INFECTION AND IMMUNITY, Oct. 1986, p. 43-49 Vol. 54, No.

1
0019-9567/86/100043-07$02 .00/0
Copyright 1986, American Society for Microbiology

Identification and Characterization of a Uroepithelial Cell Adhesin

from a Uropathogenic Isolate of Proteus mirabilis
SUSAN K. WRAY,l* SHEILA I. HULL,' RICHARD G. COOK,'2 JAMES BARRISH,3 AND RICHARD A. HULL'
Department of Microbiology and Imunttiologyl anid Howard Huighes Medical Institutte,2 Baylor College of Medicine, and
Department of Pathology, Texas Children's Hospitall,3 Houtstoni, Texas 77030
Received 27 January 1986/Accepted 17 June 1986

Proteus mirabilis is a frequent cause of urinary tract infections in rehabilitation hospitals and among persons
with structural abnormalities of the urinary tract. Adherence to uroepithelial tissues may be an important
virulence determinant in these infections because most Proteus strains adhere to desquamated uroepithelial
cells. To identify the adherence factor responsible for this phenomenon, we sheared outer membrane material

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

from 35S04-radiolabeled bacteria and allowed it to bind to uroepithelial cells. Following sodium dodecyl
sulfate-polyacrylamide gel electrophoresis, the major adherence element was a protein with an apparent
molecular weight of 17,500 and was provisionally designated as the uroepithelial cell adhesin. This adhesin was
purified by heat shock and gel filtration on Sepharose CL-4B. After purification, the adhesin was seen
assembled as long, flexible rods by electron microscopy. The N-terminal amino acid sequence of the subunit had
limited homology with that of the K99 fimbriae of Escherichia coli.

Urinary tract infections are the most commonly acquired MATERIALS AND METHODS
nosocomial infections (26). Proteuts mirabilis is a frequent
cause of these infections, especially in rehabilitation hospi- Strains, media, buffers, and reagents. The uropathogenic
tals and in persons with structural abnormalities of the strains of P. mirabilis were obtained from the Ben Taub
urinary tract (5, 20, 27). These bacteria generally produce a County Hospital, Houston, Tex.
more severe disease than does Escherichia coli, owing to the Minimal medium or L medium was used for the cultivation
production of urease, which in turn leads to struvite stone of bacteria. L medium consisted of 10 g of tryptone, 5 g of
formation (10). Several bacterial determinants are associated yeast extract, and 5 g of NaCl per liter. Minimal medium
with uropathogenic strains of P. mirabilis, including hemo- consisted of 60 mM potassium phosphate (pH 7.4); 2 mM
lytic activity, invasiveness in tissue cultures, motility, and disodium citrate; 0.8 mM MgCl,; 15 mM NH4Cl; 0.08 mM
certain proticine types (22-24, 27). (NH4)2SO4; 0.02 mg each of glycine, serine, arginine, thre-
The capacity to adhere to uroepithelium is believed to be onine, histidine, isoleucine, valine, lysine, leucine, and
another bacterial determinant related to virulence in Proteuls tryptophan per ml; 0.4% glycerol; and 0.5 mg of yeast
strains. In other genera studied to date, this adherence is extract per ml.
mediated by structures present on the outer surfaces of the Buffered saline with gelatin (BSG) contained 0.15 M NaCl,
cells (glycolayxes, capsules, and fimbriae) (6). For example, 0.004 M potassium phosphate (pH 6.0), and 0.1% gelatin.
uropathogenic E. coli uses P fimbriae to attach to the P blood Carrier-free H2350S4 was obtained from ICN Radiochem-
group antigen on human tubule cells of the kidney (12). A icals, Irvine, Calif.
similar phenomenon probably occurs in Proteuts strains, as Bacterial cell adherence assay. Uroepithelial cells were
demonstrated by electron micrographs which show organ- isolated from the urine of three healthy female volunteers by
isms attached to rat renal epithelium by fimbriae of the type centrifugation at 400 x g for 10 min. The cells were washed
IV category (7 nm in diameter) (28). In human studies, in BSG, and pellets containing 2 x 105 cells were suspended
Proteus strains have been shown to adhere to desquamated in 1 ml of BSG containing 5 x 10' bacteria. Suspensions
uroepithelium; however, attempts to correlate virulence were rotated for 1 h at 37C. The cells were washed four
with adherence were unsuccessful, owing to the high fre- times in BSG and examined under a light microscope. The
quency of adherent strains among stool isolates used as number of bacteria adhering per cell was counted.
controls (29). Certainly, the contribution of adherence fac- Hemagglutination. Bacteria were grown in L broth, in
tors to the pathogenicity of Proteuts isolates is not well minimal broth, or on L agar for 16 h, washed, and suspended
understood. in BSG to an A600 of 1. Then, 0.2 ml of bacterial suspension
In this study, we identified a protein from a uropathogenic was mixed with a equal volume of a 2% suspension of fresh
isolate of P. mirabilis which adheres to desquamated human or tanned erythrocytes in BSG. The mixtures were rocked
uroepithelial cells. This protein was purified and found to be for 5 min at room temperature and examined for agglutina-
organized as flexible rods (fimbriae) by electron microscopy. tion.
In addition, these fimbriae were characterized as to N- Preparation of radiolabeled outer membrane material. Bac-
terminal amino acid sequence and amino acid composition. teria (1010) were harvested after 16 h of growth at 37C from
We suggest that these fimbriae are the adherence element 10 ml of minimal broth containing 50 pLCi of 35SO4 per ml and
responsible for binding uropathogenic Prcoteuis strains to washed with BSG. Bacteria were suspended in 1.0 ml of
uroepithelium. BSG, and outer membrane material was sheared from the
bacterial surface by passage through a 26-gauge needle.
Whole bacteria were removed by centrifugation for 10 min at
*
Corresponding author. 10,000 x g.
43
44 WRAY ET AL. INFECT. IMMUN.

