dysfunction, which could render a host more susceptible to UTIs, Langermann and co-workers ( 27 ) demonstrated that FimC/H vaccine was equally effective in protecting the bladder against infection by the type 1-fimbriated E. coli in both neutropenic and nonneutropenic mice. This suggests that THP dysfunction can contribute significantly to the propensity of diabetic patients to developing UTIs, independent of the neutrophil status. Urinary THP reduction has also been found in acute tubular necrosis, hyperprostaglandin E syndrome, and active lupus nephritis ( 30, 50, 64 ). The consequence of such a THP reduction in these pathological conditions, however, remains to be elucidated. Thus far, only a few studies have been attempted to correlate THP with human UTIs. In two separate studies, Reinhart and co-workers ( 43, 44 ) found that the concentrations of aggregated THP were dramatically decreased in children as well as in the elderly during the episodes of UTI. Despite a demonstrated correlation in these reports, more studies are clearly needed to establish a role for THP in human UTIs. It is well known that the majority of recurrent UTIs occur in anatomically normal individuals. We speculate that physiological defects in the components of innate urinary defenses, such as THP, may account for the predisposition of some individuals to recurrent UTIs.
In summary, we showed by a gene-knockout approach that THP deficiency can lead to increased bladder colonization by type 1-fimbriated E. coli. Our results indicate that THP can serve as an effective defense factor in the urinary tract by competitively inhibiting the FimH adhesin from binding to the uroplakin receptors. Because quantitative and qualitative defects of THP are known to exist in humans, it will be extremely interesting to determine whether these defects are the underlying causes of some of the UTIs, particularly recurrent UTIs in humans. Finally, THP has been proposed to play roles in regulating renal stone formation, immunosuppression, and salt and water balances ( 24, 25 ). The availability of the THP knockout mice provides new opportunities to explore other potential roles of the THP in urinary tract physiology and diseases.
ACKNOWLEDGMENTS
The authors thank Z.-T. Zhang and J. Pak for technical support and A. Auerbach and W. Auerbach for advice on ES cell work.
GRANTS
This work was supported in part by National Institutes of Health Grants DK-56903 to X.-R. Wu and AI-42886 to D. L. Hasty and a Merit Review award from the Veterans Administration Medical Research Service to X.-R. Wu.
【参考文献】
Below AA, Chakraborty J, Khuder SH, and Haselhuhn GD. Evaluation of urinary Tamm-Horsfall protein in postmenopausal diabetic women. J Diabetes Complications 13: 204-210, 1999.
Bernard AM, Ouled AA, Lauwerys RR, Lambert A, and Vandeleene B. Pronounced decrease of Tamm-Horsfall proteinuria in diabetics. Clin Chem 33: 1264, 1987.
Blomfield IC, McClain MS, and Eisenstein BI. Type 1 fimbriae mutants of Escherichia coli K12: characterization of recognized afimbriate strains and construction of new fim deletion mutants. Mol Microbiol 5: 1439-1445, 1991.
Brown PD, Freeman A, and Foxman B. Prevalence and predictors of trimethoprim-sulfamethoxazole resistance among uropathogenic Escherichia coli isolates in Michigan. Clin Infect Dis 34: 1061-1066, 2002.
Cavallone D, Malagolini N, and Serafini-Cessi F. Mechanism of release of urinary Tamm-Horsfall glycoprotein from the kidney GPI-anchored counterpart. Biochem Biophys Res Commun 280: 110-114, 2001.
Connell I, Agace W, Klemm P, Schembri M, Marild S, and Svanborg C. Type 1 fimbrial expression enhances Escherichia coli virulence for the urinary tract. Proc Natl Acad Sci USA 93: 9827-9832, 1996.
Dall'Olio F, de Kanter FJ, van den Eijnden DH, and Serafini-Cessi F. Structural analysis of the preponderant high-mannose oligosaccharide of human Tamm-Horsfall glycoprotein. Carbohydr Res 178: 327-332, 1988.
Dulawa J, Rambausek M, Jann K, Notohamiprodjo M, and Ritz E. Abnormal radiofurosemide binding by Tamm-Horsfall glycoprotein of diabetic patients. Diabetologia 28: 827-830, 1985.
Dziukas LJ, Sterzel RB, Hodson CJ, and Hoyer JR. Renal localization of Tamm-Horsfall protein in unilateral obstructive uropathy in rats. Lab Invest 47: 185-193, 1982.
Foxman B, Zhang L, Palin K, Tallman P, and Marrs CF. Bacterial virulence characteristics of Escherichia coli isolates from first-time urinary tract infection. J Infect Dis 171: 1514-1521, 1995.
Fukuoka S, Freedman SD, Yu H, Sukhatme VP, and Scheele GA. GP-2/THP gene family encodes self-binding glycosylphosphatidylinositol-anchored proteins in apical secretory compartments of pancreas and kidney. Proc Natl Acad Sci USA 89: 1189-1193, 1992.
