Proteus Mirabilis Flagella and MR/P Fimbriae

Home , Fimbria (bacteriology)

INFECTION AND IMMUNITY, OCt. 1991, p. 3574-3580 Vol. 59, No. 10 0019-9567/91/103574-07$02.00/0 Copyright C 1991, American Society for Microbiology Proteus mirabilis Flagella and MR/P Fimbriae: Isolation, Purification, N-Terminal Analysis, and Serum Antibody Response following Experimental Urinary Tract Infection FARAH K. BAHRANI, DAVID E. JOHNSON, DAVID ROBBINS, AND HARRY L. T. MOBLEY* Division ofInfectious Diseases, Department of Medicine, University of Maryland School of Medicine, 10 South Pine Street, Baltimore, Maryland 21201 Received 4 March 1991/Accepted 24 July 1991

Urinary tract infection with Proteus mirabilis may lead to serious complications, including cystitis, acute pyelonephritis, fever, bacteremia, and death. In addition to the production of hemolysin and the enzyme urease, fimbriae and flagellum-mediated motility have been postulated as virulence factors for this species. We purified mannose-resistant/proteuslike (MR/P) fimbriae and flagella from strains CFT322 and HU2450, respectively. Electron microscopy revealed highly concentrated preparations of fimbriae and flagella. Fimbrial and flagellar structural subunits were estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis to be 18.5 and 41 kDa, respectively. N-terminal sequencing revealed that 10 of the first 20 amino acids of the major MR/P subunit matched the sequence of the P. mirabilis uroepithelial cell adhesin N terminus and 11 of 20 amino acids matched the predicted amino acid sequence of the Escherichia coli P fimbriae structural subunit, PapA. In addition, 90 and 80% homologies were found between the first 20 amino acids ofP. mirabilis flagellin and those of Salmonella typhimurium phase-i flagellin and the E. coli hag gene product, respectively. An enzyme-linked immunosorbent assay using purified antigens showed a strong reaction between the MR/P fimbriae or flagella and sera of CBA mice challenged transurethrally with P. mirabilis. A possible role for MR/P fimbriae in the pathogenesis of urinary tract infection is supported by (i) a strong immune response to the antigen in experimentally infected animals, (ii) amino acid sequence similarity to other enteric surface structure, and (iii) our previously reported observation that MR/P fimbriae are expressed preferentially as the sole fimbrial type in human pyelonephritis isolates.

Proteus mirabilis, one of the most common uropathogens, cells (8, 21, 42, 49). Heavily fimbriated organisms were is frequently isolated from the urine of elderly long-term shown to initiate infection better than lightly fimbriated catheterized patients (26, 46). Infection with this organism organisms. Principally, two fimbrial types, mannose-resis- can result in serious complications, such as stone formation tant/proteuslike (MR/P) and mannose-resistant/klebsiellalike in the kidneys and bladder (26), encrustation and obstruction (MR/K), have been associated with the ability to hemagglu- of the catheter (27), acute pyelonephritis (6, 7, 36), fever, tinate untreated or tannic acid-treated erythrocytes from and bacteremia (40). several animal species (1, 30, 31). Silverblatt and Ofek (42) Several virulence factors have been proposed to play a found that when rats were inoculated transurethrally with P. role in the pathogenesis of P. mirabilis infection, including mirabilis expressing either MR/P or MR/K fimbriae, inocula hydrolysis of urea by urease (5, 9, 15, 22), cytotoxicity by expressing MR/P fimbriae caused a higher frequency of hemolysin (25, 33, 39), avoidance of host defense by immu- cortical abscesses at 1 week than inocula expressing pre- noglobulin A protease (38), destruction of kidney epithelial dominantly MR/K fimbriae. This observation suggested that cells by invasion (34), swarming motility by flagella (4, 32), MR/P fimbriae may play an important role in the develop- and adherence to uroepithelial cells and kidney epithelium ment of pyelonephritis. by fimbriae (21, 37, 42) and other adhesins (49). Our objectives were to isolate cell surface organelles of P. The role of swarming motility in the pathogenesis of P. mirabilis and to obtain amino acid sequence data that will be mirabilis has not been well studied. However, Pazin and useful for the isolation of MR/P and flagellin gene sequences. Braude (32) demonstrated that immobilizing antibodies, In addition, we wished to determine whether these antigens raised in response to immunization with purified flagella, were recognized by the immune system during experimental prevented the spread ofP. mirabilis from one rat kidney into infection. In this study, flagella and MR/P fimbriae were which organisms had been inoculated to the other kidney via purified from two uropathogenic strains of P. mirabilis and the ureters. Motility also appears to be critical for the the N-terminal amino acid sequences of the structural sub- successful infection of burn wounds by P. mirabilis (10) and units of both organelles were determined. CBA mice in- Pseudomonas aeruginosa (28). fected transurethrally with P. mirabilis HI4320 (MR/P' A few reports suggest that the adherence ofP. mirabilis to flagella') showed a serum immune response to both antigens uroepithelial cells contributes importantly to pathogenesis. when assayed by enzyme-linked immunosorbent assay In an animal model of unobstructed ascending infection, P. (ELISA). mirabilis appeared to adhere to renal epithelial cells by fimbriae (41, 42), as well as in vitro to human uroepithelial MATERIALS AND METHODS