TABLE 1. Adherence of P. mirabilis to desquamated isopropanol-40% water. Phenylthiohydantoin amino acids

uroepithelial cellsa were eluted at a flow rate of 1 ml/min by using the following
Mean no. of bacteria/uroepithelial cell gradient: 0 to 0.5 min, 0% solvent B; 0.5 to 3.5 min, convex
Strain gradient to 35% solvent B; 3.5 to 12 min, isocratic elution at
L agar L broth Minimal broth 35% solvent B. The amino-terminal sequence was deter-
HU1069 15.5 1.6b 25.5 5 29.2 4.5b mined three times.
HU1070 5.6c NDd ND Amino acid analysis. The amino acid composition was
HU1071 6.5c ND ND determined by using the Waters Pico-Tag system after acid
a A total of 64 uroepithelial cells were examined by light microscopy, and hydrolysis of SDS-PAGE-purified P. mirabilis UCA (3).
the number of adherent bacteria was determined.
b Four experiments.
Electron microscopy. Bacteria (P. mirabilis HU1069)
c Two experiments.
grown overnight in minimal broth, in L broth, or on L agar
d
ND, Not determined. were washed and suspended in BSG. One drop of the
bacterial suspension or one drop of a 1:100 dilution of P.
mirabilis UCA in phosphate-buffered saline was placed on a
Adsorption of bacterial outer membrane material by cube of 2% agarose and stained with 0.5% uranyl acetate by
uroepithelial cells. Uroepithelial cells were isolated as de- a pseudoreplicating technique (19). Specimens were exam-
scribed above. Pellets containing 2 x 105 cells were sus- ined with a JEOL 100CX electron microscope. Bacteria