Fukuoka S and Kobayashi K. Analysis of the C-terminal structure of urinary Tamm-Horsfall protein reveals that the release of the glycosyl phosphatidylinositol-anchored counterpart from the kidney occurs by phenylalanine-specific proteolysis. Biochem Biophys Res Commun 289: 1044-1048, 2001.
Geerlings SE, Erkelens DW, and Hoepelman IM. Urinary tract infections in patients with diabetes mellitus. Ned Tijdschr Geneeskd 141: 372-375, 1997.
Gunther IN, Snyder JA, Lockatell V, Blomfield I, Johnson DE, and Mobley HL. Assessment of virulence of uropathogenic Escherichia coli type 1 fimbrial mutants in which the invertible element is phase-locked on or off. Infect Immun 70: 3344-3354, 2002.
Hanson MS and Brinton CC Jr. Identification and characterization of E. coli type-1 pilus tip adhesion protein. Nature 332: 265-268, 1988.
Hawthorn LA, Bruce AW, and Reid G. Ability of uropathogens to bind to Tamm-Horsfall protein-coated renal tubular cells. Urol Res 19: 301-304, 1991.
Hession C, Decker JM, Sherblom AP, Kumar S, Yue CC, Mattaliano RJ, Tizard R, Kawashima E, Schmeissner U, Heletky S, Chow EP, Burne CA, Shaw A, and Muchmore A. Uromodulin (Tamm-Horsfall glycoprotein): a renal ligand for lymphokines. Science 237: 1479-1484, 1987.
Hooper LV and Gordon JI. Commensal host-bacterial relationships in the gut. Science 292: 1115-1118, 2001.
Joshi N, Caputo GM, Weitekamp MR, and Karchmer AW. Infections in patients with diabetes mellitus. N Engl J Med 341: 1906-1912, 1999.
Jovine L, Qi H, Williams Z, Litscher E, and Wassarman PM. The ZP domain is a conserved module for polymerization of extracellular proteins. Nat Cell Biol 4: 457-461, 2002.
Joyner AL. Gene Targeting: A Practical Approach, edited by Hames BD. Oxford, UK: Oxford Univ. Press, 1993, p. 101-131.
Kachar B, Liang F, Lins U, Ding M, Wu XR, Stoffler D, Aebi U, and Sun TT. Three-dimensional analysis of the 16 nm urothelial plaque particle: luminal surface exposure, preferential head-to-head interaction, and hinge formation. J Mol Biol 285: 595-608, 1999.
Klemm P, Schembri M, and Hasty DL. Toward Anti-Adhesin Therapy of Microbial Diseases, edited by Kahane I and Ofek I. New York: Plenum, 1996, p. 81-94.
Kokot F and Dulawa J. Tamm-Horsfall protein updated. Nephron 85: 97-102, 2000.
Kumar S and Muchmore A. Tamm-Horsfall protein-uromodulin (1950-1990). Kidney Int 37: 1395-1401, 1990.
Lamont JT. Mucus: the front line of intestinal mucosal defense. Ann NY Acad Sci 664: 190-201, 1992.
Langermann S, Palaszynski S, Barnhart M, Auguste G, Pinkner JS, Burlein J, Barren P, Koenig S, Leath S, Jones CH, and Hultgren SJ. Prevention of mucosal Escherichia coli infection by FimH-adhesin-based systemic vaccination. Science 276: 607-611, 1997.
Leeker A, Kreft B, Sandmann J, Bates J, Wasenauer G, Muller H, Sack K, and Kumar S. Tamm-Horsfall protein inhibits binding of S- and P-fimbriated Escherichia coli to human renal tubular epithelial cells. Exp Nephrol 5: 38-46, 1997.
Martinez JJ, Mulvey MA, Schilling JD, Pinkner JS, and Hultgren SJ. Type 1 pilus-mediated bacterial invasion of bladder epithelial cells. EMBO J 19: 2803-2812, 2000.
McLaughlin PJ, Aikawa A, Davies HM, Ward RG, Bakran A, Sells RA, and Johnson PM. Uromodulin levels are decreased in urine during acute tubular necrosis but not during immune rejection after renal transplantation. Clin Sci (Lond) 84: 243-246, 1993.
Measley RE Jr and Levison ME. Host defense mechanisms in the pathogenesis of urinary tract infection. Med Clin North Am 75: 275-286, 1991.
Min G, Stolz M, Zhou G, Liang F, Sebbel P, Stoffler D, Glockshuber R, Sun TT, Aebi U, and Kong XP. Localization of uroplakin Ia, the urothelial receptor for bacterial adhesin FimH, on the six inner domains of the 16 nm urothelial plaque particle. J Mol Biol 317: 697-706, 2002.