Bacterial strains. P. mirabilis CFT322 was isolated from * Corresponding author. the urine of a patient with the clinical symptoms of acute 3574 VOL. 59, 1991 P. MIRABILIS FLAGELLA AND MR/P FIMBRIAE 3575 pyelonephritis who was admitted to the University of Mary- from crude extracts by treating them with bovine pancreatic land Medical Systems Hospital in Baltimore (24). P. mirab- RNase (40 ,ug/ml), DNase I (100 jig/ml), and proteinase K (10 ilis HU2450 and HI4320 were isolated from the urine of two j,g/ml) (all obtained from Sigma) at 60°C for 1 h. Further elderly (.65 years old) long-term catheterized patients with purification was performed by repeated ultracentrifugation significant bacteriuria (2105 CFU/ml) (27, 46). At the time of at 105,000 x g for 2 h, until the optical density of the isolation, these two patients were asymptomatic. supernatant was less than 0.01 at 260 and 280 nm (12). The Hemagglutination. Bacteria were passaged statically three purified LPS was dissolved in water and lyophilized. times for 48 h in nutrient broth at 37°C (30) and then Electron microscopy. One drop ofthe specimen was placed harvested by centrifugation (5,000 x g, 10 min, 4°C). Cell on a Formvar-coated grid and negatively stained with 1% pellets were resuspended in phosphate-buffered saline (PBS; sodium phosphotungstic acid (pH 6.8). The specimens were per liter, 8 g of NaCl, 0.2 g of KCl, 1.44 g of Na2HPO4, and then examined in a JEOL 100B transmission electron micro- 0.24 g of KH2PO4; pH 7.3) to approximately 109 CFU/ml and scope at an accelerating voltage of 80 kV. then mixed with an equal volume of a 3% (vol/vol) erythro- SDS-PAGE. Flagellum preparations (10 ;ig) were sus- cyte suspension. Hemagglutination was defined as visible pended in the sodium dodecyl sulfate (SDS) gel sample clumping of erythrocytes as follows: MR/K hemagglutina- buffer and treated at 100°C for 10 min. Preparations of MR/P tion was shown by the agglutination of tannic acid-treated ox fimbriae were not denatured under these conditions and erythrocytes but not untreated ox erythrocytes, and MR/P therefore did not run into the gel. Treatment with 10% hemagglutination was demonstrated by the hemagglutination trichloroacetic acid was necessary for the denaturation of of untreated and tannic acid-treated chicken, sheep, guinea MR/P fimbriae into fimbrial subunits. SDS-polyacrylamide pig, horse, ox, and human erythrocytes, with chicken cells gel electrophoresis (SDS-PAGE) was performed with a giving the strongest reaction. The reactions were not in- 1.5-mm 3.75% stacking gel and a 15 or 12.5% polyacrylamide hibitable by 50 mM mannose. Ox, guinea pig, horse, and resolving gel and with the buffer system described by Laemm- sheep erythrocytes were purchased from Cleveland Scien- li (20). Protein bands were stained with Coomassie brilliant tific Co., Cleveland, Ohio. Fowl erythrocytes were obtained blue (Bio-Rad). from Pelfreez Biological, Rogers, Ark. Human erythrocytes N-terminal analysis. Solubilized flagella (10 ,ug) or MR/P were provided by H. L. T. Mobley. fimbriae (10 jg) were electrophoresed on an SDS-10% Isolation and purification of flagella. Flagella were pre- polyacrylamide gel. Protein subunits were transferred to pared by differential centrifugation as described by Montie et polyvinylidene difluoride membrane (Immobilon-P; Mili- al. (28). An overnight culture (300 ml) of P. mirabilis pore Corp., Bedford, Mass.) and stained with Coomassie HU2450 was used to inoculate 60 Luria agar plates (2% agar, brilliant blue. Stained bands were excised and subjected to 15 by 150 mm). Cells were grown for 48 h at 37°C, harvested N-terminal amino acid analysis on an Applied Biosystems with a glass rod in 10 mM potassium phosphate (pH 7.0), pulsed liquid-phase sequencer (model 477A) (29). centrifuged (5,000 x g, 15 min, 4C), washed, and resus- CBA mouse model of ascending urinary tract infection. pended in the phosphate buffer (100 ml/6 g [wet weight] of Eleven 6- to 8-week-old female inbred CBA mice (Jackson cells). The suspension was blended at setting 4 in a Waring Laboratory, Bar Harbor, Maine) were challenged transure- commercial blender (model 34BL97) for 5 min and then thrally with P. mirabilis HI4320 (108 CFU in 0.1 ml of PBS) centrifuged (16,000 x g, 15 min, 4°C). The supernatant was as previously described (13). Serum was collected from the centrifuged (40,000 x g, 3 h, 4°C), and the pellet was washed tail vein prior to challenge and after 2 days and 1, 2, 3, 4, 5, twice in phosphate buffer and centrifuged (40,000 x g, 2 h, and 6 weeks. 