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

pended in BSG containing radiolabeled outer membrane were also fixed in 3% glutaraldehyde prior to being stained.
material. Reaction volumes were 1.0 ml. After incubation for Fixed and unfixed bacteria appeared identical when
1 h at 37C, the reaction mix was centrifuged, and both the examined.
cell pellet and the supernatant were retained. The cells were
then washed extensively with BSG, and a sample was RESULTS
counted in a Beckman scintillation counter. The number of
counts in the entire sample was calculated, and an equal Adherence of strain HU1069 to uroepithelium. Three
number of counts was removed from the supernatant frac- uropathogenic isolates of P. mirabilis were grown on L agar
tion for further processing. The cells with adhered bacterial for 16 h and tested for their capacity to bind desquamated
components and the counts from the supernatant fraction human uroepithelial cells as described in Materials and
containing unattached bacterial components were dialyzed Methods. The strains tested bound uroepithelium to various
against distilled H20 and lyophilized. Samples were electro- degrees (Table 1). Strain HU1069 attached to the uroepithe-
phoresed by sodium dodecyl sulfate-polyacrylamide gel lial cells in the greatest numbers (15 bacteria per cell) and
electrophoresis (SDS-PAGE) by the procedure of Laemmli was chosen for subsequent experiments.
(18). Gels were dried and autoradiographed. Strain HU1069 was then tested for its capacity to adhere
Purification of UCA. Bacteria were grown in minimal broth to uroepithelium after growth in various types of media.
and harvested after 18 h. Cells were washed in 0.05 M Tris Minimal broth cultures yielded bacteria which adhered bet-
(pH 8.0). Bapteria were suspended in 50 mM sodium phos- ter than bacteria grown on L agar or in L broth.
phate (pH 7.2) with 2 M urea and incubated at 65C for 20 Bacteria grown under these three conditions were also
min. Whole cells were removed by centrifugation at 7,500 x tested for their capacity to agglutinate fresh erythrocytes
g for 10 min. An equal volume of saturated ammonium from humans, guinea pigs, and sheep and tanned human
sulfate was added to the supernatant, and the solution was erythrocytes. No agglutination was detected.
stirred for 2 h at 4C. The precipitate was collected by Identification of strain HU1069 UCA. Because adherence is
centrifugation at 12,000 x g for 20 min. The pellet was commonly mediated by proteinaceous outer membrane ap-
suspended in 0.05 M Tris (pH 7.8) and dialyzed against the pendages among members of the family Enterobacteriaceae,
same buffer for 16 h. Insoluble material was removed by a similar phenomenon was expected to occur in P. mirabilis.
centrifugation at 25,000 x g for 20 min. Sodium deoxycho- To identify this uroepithelial cell adhesin (UCA), we labeled
late was added to a concentration of 0.5%, and the sample bacterial cells by growth in minimal broth (conditions opti-
was dialyzed against this buffer for 48 h. The deoxycholate- mal for adherence) containing 35 o4 and removed their outer
insoluble material was removed by centrifugation at 12,000 membrane appendages by passing the suspension through a
x g for 20 min, and the supernatant was dialyzed against 26-gauge needle. The whole cells were removed by centrif-
0.05 M Tris (pH 7.8) for 72 h. The protein was then dialyzed ugation, and the supernatant was incubated with desqua-
against water and lyophilized. Samples were suspended in 50 mated human uroepithelial cells in BSG. Following incuba-
mM sodium phosphate (pH 7.2) with 6 M urea and chro- tion, the cells and attached bacterial components were
matographed in the same buffer over a Sepharose CL-4B subjected to SDS-PAGE followed by autoradiography (Fig.
column. The flow rate was 5 ml/h. The void volume con- 1). The radiolabeled bacterial components which were re-
tained the uroepithelial cell adhesin (UCA). The yield from 3 moved by shearing of the bacteria are shown in Fig. 1, lane
liters of mninimal broth was approximately 0.5 mg of UCA. A. The profile is similar to that reported by Driver and
N-terminal amino acid sequencing. Approximately 20 p.g of Lambert (9). The mixture was heterogenous; however, the
protein was further purified by SDS-PAGE and electroeluted label was concentrated in two proteins of Mr 17,500 and Mr
as described previously (11). The protein was lyophilized 39,000. The radiolabeled bacterial components which re-
and sequenced by using the Applied Biosystems model 470A mained in the supernatant following adsorption by uroepi-
protein sequencer and the 02RPTH program with trifluoro- thelial cells were electrophoresed in Fig. 1, lane B. The
acetic acid conversion to phenylthiohydantoin derivatives profile is identical to that in Fig. 1, lane A. Figure 1, lane C,
(7). Phenylthiohydantoin amino acids were identified by contains the bacterial components adsorbed by the
high-performance liquid chromatography on a Waters Nova uroepithelial cells. The major protein which adhered to
Pac C18 column by using a Waters high-performance liquid uroepithelium (i.e., the UCA) was the Mr-17,500 protein.
chromatography system. Solvent A was 84% sodium acetate The other radiolabeled proteins that were represented in the
(33 mM, pH 5.0)-16% acetonitrile; solvent B was 60% supernatant, including the Mr-39,000 protein, were also
VOL. 54, 1986 ADHESIN OF PROTEUS MIRABILIS 45