Mulvey MA. Adhesion and entry of uropathogenic Escherichia coli. Cell Microbiol 4: 257-271, 2002.
Mulvey MA, Lopez-Boado YS, Wilson CL, Roth R, Parks WC, Heuser J, and Hultgren SJ. Induction and evasion of host defenses by type 1-piliated uropathogenic Escherichia coli. Science 282: 1494-1497, 1998.
Mulvey MA, Schilling JD, and Hultgren SJ. Establishment of a persistent Escherichia coli reservoir during the acute phase of a bladder infection. Infect Immun 69: 4572-4579, 2001.
Mulvey MA, Schilling JD, Martinez JJ, and Hultgren SJ. Bad bugs and beleaguered bladders: interplay between uropathogenic Escherichia coli and innate host defenses. Proc Natl Acad Sci USA 97: 8829-8835, 2000.
Orskov I, Ferencz A, and Orskov F. Tamm-Horsfall protein or uromucoid is the normal urinary slime that traps type 1 fimbriated Escherichia coli. Lancet 1: 887, 1980.
Pak J, Pu Y, Zhang ZT, Hasty DL, and Wu XR. Tamm-Horsfall protein binds to type 1 fimbriated Escherichia coli and prevents E. coli from binding to uroplakin Ia and Ib receptors. J Biol Chem 276: 9924-9930, 2001.
Parkkinen J, Virkola R, and Korhonen TK. Identification of factors in human urine that inhibit the binding of Escherichia coli adhesins. Infect Immun 56: 2623-2630, 1988.
Pennica D, Kohr WJ, Kuang WJ, Glaister D, Aggarwal BB, Chen EY, and Goeddel DV. Identification of human uromodulin as the Tamm-Horsfall urinary glycoprotein. Science 236: 83-88, 1987.
Rambausek M, Dulawa J, Jann K, and Ritz E. Tamm-Horsfall glycoprotein in diabetes mellitus: abnormal chemical composition and colloid stability. Eur J Clin Invest 18: 237-242, 1988.
Reinhart H, Obedeanu N, Merzbach D, and Sobel JD. Effect of Tamm-Horsfall protein on chemiluminescence response of polymorphonuclear leukocytes to uropathogenic Escherichia coli. J Infect Dis 164: 404-406, 1991.
Reinhart HH, Obedeanu N, Robinson R, Korzeniowski O, Kaye D, and Sobel JD. Urinary excretion of Tamm-Horsfall protein in elderly women. J Urol 146: 806-808, 1991.
Reinhart HH, Spencer JR, Zaki NF, and Sobel JD. Quantitation of urinary Tamm-Horsfall protein in children with urinary tract infection. Eur Urol 22: 194-199, 1992.
Rindler MJ, Naik SS, Li N, Hoops TC, and Peraldi MN. Uromodulin (Tamm-Horsfall glycoprotein/uromucoid) is a phosphatidylinositol-linked membrane protein. J Biol Chem 265: 20784-20789, 1990.
Roberts JA. Factors predisposing to urinary tract infections in children. Pediatr Nephrol 10: 517-522, 1996.
Ronald A and Ludwig E. Urinary tract infections in adults with diabetes. Int J Antimicrob Agents 17: 287-292, 2001.
Ronald AR. Editorial comment: sexually transmitted diseases and urinary tract infections. Curr Opin Infect Dis 8: 1-3, 1995.
Schoolnik GK, O'Hanley P, Lark D, Normark S, Vosti K, and Falkow S. Uropathogenic Escherichia coli : molecular mechanisms of adherence. Adv Exp Med Biol 224: 53-62, 1987.
Schroter J, Timmermans G, Seyberth HW, Greven J, and Bachmann S. Marked reduction of Tamm-Horsfall protein synthesis in hyperprostaglandin E-syndrome. Kidney Int 44: 401-410, 1993.
Serafini-Cessi F, Dall'Olio F, and Malagolini N. High-mannose oligosaccharides from human Tamm-Horsfall glycoprotein. Biosci Rep 4: 269-274, 1984.
Serafini-Cessi F, Malagolini N, Hoops TC, and Rindler MJ. Biosynthesis and oligosaccharide processing of human Tamm-Horsfall glycoprotein permanently expressed in HeLa cells. Biochem Biophys Res Commun 194: 784-790, 1993.
Sherblom AP and Smagula RM. High-mannose chains of mammalian glycoproteins. Methods Mol Biol 14: 143-149, 1993.
Sikri KL, Foster CL, MacHugh N, and Marshall RD. Localization of Tamm-Horsfall glycoprotein in the human kidney using immunofluorescence and immunoelectron microscopical techniques. J Anat 132: 597-605, 1981.