4°C). ELISA. The standard procedure of Voller et al. (45) was For further purification, flagellum-containing pellets were used for the ELISA. Briefly, purified flagella and purified suspended in the phosphate buffer, to which was added CsCl MR/P fimbrial antigens were suspended in 0.05 M carbonate- (1.3 g/ml) and 1.5% sodium N-lauryl sarcosine, and centri- bicarbonate buffer (pH 9.6) and used to coat the wells of fuged (117,000 x g, 44 h, 20°C) in an SW41 Ti rotor. Visible microtiter plates (Cooke Dyna Tech Labs) (100 ng of protein protein bands were collected by needle aspiration and dia- per well), and they were incubated overnight at 4°C. Plates lyzed overnight against distilled water at room temperature. were washed with PBS and blocked with 5% fetal bovine Isolation and purification of MR/P fimbriae. Luria broth (6 serum in PBS (1 h at 37°C). Plates were washed, and mouse liters) was inoculated with a 50-ml stationary culture of P. sera were added at dilutions of 1/50, 1/100, and 1/200 and mirabilis CFT322 which had been passaged statically three incubated at 37°C for 1 h. After a washing with PBS-Tween, times for 48 h at 37°C (30), conditions which favor the an alkaline phosphatase derivative of a rabbit anti-mouse expression of MR/P fimbriae. After 48 h of static growth at polyvalent imnmunoglobulin (Sigma) was added, and the 37°C, cells were harvested and washed by centrifugation, plates were incubated at 37°C for 1 h. The substrate, suspended in 250 ml of 10 mM Tris HCl (pH 7.2), sheared by p-nitrophenylphosphate, was added, and after an incubation blending (5 min on setting 4), and centrifuged (8,000 x g, 20 (37°C for 30 min), the reaction was stopped with 3 M NaOH min, 4°C). The supernatant was centrifuged (39,000 x g, 30 and absorbance was read at 405 nm. min, 4C), and the supernatant from this spin was centri- Immobilization assay. Mouse serum (1/25 dilution) was fuged again (180,000 x g, 1.5 h, 4°C). The pellet was filter sterilized through a Low Protein Binding Filter (0.2-,um resuspended in and dialyzed against distilled water and pore diameter; Acrodisc, Gelman Sciences) and then mixed centrifuged in an SW41 Ti rotor (117,000 x g, 44 h, 20°C) on with an equal volume of 2x concentrated motility agar (final a CsCl (1.3 g/ml) gradient containing 1.5% sodium N-lauryl concentration: 1 g of tryptone, 0.5 g of NaCl, and 0.5 g of sarcosine. Bands were extracted by needle aspiration and agar [each per 100 ml]). P. mirabilis, grown overnight on a dialyzed against distilled water at room temperature (43). Luria agar plate, was used to inoculate tubes of agar by LPS purification. Lipopolysaccharide (LPS) was prepared stabbing. Growth was monitored at 4, 8, and 24 h. frotn P. mirabilis H14320 by hot phenol-water extraction as Western blotting. Purified proteins and LPS were electro- described by Westphal and Jann (48), using acetone-dried phoresed on 10 to 15% gradient polyacrylamide gels and cells. RNA, DNA, and protein contamination was removed blotted onto nitrocellulose. Western blots (immunoblots) 3576 BAHRANI ET AL. INFECT. IMMUN.

A97 vf * >1 ' l t $ot I 9g7 .....8 . A 6

66 A5' :" , ^$~~~ 43

tSi .. 4 31: :.U r.,41 'II...,. 43'4'"I :. ._ ka ;i. '..::

4. .1, T

22_ 31 _Ue

1441W.w FIG. 1. Purified P. mirabilis flagellin. (A) Purified protein (10 ,ug) was denatured in gel sample buffer and analyzed by SDS-PAGE (15% polyacrylamide). Standard proteins (Bio-Rad) were hen egg 22-40~ white lysozyme (14.4 kDa), soybean trypsin inhibitor (21.5 kDa), bovine carbonic anhydrase (31 kDa), hen egg white ovalbumin (42.7 kDa), bovine serum albumin (66.2 kDa), and rabbit muscle phos- phorylase b (97.4 kDa). (B) Purified flagella were negatively stained with 1% sodium phosphotungstic acid and observed by electron 14 microscopy at a magnification of x60,000. FIG. 2. Purified P. mirabilis MR/P fimbriae. (A) Purified protein (10 p.g) was precipitated in 10% trichloroacetic acid, denatured in gel sample buffer, and analyzed by SDS-PAGE (12.5% polyacryl- were developed as described by Towbin et al. (44) with the amide). Standard proteins are the same as those described in the modification of Batteiger et al. (3). Nitrocellulose filters were legend to Fig. 1. (B) Purified fimbriae were negatively stained with incubated in 1:200 dilutions of mouse serum for 1 h. After sodium phosphotungstic acid and observed by electron microscopy being washed in 0.05% Tween 20 in PBS, filters were at a magnification of x60,000. incubated in conjugated goat anti-mouse immunoglobulin (diluted 1:1,000) for 1 h. Filters were washed, and the color reaction was developed with Nitro Blue Tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate. the main 18.5-kDa fimbrial subunit. These bands could be attributed to outer membrane proteins, or, perhaps more likely, they represent accessory polypeptides and minor RESULTS pilins associated with enteric fimbriae such as those observed following the purification of the P fimbriae of uro- Strains used for purification of flagella and fimbriae. On the pathogenic Escherichia coli (35). basis of electron microscopic observations and hemaggluti- N-terminal analysis of P. mirabilis flagellar and MR/P nation pattern, strain CFT322 was found to be heavily fimbrial subunits. Of the first 20 amino acids of the flagellin fimbriated (MR/P+ MR/K-) and lightly flagellated and was subunit determined (Fig. 3), 18 residues were exact matches used for the preparation of fimbriae. Strain HU2450 was flagellated but not fimbriated and was used for the preparation of flagella. This selection prevented cross-contamina- fimbrial Strain HI4320 was tion of flagellar and preparations. 