present among the attached bacterial components, although A B C D

in a reduced quantity, suggesting they have a low affinity for
uroepithelium.
Purification of the Mr-17,500 protein from strain HU1069.
The Mr-17,500 major adherence protein (UCA) was believed
to mediate the attachment of bacteria to uroepithelial cells 66 u>
because of its capacity to bind uroepithelium in vitro. A
protein having an identical migration in SDS-PAGE was
purified from strain HU1069 by using fimbria isolation tech- 45 -*
niques (17). Bacteria were grown in minimal broth, washed, _
suspended in 50 mM P04 with 2 M urea, and incubated at 36 _, _
65C for 20 min. The whole bacteria were removed by
centrifugation. This procedure yielded a complex mixture of 29 __ _
proteins (Fig. 2, lane C). This mixture was concentrated by
ammonium sulfate precipitation and dialyzed against 50 mM 24 *
Tris (pH 7.8). Insoluble material was removed by centrifu-
gation, the supernatant was treated with 0.5% deoxycholate,

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

and insoluble material was again removed. The deoxycho- 20 s
late-soluble material was chromatographed over a Sepharose
CL-4B column. The void volume from this column contained
the Mr-17,500 protein and a contaminant Of Mr 37,000 (Fig. 2,
lane D). An alternative isolation procedure, in which strain
HU1069 was grown on complex medium (L agar) and outer 14 vi .
ms
membranes appendages were blended off in a Waring
blender, yielded principally the Mr-39,000 protein (Fig. 2,

FIG. 2. Purification of the Mr-17,500 protein from P. mirabilis

A B C HU1069. Lanes: A, standard marker proteins; B, sheared outer
membrane material from HU1069 grown on complex medium; C,
material derived from heat shock of HU1069 grown in minimal
medium; D, purified Mr-17,500 protein. Numbers on the left are the
apparent molecular weights (in thousands) of the marker proteins;
the arrow indicates an Mr of 17,500.

lane B). Significant amounts of the Mr-17,500 protein could

not be isolated by this method.
43--. Binding of purified UCA to uroepithelial cells. To determine
if the purified Mr-17,500 protein possessed the same func-
tional characteristics as the UCA (as identified by adsorption
28-n studies), we purified the Mr-17,500 protein from bacterial
cells grown in minimal broth containing 30 ,uCi of 15 O4 per
ml by using the technique described above (i.e., Sepharose
CL-4B column chromatography). The radiolabeled
Mr-17,500 protein (82,154 cpm) was incubated with 105
uroepithelial cells. After washing, 7,335 cpm remained at-
18.4-_ tached to the cells (approximately 9% binding). For compar-
ison, an equal number of counts contained in bacterial
proteins which were retained on the Sepharose CL-4B
14.3-' column (nonfimbrial proteins) were also incubated with
uroepithelial cells. These proteins attached less well to
uroepithelial cells (1.5% binding). To determine if the in-
creased binding (9%) by fimbrial proteins was due to the
adherence of the Mr-17,500 protein, we subjected the cells in
this assay to SDS-PAGE (Fig. 3). The majority of the counts
FIG. 1. Adherence of P. mirabilis outer membrane material to were due to the attachment of the Mr-17,500 protein to
human uroepithelial cells. Radiolabeled outer membrane material uroepithelium (Fig. 3, lane B). Thus, the purified Mr-17,500
was sheared from strain HU1069 after growth in minimal medium protein possessed the adhesive capacity characteristic of the
containing 50 ,uCi of 35SO4 per ml and incubated with 2 x 105 I JCA in addition to having an identical migration in SDS-
desquamated uroepithelial cells in BSG. Uroepithelial cells and PAGE. Hence, we assumed that the purified Mr-17,500
adsorbed material were removed by centrifugation and washed.
Samples were lyophilized and electrophoresed on SDS gels. Gels protein was the UCA.
were dried and autoradiographed. Lanes: A, sheared material from Organization of UCA as fimbriae. The column-purified
HU1069; B, outer membrane material remaining in the supernatant UCA was stained with uranyl acetate by a pseudoreplicating
after adsorption to uroepithelial cells; C, outer membrane material technique and viewed with a JEOL 100CX electron micro-
adsorbed to uroepithelial cells. The positions of molecular weight scope. The UCA was seen organized as long, flexible rods
markers (in thousands) are indicated on the left. (Fig. 4A). The rods in the preparation varied in diameter. A
46 WRAY ET AL. INFECT. IMMUN.