Slomiany BL and Slomiany A. Role of mucus in gastric mucosal protection. J Physiol Pharmacol 42: 147-161, 1991.
Smagula RM, Van Halbeek H, Decker JM, Muchmore AV, Moody CE, and Sherblom AP. Pregnancy-associated changes in oligomannose oligosaccharides of human and bovine uromodulin (Tamm-Horsfall glycoprotein). Glycoconj J 7: 609-624, 1990.
Sokurenko EV, Chesnokova V, Doyle RJ, and Hasty DL. Diversity of the Escherichia coli type 1 fimbrial lectin. Differential binding to mannosides and uroepithelial cells. J Biol Chem 272: 17880-17886, 1997.
Sokurenko EV, Chesnokova V, Dykhuizen DE, Ofek I, Wu XR, Krogfelt KA, Struve C, Schembri MA, and Hasty DL. Pathogenic adaptation of Escherichia coli by natural variation of the FimH adhesin. Proc Natl Acad Sci USA 95: 8922-8926, 1998.
Sokurenko EV, Courtney HS, Maslow J, Siitonen A, and Hasty DL. Quantitative differences in adhesiveness of type 1 fimbriated Escherichia coli due to structural differences in fimH genes. J Bacteriol 177: 3680-3686, 1995.
Thankavel K, Madison B, Ikeda T, Malaviya R, Shah AH, Arumugam PM, and Abraham SN. Localization of a domain in the FimH adhesin of Escherichia coli type 1 fimbriae capable of receptor recognition and use of a domain-specific antibody to confer protection against experimental urinary tract infection. J Clin Invest 100: 1123-1136, 1997.
Torffvit O. Urinary sulphated glycosaminoglycans and Tamm-Horsfall protein in type 1 diabetic patients. Scand J Urol Nephrol 33: 328-332, 1999.
Torffvit O and Agardh CD. Tubular secretion of Tamm-Horsfall protein is decreased in type 1 (insulin-dependent) diabetic patients with diabetic nephropathy. Nephron 65: 227-231, 1993.
Torffvit O, Agardh CD, and Thulin T. A study of Tamm-Horsfall protein excretion in hypertensive patients and type 1 diabetic patients. Scand J Urol Nephrol 33: 187-191, 1999.
Tsai CY, Wu TH, Yu CL, Lu JY, and Tsai YY. Increased excretions of 2 -microglobulin, IL-6, and IL-8 and decreased excretion of Tamm-Horsfall glycoprotein in urine of patients with active lupus nephritis. Nephron 85: 207-214, 2000.
Tuomanen E. A spoonful of sugar to control inflammation? J Clin Invest 93: 917-918, 1994.
Van Rooijen JJ, Voskamp AF, Kamerling JP, and Vliegenthart JF. Glycosylation sites and site-specific glycosylation in human Tamm-Horsfall glycoprotein. Glycobiology 9: 21-30, 1999.
Wallace JL and Granger DN. The cellular and molecular basis of gastric mucosal defense. FASEB J 10: 731-740, 1996.
Walz T, Haner M, Wu XR, Henn C, Engel A, Sun TT, and Aebi U. Towards the molecular architecture of the asymmetric unit membrane of the mammalian urinary bladder epithelium: a closed "twisted ribbon" structure. J Mol Biol 248: 887-900, 1995.
Winberg J, Mollby R, Bergstrom J, Karlsson KA, Leonardsson I, Milh MA, Teneberg S, Haslam D, Marklund BI, and Normark S. The PapG-adhesin at the tip of P-fimbriae provides Escherichia coli with a competitive edge in experimental bladder infections of cynomolgus monkeys. J Exp Med 182: 1695-1702, 1995.
Wu XR, Sun TT, and Medina JJ. In vitro binding of type 1-fimbriated Escherichia coli to uroplakins Ia and Ib: relation to urinary tract infections. Proc Natl Acad Sci USA 93: 9630-9635, 1996.
Zhou G, Mo WJ, Sebbel P, Min G, Neubert TA, Glockshuber R, Wu XR, Sun TT, and Kong XP. Uroplakin Ia is the urothelial receptor for uropathogenic Escherichia coli : evidence from in vitro FimH binding. J Cell Sci 114: 4095-4103, 2001.
Zhu X, Cheng J, Gao J, Lepor H, Zhang ZT, Pak J, and Wu XR. Isolation of mouse THP gene promoter and demonstration of its kidney-specific activity in transgenic mice. Am J Physiol Renal Physiol 282: F608-F617, 2002.
作者单位：Departments of 1 Urology and 3 Microbiology, Kaplan Comprehensive Cancer Center, New York University School of Medicine, New York 10016; 4 Veterans Affairs Medical Center in New York, New York 10010; and 2 Department of Anatomy and Neurobiology, University of Tennessee, Memphis 38163 and Veterans Af