1 20 heavily flagellated and fimbriated (MR/P+ MR/K+) and was used for transurethral challenge in mice. S. typhnurin Ala Gn Val li Asn Tlr Asn Ser Leu Ser Leu Leu ThrGkn Asn Asn Lou Asn Lys Ser Characterization of flagella and MR/P fimbriae. Electron microscopy revealed highly concentrated flagellar (Fig. 1B) P. n*abm Ala Gn Val li Asn Tr Asn Tyr Leu Ser Leu Val Tr Gn Asn Asn Leu Asn Lys Ser and fimbrial (Fig. 2B) preparations, usually with no particu- late contamination. Sarcosine was found to be helpful in the E coa a Gi Val le AsnmrMAsn Ser Leu Ser Lou le TrGh Asn Asn le Asn Lys Asn removal of membrane vesicles. Flagella disaggregated in upon readily SDS and ,B-mercaptoethanol boiling, but :oxci ainho acid match wh P. miabis flgean MR/P fimbriae were not solubilized under these conditions consovafive amino acid and therefore did not enter the stacking gel. However, replacement FIG. 3. Comparison of N-terminal amino acid sequence of P. precipitation in 10% trichloroacetic acid followed by boiling mirabilis flagellar subunits with those of S. typhimurium phase-1 i in SDS gel sample buffer disaggregated the MR/P fimbriae. flagellin and E. coli hag gene product. Solubilized proteins were Electrophoretic analyses showed apparent molecular masses analyzed by SDS-PAGE, transferred to Immobilon-P, and subjected of 41 kDa for flagellin and 18.5 kDa for the MR/P fimbrial to N-terminal analysis. The Proteus sequence is compared with structural subunit (Fig. 1A and 2A). Occasionally, SDS- published sequences for the flagella of S. typhimurium (17, 47) and PAGE of fimbriae revealed several faint bands in addition to E. coli (18). VOL. 59, 1991 P. MIRABILIS FLAGELLA AND MR/P FIMBRIAE 3577 8- MrpA 1 Ala Ser Asp Gly TNTlierlb PheTr Gl Lys Val Ie s2 Ala Pro Cys GlyE[ 20 a 0 A UCA 1 TrGlyTv lbThr TrTPhe Gly Lys Val VW Aa G lThr X Ser Val 18 23 PapA 27 Pro Gin Gly Va]hrPVa4Asn Pro Cysoe 46 6- 0 8) K99 1 GlyrTlA1ieAsnrLyS 18 a 0 0 0 FIG. 4. Comparison of N-terminal amino acid sequence of P. 4- mirabilis MR/P fimbrial subunit with other fimbrial subunits and adhesins. Solubilized proteins were denatured, analyzed by SDS- 0 PAGE, transferred to Immobilon-P, and subjected to N-terminal -J analysis. The P. mirabilis MrpA sequence is compared with the 2- predicted amino acid sequence of the E. coli P fimbria major pilin, PapA (2), the N-terminal sequence of the P. mirabilis uroepithelial cell adhesin (UCA; 49), and the sequence of the E. coli K99 fimbrial subunit (49). 0 10 20 30 40 50 2.0 - B with those of Salmonella typhimurium phase-1 i flagellin (17, C 0 47) and 2 residues represented conservative replacements. E 0 Also, 16 of 20 residues were exact matches with those of the a E. coli K-12 hag gene product; 4 residues were conserva- an._.0 . tively replaced (18) (Fig. 3). Moreover, the first amino acid C 1.0 0 of the P. mirabilis flagellar subunit aligned with the first 0 amino acid of E. coli flagellin and of S. typhimurium phase-1 0 i flagellin. This suggests that the encoded amino terminal 0 5 sequences of these flagellins are the same as those of the HI A~~~~~~ mature proteins and, therefore, that no signal peptide, typi- I- cal of other exported proteins, is present. The N-terminal analysis of the first 20 amino acids of the 0 10 23 30 40 so MR/P fimbrial subunit revealed a stretch of 10 amino acids c (residues 4 to 13) that matched exactly with 10 amino acids is (residues 2 to 11) of the P. mirabilis uroepithelial cell adhesin c 0 isolated by Wray et al. (49) (Fig. 4). In addition, the MR/P La A fimbrial subunit showed 11 exact matches with the amino 0 0 0%.o acid sequence predicted from papA, the gene encoding the C structural subunit of P fimbriae of pyelonephritogenic E. coli. The first amino acid of the MR/P subunit corresponded UN. to the 27th amino acid of the predicted sequence of PapA (2). a.r0- These data suggest that the native MR/P fimbrial subunit may also possess a signal sequence which is cleaved prior to 0 0 secretion and assembly. Also, the MR/P fimbrial subunit shared nine exact amino acid matches with the K99 fimbrial subunit isolated from the diarrheagenic E. coli that afflicts 0 10 20 30 40 calves. 