A B the most frequent amino acids. Arginine, cysteine, and

methionine were detected only in small amounts (one resi-
due per molecule). The concentration of tryptophan was not
determined.

DISCUSSION
These studies identified a protein from a uropathogenic
strain of P. mirabilis which may be responsible for the
attachment of the bacteria to human uroepithelium. This
main characteristic of the UCA was its capacity to adhere to
uroepithelial cells. Furthermore, the UCA was seen orga-
nized as rod-shaped structures or fimbriae following purifi-
cation. Similar structures were seen radiating from intact
organisms. The expression of the fimbriae on the organisms
was greatest under conditions allowing maximal adherence
of the bacteria to uroepithelium. These data suggest that the

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

UCA is organized on the bacterial cell surface as fimbriae
which mediate the attachment of the bacteria to uroepithe-
lium.
The UCA was purified to homogeneity by a combination
m of deoxycholate treatment, gel filtration, and SDS-PAGE
from bacteria grown in minimal broth cultures. Following
purification, the UCA was physically characterized as to
amino acid composition, subunit size, and N-terminal amino
acid sequence. The UCA has an amino acid composition
similar to that reported for the fimbriae of E. coli, i.e.,
FIG. 3. Binding of the purified Mr-17,500 protein to uroepithelial threonine was unusually frequent and arginine was rare (for
cells. Lanes: A, radiolabeled proteins contained in the void volume
after Sepharose CL-4B chromatography; B, radiolabeled proteins a review, see reference 16). The relative mobility of the UCA
from the column void volume which attached to uroepithelial cells. in SDS-PAGE was 17,500, which is also similar to that
reported for other fimbrial subunits. Those studied to date
range in molecular weight from 15,000 (colonization factor
antigen 1) to approximately 29,500 (F41) (8, 14). N-terminal
thin filament of 4 nm predominated, although a thicker amino acid sequencing of the UCA fimbriae revealed limited
filament of 6 nm was also present (approximately 10% of the homology to the K99 pilus of E. coli. The significance of this
filaments). Similar structures were seen radiating from intact homology is uncertain. Certainly these two adhesins bind to
organisms. These fimbrial structures were expressed most very different tissues: K99 to the intestional epithelium of
frequently on cells grown in minimal broth (Fig. 4B), al- calves and sheep erythrocytes and UCA to desquamated
though the majority of the cells were heavily flagellated and epithelium from human urinary tracts. It is interesting that
not fimbriated. Organisms grown in L broth expressed there is much less homology between UCA and the urinary
fimbriae less frequently than those grown in minimal broth, tract adhesins of E. coli (P. fimbriae, 20% homology; type 1
and L agar-grown organisms rarely expressed fimbriae (data fimbriae, 25% homology) in the first 20 residues. This may
not shown). indicate that UCA is, in fact, primarily an adhesin for
Characterization of purified UCA. The UCA was purified intestional epithelium and that patients acquire P. mirabilis
to homogeneity by electroelution from a Coomassie blue- urinary tract infections from natural bowel flora. In support
stained 12.5% SDS-polyacrylamide gel. The protein was of this theory is the observation of Svanborg-Eden et al.,
lyophilized, ethanol precipitated, and subjected to N- who showed that Proteus isolates from normal stools have
terminal amino acid sequence analysis (Fig. 5). The P. the same adhesive capacity for exfoliated human uroepithe-
mirabilis UCA has significant homology to the K99 pilus lium as do uropathogenic Proteus isolates (29). Prospective
associated with diarrhea-causing E. coli of calves. Among studies are in progress to investigate whether intestinal
the first 15 residues, there is 46% homology (7 residues are Proteus strains from individual patients are responsible for
identical). Among the eight remaining positions, one has a subsequent urinary tract infections in the same patients.
very conservative isoleucine-to-valine change, four other Previous ultrastructural studies of Proteus strains have
changes can also be attributed to 1-base-pair differences in identified two types of fimbria, a thick filament approxi-
the DNA, and only three positions represent more radical mately 7 nm in diameter and a thin filament 4 nm in diameter
changes in the DNA (aspartic acid to threonine, valine to (4, 28). These are, in addition to the numerous flagella (12
threonine, and glutamine to alanine). Homology to other nm), present on the surface. Silverblatt has reported that the
fimbriae of E. coli is very limited, ranging from 26% for thicker pili appear to be related to infection, i.e., cells
colonization factor antigen to 13% for P fimbriae in the first expressing 7-nm-thick fimbriae were more virulent in a rat
15 residues. model of pyelonephritis (28). Other workers have attempted
Amino acid composition analysis was performed on SDS- to associate patterns of hemagglutination with expression of
PAGE-purified UCA (Table 2). The moles percent were these fimbriae (1, 21). Old and Adegbola (21) reported that
calculated from the peak area following high-performance the thin filaments were responsible for Klebsiella-like hem-
liquid chromatography and, based on a molecular weight of agglutination (agglutination of tanned ox erythrocytes) and
17,600, the number of residues of each amino acid was that the thicker filaments may be associated with Proteus-
calculated. Threonine and asparagine or aspartic acid were like hemagglutination (agglutination of fresh fowl, guinea
VOL. 54, 1986 ADHESIN OF PROTEUS MIRABILIS 47

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

FIG. 4. (A) Electron micrograph of purified UCA. (B) Electron micrograph of P. mir' bilis HU1069 grown for 16 h in minimal broth.
48 WRAY ET AL. INFECT. IMMUN.

5 10 15 2k ? 2r re f.

UCA Tyr Asp Gly Thr Thr Phe Thr Gly Lys ViJ Val Ala Gln h X Ser AnLThr Ast tAs s A Leu
i , h s 5a rt i 1e

K99 Asn Thr Gly Thr Asn Phe Asn Gly Lys Ile Thr Ser Ala Thr|Cys Thr I'e 'li pro -, Va' Asn st %i 'sn ?
L_..
CFAI Val Glu Lys Asn Ile IThrI Val IThrI Al a Se r ivalI Asp Pro Val Ile Asp Leu Le. G'In Ald Asp r,y Asn A a Le !4

K88 Trp Met1Th .lGly AspfPhe Asn ISe r Val IAsp Ile Gly Gly Ser Ile Thr Ala Asp Asp Tyr Arg Gin Lys

type I Ala AlaIThrjThr Val Asn GlyI6lylThrIVal His Phe Lys Gly Gluj Val EValj nAla Ala fys Al a VVal Asp Ir

P type Ala Pro Thr Ile Pro GIn G7y Gn IGlyILysIVal Thr Phe Asn Gly Thr Val |Val Asp Ala Pro Cys Ser lit Ser 2
t f ~~J
FIG. 5. N-terminal amino acid sequence of UCA from P. mircabilis. The N-terminal amino acid sequence of UCA is compared with those

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

of various fimbriae isolated from E. coli. Residues that are common between UCA and other fimbriae are shaded. Numbering corresponds
to the UCA sequence. CFAl, Colonization factor antigen 1.