90 Immune response to flagella and MR/P fimbriae following Days after Challenge FIG. 5. Serum antibody response to MR/P fimbriae and flagella experimental infection. Eleven CBA mice inoculated with 108 in mice transurethrally inoculated with P. mirabilis H14320. (A) CFU ofP. mirabilis H14320 developed significant bacteriuria Quantitative urine cultures were taken from seven CBA mice (.106 CFU/ml of urine) which was maintained in all mice for transurethrally inoculated with 108 CFU of P. mirabilis H14320. The 6 days. Three mice died by day 7, and a fourth died by day line connects the geometric means of log1o CFU/ml of urine. (B) 9. There was no additional mortality throughout the 6-week Serum antibody levels to purified MR/P fimbriae were measured by course. Colony counts, shown in Fig. 5A for the surviving ELISA. Sera from seven mice were tested at all time points. (C) seven mice, dropped below 104 CFU/ml of urine between Serum antibody levels to purified flagella were measured by ELISA. days 10 and 14. Two mice remained heavily colonized for The lines in panels B and C connect the arithmetic means of the most of the 6-week course, whereas the five remaining mice optical densities at 405 nm. For all three graphs, each mouse is appeared to clear the infection. identified by the same unique symbol. Serum collected from mice at intervals after experimental urinary tract infection showed elevated serum immunoglobulin to both antigens when compared with prechallenge demonstrated the strongest immunoglobulin responses. The serum. ELISAs using purified MR/P fimbriae (Fig. 5B) and reaction of prechallenge sera against both antigens was flagella (Fig. 5C) as antigens were used to measure the negligible. mouse serum immunoglobulin levels of the seven mice that Serum (6-week postchallenge) from one experimentally survived the 6-week period. The antibody response (mean infected mouse was tested by Western blot and found to ELISA values) to MR/P fimbriae steadily rose throughout react with denatured flagellar preparations (Fig. 6). A lower- the 6 weeks. The flagellar response peaked at 1 week but molecular-weight band not visible on Coomassie blue- maintained significant levels throughout the course of the stained gels was also recognized. In addition, purified LPS infection. The two mice that developed chronic infections from P. mirabilis H14320 was recognized by the serum. 3578 BAHRANI ET AL. INFECT. IMMUN.

DISCUSSION A B C P. mirabilis, a very common uropathogen in catheterized elderly patients, can cause acute pyelonephritis and renal stone formation (26, 36). Also, this species is a prevalent cause of urinary tract infection in acute care hospitals, with the assumed risk of gram-negative bacteremia and death. Unlike the well-studied uropathogenic E. coli, not all virulence factors ofP. mirabilis have necessarily been identified. '46 However, using isogenic strains, we have previously demonstrated that the urease of this species appears absolutely essential for the establishment of acute pyelonephritis in the kidneys of CBA mice after transurethral challenge (15). Although urease is necessary for pathogenesis, it may indeed *30 not be sufficient. That is, other proposed virulence determinants may be required in the process. Therefore, we have begun to examine two additional putative virulence factors, MR/P fimbriae, which may aid in colonization, and flagella, *21.5 which may allow the bacteria to ascend via the ureters to the kidneys. We report here that both the MR/P fimbriae and flagella of P. mirabilis are well-recognized antigens in experimental .14.3 urinary tract infections. CBA mice infected transurethrally with P. mirabilis developed significant immunoglobulin titers to these antigens after infection. Levels of serum antibody to MR/P fimbriae and flagella rose most dramatically in the chronically infected mice, that is, in those mice that were *6.5 unable to clear the infection during the 6-week period that was monitored with quantitative urine cultures. Previous studies by our group indicate that the expression .3.4 of MR/P fimbriae as a single fimbrial type is epidemiologi- cally correlated with strains isolated from patients with acute pyelonephritis (24). That is, pyelonephritogenic strains were FIG. 6. Western blot of purified MR/P fimbriae and flagella, much more likely to express MR/P alone than were fecal using serum from a mouse transurethrally infected with P. mirabilis strains or those strains isolated from patients with catheter- H14320. Purified LPS (20 ,ug) (A), purified flagella (20 ,ug) (B), associated bacteriuria, which more often expressed both and purified MR/P fimbriae (20 ,ug) (C) were electrophoresed MR/P and MR/K fimbriae. These studies implied that MR/K on an SDS-polyacrylamide gel, electroblotted to nitrocellulose, fimbriated organisms may be cleared more rapidly from the and reacted with a 1/200 dilution of 6-week postchallenge mouse urinary tract. serum. Western blots were developed as described in Materials and Methods. Standard proteins (Amersham) are ovalbumin (46 Morphologically, these fimbriae resemble typical fimbriae kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (21.5 kDa), isolated from other representatives of the family Enterobac- lysozyme (14.3 kDa), aprotinin (6.5 kDa), and insulin B chain (3.4 teriaceae (23). The major subunit of the structure, likely to kDa). be designated MrpA, has an apparent molecular mass of 18.5 kDa. This value is typical of enteric fimbriae and falls well within the range of other major fimbrial subunits thus far described. This subunit size is similar to that of the uroepithelial cell adhesin (17.5 kDa) described by Wray et al. (49), However, purified flagellar and fimbrial preparations did not and the fimbrial subunit shares with this adhesin a homolo- appear to contain contaminating LPS. The serum recognized gous 10-amino-acid sequence at the N termini. The resis- native, nondenatured fimbriae, but it did not react with tance of MR/P fimbriae to dissociation into subunits by SDS TCA-denatured fimbrial subunits. No contaminating poly- is reminiscent of type I pili of E. coli, which resist various peptides in the denatured fimbrial preparations were recog- chaotropic agents such as urea and SDS (21). nized by the serum. Recently Sareneva et al. (37) isolated MR/P fimbriae from To determine whether mice challenged with P. mirabilis a clinical strain of P. mirabilis and reported a single peptide H14320 developed immobilizing antibodies, motility agar with an apparent molecular mass of 21 kDa. This discrep- containing mouse sera (1/50 final dilution) from prechal- ancy with our result (18.5 kDa) indicates that although lenged and 6-week postchallenged mice was prepared. Tubes fimbriae have functional and morphological similarity, struc- were stabbed with P. mirabilis HI4320. In tubes prepared turally they may be heterogeneous. Because the receptor for with no antiserum or prechallenge (optical density at 405 nm MR/P fimbriae has not been identified, the possibility exists of 1/50-dilution ELISA, 0.072) antiserum, spreading of bac- that the mannose-resistant hemagglutination observed may teria from the stab was observed beginning at 4 h and actually include several distinct fimbrial types. This situation resulted in high turbidity throughout the tube at 24 h. In may be similar to that of the fimbriae of E. coli associated tubes containing 6-week postchallenge serum (optical den- with mannose-resistant hemagglutination, which have now sity at 405 nm of 1/50-dilution ELISA, 0.708), bacteria were been subclassified as P, F, S, Dr, M, and G (14). Alterna- retained within the stab at 4 and 8 h, and minimal spreading tively, MR/P may represent fimbriae with identical receptor was observed at 24 h. specificity, but they may display antigenic diversity with VOL. 59, 1991 P. MIRABILIS FLAGELLA AND MR/P FIMBRIAE 3579 respect to the major pilin, MrpA. Our studies demonstrate, for the isolation of genes encoding these putative virulence however, that experimental infection with strain H14320 determinants. results in a serum antibody response to the purified MR/P fimbriae of strain CFT322. These limited observations, of ACKNOWLEDGMENTS course, do not allow us to draw conclusions about the We thank David Maneval for helpful advice on fimbrial purifica- antigenic relatedness of all MR/P fimbriae; therefore, appro- tion, Dan Edelman for technical assistance, and Ben Tall for priate caution should be taken in interpreting these results. assisting with electron microscopy. Significant conservation of the N-terminal amino acid This work was supported in part by Public Health Service grants sequence of the P. mirabilis MR/P fimbrial subunit compared Al 23328 and AG 04393 from the National Institutes of Health. with that of the P. mirabilis uroepithelial cell adhesin described by Wray et al. (49) is observed. These two polypep- REFERENCES tides share a stretch of 10 identical amino acids. There were 1. Adegbola, R. A., D. C. Old, and B. W. Senior. 1983. The no other matches adhesins and fimbriae of Proteus mirabilis strains associated outside of this region, suggesting that these with high and low affinity for the urinary tract. J. Med. Micro- polypeptides are distinct in origin but simply share a con- biol. 16:427-431. served sequence. Amino acid similarity was also seen when 2. Baga, M., S. Normark, J. Hardy, P. O'Hanley, D. Lark, 0. the MR/P subunit was compared with the predicted amino Olsson, G. Schoolnik, and S. Falkow. 1984. Nucleotide sequence acid sequence of an E. coli P fimbrial subunit, suggesting that of the papA gene encoding the Pap pilus subunit of human the genes encoding these polypeptides may have originated uropathogenic Escherichia coli. J. Bacteriol. 157:330-333. from a common ancestral gene. Homology at the N terminus 3. Batteiger, B., W. J. Newhall, and R. B. Jones. 1982. The use of is not unusual for numbers of of E. fimbrial Tween-20 as a blocking agent in the immunological detection of types coli proteins transferred to nitrocellulose membranes. J. Immunol. subunits (16); however, this sequence of MrpA shared Methods 55:297-307. homology only with uroepithelial cell adhesin, PapA, and the 4. Bisset, K. A., and C. W. I. Douglas. 