pig, horse, human, and sheep erythrocytes). Unfortunately, possibility seems unlikely because flagella were not seen in
these experiments were hampered by the inability to isolate the fimbria preparations.
strains expressing only one type of fimbria. In the present P. mirabilis HU1069 was seen to be heavily flagellated.
study, these two types of fimbria were both excluded on the Even after growth in minimal medium, only a minority of the
Sepharose CL-4B column and could be observed in the cells expressed fimbriae. Despite this relatively low level of
electron micrographs. expression, the UCA fimbriae could be isolated by heat
Complete purification of the UCA required SDS-PAGE shock followed by gel filtration. The fimbrial subunit was not
because a contaminant of Mr-37,000 copurified with the detected by SDS-PAGE on cells grown on complex medium,
UCA during gel filtration. This larger protein was not a dimer due to reduced expression under these conditions. This
of the UCA subunit, as its N-terminal sequence was different relatively low level of expression of fimbriae in this strain
from that of the UCA (data not shown). The second protein may explain the inability to detect any agglutination of
may be the subunit of the second type of fimbria seen in the sheep, guinea pig, or human erythrocytes. Alternatively,
micrographs, although the molecular weight of the protein Adegbola et al. (1) reported that a small percentage of
appears to be large for a fimbrial subunit. Alternatively, this Proteus isolates agglutinate only fowl erythrocytes, which
contaminant may be a dimer of flagellin. Driver and Lambert were not tested here; strain HU1069 may be similar to these
(9) reported that the dimer of flagellin migrates in SDS- strains. We believe that this is a more plausible explanation
PAGE with an apparent molecular weight of 40,000, which is because strain HU1069 adhered well to uroepithelium, sug-
similar to the molecular weight of the contaminant. This gesting that sufficient fimbriae were present for interaction
with mammalian tissues if the appropriate receptors were
present.
In summary, we have detected a protein from a uropath-
TABLE 2. Amino acid composition of P. mirabilis UCA' ogenic isolate of P. mirabilis which is related to the K99
fimbriae of enterotoxigenic E. (coli. The functional and
Amino Mol No. of physical characteristics of this protein suggest that it may be
acid Y residues
responsible for the attachment of P. mirabilis to human
Thr 0.158 25 uroepithelium during urinary tract infections.
Asx 0.146 23
Ala 0.112 18 ACKNOWLEDGMENTS
Glx 0.086 14
Val 0.086 14 This work was supported by a research fellowship to S.K.W. from
Gly 0.074 12 the National Kidney Foundation and by Public Health Service
Leu 0.067 11 research grants Al 18462, Al 21009. and Al 17897 from the National
Ser 0.057 9 Institutes of Health.
Lys 0.036 6
Ile 0.036 6 LITERATURE CITED
Pro 0.039 6 1. Adegbola, R. A., D. C. Old, and B. W. Senior. 1983. The
Tyr 0.034 6 adhesins and fimbriae of Proteus mirabilis strains associated
Phe 0.025 4 with high and low affinity for the urinary tract. J. Med. Micro-
His 0.020 3 biol. 16:427-431.
Met 0.007 1 2. Baga, M., S. Normark, J. Hardy, P. O'Hanley, D. Lark, 0.
Arg 0.005 1 Olsson, G. Schoolnik, and S. Falkow. 1984. Nucleotide sequence
Cys 0.005 1 of the papA gene encoding the Pap pilus subunit of human
a SDS-PAGE-purified UCA was hydrolyzed and analyzed by using the uropathogenic Escherichia coli. J. Bacteriol. 157:330-333.
Waters Pico-Tag system. The number of residues was calculated by assuming 3. Bidlingmeyer, B. A., S. A. Cohen, and T. L. Tarvin. 1984. Rapid
a molecular weight of 17,600 for the UCA and an average molecular weight of analysis of amino acids using pre-column derivatization. J.
110 per amino acid. The number of residues was rounded to the nearest Chromatogr. 336:93-104.
integer. 4. Brinton, C. C. 1965. The structure, function, synthesis and
VOL. 54, 1986 ADHESIN OF PROTEUS MIRABILIS 49