1976. A continuous study of K99 fimbrial subunit. morphological phase in the swarm of Proteus. J. Med. Micro- We observed that the homology of the amino acid se- biol. 9:229-231. quence native MrpA and the predicted amino acid sequence 5. Braude, A. I., and J. Siemienski. 1960. Role of bacterial urease of PapA begins at the 27th amino acid of the Proteus fimbrial in experimental pyelonephritis. J. Bacteriol. 80:171-179. subunit. This is consistent with the fact that the first 26 6. Eriksson, S., J. Zbornik, H. Dahnsjo, P. Erlanson, 0. Kahlme- ter, H. Fritz, and C. A. Bauer. 1986. The combination of amino acids of PapA represent the signal sequence which is pivampicillin and pivmecillinam versus pivampicillin alone in cleaved during the secretion and assembly of P fimbriae. It is the treatment of acute pyelonephritis. Scand. J. Infect. Dis. likely that MrpA also has a signal sequence which is proc- 18:431-438. essed during secretion and assembly. Confirmation of this 7. File, T., J. Tan, S.-J. Salstrom, and J. Johnson. 1985. Timentin supposition awaits nucleotide sequence data. versus piperacillin in the therapy of serious urinary tract infec- Swarming motility has been shown to be a means of tions. Am. J. Med. 79:91-95. dissemination for P. a process for suc- 8. Fowler, J. E., Jr., and T. A. Stamey. 1978. Studies of introital mirabilis, required colonization in women with recurrent urinary infections. X. cessful infection. Pazin and Braude (32) demonstrated that Adhesive properties of Escherichia coli and Proteus mirabilis: when rats were immunized with flagellar preparations, the lack of correlation with urinary pathogenicity. J. Urol. 120:315- spread of P. mirabilis from one kidney directly inoculated 318. with bacteria to the other kidney via the ureter was pre- 9. Griffith, D. P., D. M. Musher, and C. Itin. 1976. Urease: the vented. This observation suggested a role for immobilizing primary cause of infection-induced urinary stones. Invest. Urol. antibodies in protecting against Proteus pyelonephritis. 13:346-350. Also, monoclonal antibodies specific for flagella have been 10. Harmon, R. C., R. L. Rutherford, H.-M. Wu, and M. S. Collins. shown to be in burn wound 1989. Monoclonal antibody-mediated protection and neutraliza- highly protective experimental tion of motility in experimental Proteus mirabilis infection. infections caused by P. mirabilis (10). Our results indicate Infect. Immun. 57:1936-1941. the recognition of flagellar antigens expressed on the surface 11. Homma, M., H. Fujita, S. Yamaguchi, and T. Lino. 1987. of P. mirabilis in experimental urinary tract infections, Regions of Salmonella typhimurium flagellin essential for its which suggests that immobilizing antibodies may be elicited polymerization and excretion. J. Bacteriol. 169:291-296. during natural infection with this species. 12. Inzana, T. J., B. Iritani, R. P. Gogolewski, S. A. Kanial, and Observation of over 80% homology between the N termini L. B. Corbell. 1988. Purification and characterization of lipooli- of the flagellin subunits ofP. mirabilis and those of two other gosaccharides from four strains of "Haemophilus somnus." members of the family Enterobacteriaceae suggests evolu- Infect. Immun. 56:2830-2837. conservation at the N terminus. This is known 13. Johnson, D. E., C. V. Lockatell, M. Hall-Craigs, H. L. T. tionary region Mobley, and J. W. Warren. 1987. Uropathogenicity in rats and to be essential for transport (19) and filament construction in mice of Providencia stuartii from long-term catheterized pa- S. typhimurium (11) and E. coli (18); therefore, this conser- tients. J. Urol. 138:632-635. vation is not surprising. Further characterization of P. 14. Johnson, J. R. 1991. Virulence factors in Escherichia coli mirabilis flagellin subunits will require complete amino acid urinary tract infection. Clin. Microbiol. Rev. 4:80-128. or nucleotide sequencing. 15. Jones, B. D., C. V. Lockatell, D. E. Johnson, J. W. Warren, and These studies provide a characterization of two putative H. L. T. Mobley. 1990. Construction of a urease negative mutant virulence factors of uropathogenic P. mirabilis and demon- of Proteus mirabilis: analysis of virulence in a mouse model of strate that these proteins elicit immune response during ascending urinary tract infection. Infect. Immun. 58:1120-1123. infection. the N termini of the 16. Jones, G. W., and R. E. Jaacson. 1983. Proteinaceous bacterial experimental Also, polypep- adhesions and receptors. Crit. Rev. Microbiol. 10:229-260. tides (MR/P and flagellin) share homologous sequences with 17. Joys, T. M. 1985. The covalent structure of the phase-1 flagellar the N termini of (i) the uroepithelial cell adhesin of P. filament protein of Salmonella typhimurium and its comparison mirabilis; (ii) P fimbriae, a critical virulence determinant of with other flagellins. J. Biol. Chem. 260:15758-15761. uropathogenic E. coli; and (iii) the flagella of both E. coli and 18. Kuwajima, G. 1988. Construction of a minimum-size functional S. typhimurium. These data provide necessary information flagellin of Escherichia coli. J. Bacteriol. 170:3305-3309. 3580 BAHRANI ET AL. INFECT. IMMUN.

19. Kuwajima, G., I. Kawagishi, M. Homma, E. Asaka, and R. M. ization of P fimbriae on 01, 07, 075, rough, and nontypable Macnab. 1989. Export of an N-terminal fragment of Escherichia strains of Escherichia coli. Infect. Immun. 56:1288-1294. coli flagellin by a flagellum-specific pathway. Proc. Natl. Acad. 36. Rubin, R. H., N. E. Tolkoff-Rubin, and R. S. Cotran. 1986. Sci. USA 86:4953-4957. Urinary tract infection, pyelonephritis, and reflux nephropathy, 20. Laemmli, U. K. 1970. Cleavage of structural proteins during the p. 1085-1141. In B. M. Brenner and F. C. Rector (ed.), The assembly of the head of bacteriophage T4. Nature (London) kidney. W. B. Saunders Co., Philadelphia. 227:680-685. 37. Sareneva, T., H. Holthofer, and T. K. Korhonen. 1990. Tissue- 21. Lomberg, H., P. Larsson, H. Leffler, and C. Svanborg-Eden. binding affinity of Proteus mirabilis fimbriae in the urinary tract. 1982. Different binding specificities of P. mirabilis compared to Infect. Immun. 58:3330-3336. E. coli. Scand. J. Infect. Dis. Suppl. 33:37-42. 38. Senior, B. W., M. Alberechtsen, and M. A. Kerr. 1987. Proteus 22. MacLaren, D. M. 1968. The significance of urease in Proteus mirabilis strains of diverse type have IgA protease activity. J. pyelonephritis: a bacteriological study. J. Pathol. Bacteriol. Med. Microbiol. 24:175-180. 96:45-56. 39. Senior, B. W., and C. Hughes. 1987. Production and properties 23. McMichael, J. C., and J. T. Ou. 1979. Structure of common pili from clinical isolates of the Proteeae. J. Med. from Escherichia coli. J. Bacteriol. 138:969-975. of haemolysins 24. Mobley, H. L. T., and G. R. Chippendale. 1990. Hemagglutinin, Microbiol. 24:17-25. urease, and hemolysin production by Proteus mirabilis from 40. Setia, U., I. Serventi, and P. Lorenz. 1984. Bacteremia in a clinical sources. J. Infect. Dis. 161:525-630. long-term care facility: spectrum and mortality. Arch. Intern. 25. Mobley, H. L. T., G. R. Chippendale, K. G. Swihart, and R. A. Med. 144:1633-1635. Welch. 1991. Cytotoxicity of HpmA hemolysin and urease of 41. Silverblatt, F. 1974. Host-parasite interaction in the renal pelvis. Proteus mirabilis and Proteus vulgaris for cultured human renal A possible role for pili in the pathogenesis of pyelonephritis. J. proximal tubular epithelial cells. Infect. Immun. 59:2026-2042. Exp. Med. 140:1696-1711. 26. Mobley, H. L. T., and R. P. Hausinger. 1989. Microbial ureases: 42. Silverblatt, F. J., and I. Ofek. 1978. Influence of pili on the significance, regulation, and molecular characterization. Micro- virulence of Proteus mirabilis in experimental hematogenous biol. Rev. 53:85-108. pyelonephritis. J. Infect Dis. 138:664-667. 27. Mobley, H. L. T., and J. W. Warren. 1987. Urease-positive 43. Tacket, C. O., D. R. Maneval, and M. M. Levine. 1987. bacteriuria and obstruction of long-term urinary catheters. J. Purification, morphology, and genetics of a new fimbrial puta- Clin. Microbiol. 25:2216-2217. tive colonization factor of enterotoxigenic Escherichia coli 28. Montie, T. C., D. Doyle-Huntzinger, R. C. Craven, and I. A. 0159:H4. Infect. Immun. 55:1063-1069. Holder. 1982. Loss of virulence associated with absence of 44. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic flagellum in an isogenic mutant of Pseudomonas aeruginosa in transfer of proteins from polyacrylamide gels to nitrocellulose the burned-mouse model. Infect. Immun. 38:1296-1298. sheets: procedure and some applications. Proc. Natl. Acad. Sci. 29. Moos, M., N. Y. Nguyen, and T. Y. Liu. 1988. Reproducible high USA 76:4350-4354. yield sequencing of proteins electrophoretically separated and 45. Voller, A., D. E. Bidwell, G. Huldt, and E. Engvall. 1974. A transferred to an inert support. J. Biol. Chem. 263:6005-6008. microplate method of ELISA and its application to malaria. 30. Old, D., and R. Adegbola. 1982. Hemagglutinins and fimbriae of Bull. W.H.O. 51:209-221. Morganella, Proteus, and Providencia. J. Med. Microbiol. 46. Warren, J. W., J. H. Tenney, J. M. Hoopes, H. L. Muncie, and 15:551-564. W. C. Anthony. 1982. A prospective microbiologic study of 31. Old, D. C., and R. A. Adegbola. 1985. Antigenic relationships bacteriuria in patients with chronic indwelling urethral cathe- among type-3 fimbriae of Enterobacteriaceae revealed by im- ters. J. Infect. Dis. 146:719-723. munoelectron microscopy. J. Med. Microbiol. 20:113-121. 47. Wei, L., and T. M. Joys. 1985. Covalent structure of three 32. Pazin, G. J., and A. I. Braude. 1969. Immobilizing antibodies in phase-1 flagellar filament proteins of Salmonella. J. Mol. Biol. pyelonephritis. J. Immunol. 102:454-465. 186:791-803. 33. Peerbooms, P., A. VerweiJ, and D. MacLaren. 1983. Investiga- 48. Westphal, O., and K. Jann. 1965. Bacterial lipopolysaccharides. tion of the haemolytic activity of Proteus mirabilis strains. Extraction with phenol-water and further applications of the Antonie Leeuwenhoek 49:1-11. procedures. Methods Carbohydr. Chem. 5:83-91. 34. Peerbooms, P. G. H., A. M. J. J. Verwei, and D. M. MacLaren. 49. Wray, S., S. Hull, R. Cook, J. Barrish, and R. Hull. 1986. 1984. Vero cell invasiveness of Proteus mirabilis. Infect. Im- Identification and characterization of a urinary cell adhesin from mun. 43:1068-1071. a uropathogenic isolate of Proteus mirabilis. Infect. Immun. 35. Pere, A., R. K. Selander, and T. K. Korhonen. 1988. Character- 54:43-49.