genetic control of bacterial pili and a molecular model for DNA Infect. Dis. 7:321-340.
and RNA transport in gram negative bacteria. Trans. N.Y. 17. Korhonen, T. K., E. Nurmiaho, H. Ranta, and C. Svanborg-
Acad. Sci. 27:1003-1054. Eden. 1980. New method for isolation of immunologically pure
5. Brisset, J. M., M. Charton, and J. F. Girard. 1983. Infection and pili from Escherichia coli. Infect. Immun. 27:569-575.
upper urinary tract lithiasis, p. 101-113. In B. Francois and P. 18. Laemmli, U. K. 1970. Cleavage of structural proteins during the
Perrin (ed.), Urinary infections. Butterworth & Co. (Publishers) assembly of the head of bacteriophage T4. Nature (London)
Ltd., London. 227:680-685.
6. Bruce, A. W., R. C. Y. Chan, D. Pinkerton, A. Morales, and P. 19. McCombs, W. B., III, C. E. McCoy, and 0. D. Holton III. 1980.
Chadwick. 1983. Adherence of gram-negative uropathogens to Electron microscopy for rapid viral diagnosis. Tex. Soc. Elec-
human uroepithelial cells. J. Urol. 130:293-298. tron Microsc. J. 11(2):9-15.
7. Craigen, W. J., R. G. Cook, W. P. Tate, and C. T. Caskey. 1985. 20. Nitzan, Y., M. Maayan, C. Wajsman, and M. Drucker. 1983.
Bacterial peptide chain release factors: conserved primary Urinary tract infection and drug-resistant bacteria in different
structure and possible frameshift regulation of release factor 2. patient populations. Isr. J. Med. Sci. 19:1039-1045.
Proc. Natl. Acad. Sci. USA 82:3616-3620. 21. Old, D. C., and R. A. Adegbola. 1982. Haemagglutinins and
8. de Graaf, F. K., and I. Roorda. 1982. Production, purification, fimbriae of Morganella, Proteus and Providencia. J. Med.
and characterization of the fimbrial adhesive antigen F41 iso- Microbiol. 15:551-564.
lated from calf enteropathogenic Escherichia coli strain B41M. 22. Pazin, G. J., and A. I. Braude. 1974. Immobilizing antibodies in
Infect. Immun. 36:751-758. urine. II. Prevention of ascending spread of Proteus mirabilis.
9. Driver, K., and P. A. Lambert. 1984. Surface antigens of Invest. Urol. 12:129-134.

Downloaded from http://iai.asm.org/ on June 10, 2017 by guest

Proteus mirabilis revealed by electroblotting from sodium do- 23. Peerbooms, P. G. H., A. M. J. J. Verweij, and D. M. MacLaren.
decyl sulphate-polyacrylamide gels. Microbios. 41:87-98. 1983. Investigation of the haemolytic activity of Proteus
10. Griffith, D. P., D. M. Musher, and C. Itin. 1976. Urease: the mirabilis strains. Antonie von Leeuwenhoek J. Microbiol.
primary cause of infection-induced urinary stones. Invest. Urol. 49:1-11.
13:346-350. 24. Peerbooms, P. G. H., A. M. J. J. Verweij, and D. M. MacLaren.
11. Hunkapiller, M. W., E. Lujan, F. Ostrander, and L. E. Hood. 1984. Vero cell invasiveness of Proteus mirabilis. Infect. Im-
1983. Isolation of microgram quantities of proteins from poly- mun. 43:1068-1071.
acrylamide gels for amino acid sequence analysis. Methods 25. Roosendaal, B., W. Gaastra, and F. K. de Graaf. 1984. The
Enzymol. 91:227-236. nucleotide sequence of the gene encoding the K99 subunit of
12. Kallenius, G., R. Mollby, S. B. Svenson, J. Winberg, A. enterotoxigenic Escherichia coli. FEMS Micribiol. Lett.
Lundblad, S. Svensson, and B. Cedergren. 1980. The pk antigen 22:253-258.
as receptor for the hemagglutination of pyelonephritic Esche- 26. Rutledge, K. A., and H. P. McDonald, Jr. 1985. Costs of treating
richia coli. FEMS Microbiol. Lett. 7:297-302. simple nosocomial urinary tract infection. Urology 26(Suppl.):
13. Klemm, P. 1981. The complete amino acid sequence of the K88 24-26.
antigen, a fimbrial protein from Escherichia coli. Eur. J. 27. Senior, B. W. 1978. The special affinity of particular types of
Biochem. 117:617-627. Proteus mirabilis for the urinary tract. J. Med. Microbiol.
14. Klemm, P. 1982. Primary structure of the CFA/I fimbrial protein 12:1-8.
from human enterotoxigenic Escherichia coli strains. Eur. J. 28. Silverblatt, F. 1974. Host-parasite interaction in the rat renal
Biochem. 124:339-348. pelvis. J. Exp. Med. 140:1696-1711.
15. Klemm, P. 1984. The fimA gene encoding the type-1 fimbrial 29. Svanborg-Eden, C., P. Larsson, and H. Lomberg. 1980. Attach-
subunit of Escherichia coli: nucleotide sequence and primary ment of Proteus mirabilis to human urinary sediment epithelial
structure of the protein. Eur. J. Biochem. 143:395-396. cells in vitro is different from that of Escherichia coli. Infect.
16. Klemm, P. 1985. Fimbrial adhesins of Escherichia coli. Rev. Immun. 27:804-807.