INFECTION AND IMMUNITY, June 1993, P. 2570-25770019-9567/93/062570-08$02.00/0Copyright X) 1993, American Society for Microbiology

Proteus mirabilis Urease: Histidine 320 of UreC Is Essentialfor Urea Hydrolysis and Nickel Ion Binding within

the Native EnzymeBUSARAWAN SRIWANTHANA AND HARRY L. T. MOBLEY*

Division of Infectious Diseases, Department ofMedicine, University of MarylandSchool ofMedicine, 10 South Pine Street, Baltimore, Maryland 21201

Received 2 October 1992/Accepted 23 March 1993

Proteus mirabilis urease, a nickel-containing enzyme, has been established as a critical virulence determinantin urinary tract infection. An amino acid sequence (residues 308 to 327: TVDEHLDMLMVCHHLDPSIP)within the large urease subunit, UreC, is highly conserved for every urease examined thus far and has beensuggested to reside within the enzyme active site. Histidine residues have been postulated to play a role incatalysis by coordinating Ni2+ ions. To test this hypothesis, oligonucleotide-directed mutagenesis was used tochange amino acid His-320 to Leu-320 within UreC. The base change (CAT for His-320 to CIT for Leu-320)was confirmed by DNA sequencing. The recombinant and mutant proteins were expressed at similar levels inEscherichia coli as detected by Western blotting (immunoblotting) of denaturing and nondenaturing gels.Specific activities of the enzymes were quantitated after partial purification. Strains expressing the mutantenzyme showed no detectable activity, whereas strains expressing the recombinant enzyme hydrolyzed urea at149 ,umol of NH3 per min per mg of protein. In addition, the mutant enzyme was able to incorporate only aboutone-half (58%) of the amount of '3Ni2+ incorporated by the active recombinant enzyme. While the mutationof His-320 to Leu-320 within UreC does not affect expression or assembly of urease polypeptide subunits UreA,UreB, and UreC His-320 of UreC is required for urea hydrolysis and proper incorporation of Ni2+ intoapoenzyme.

Proteus mirabilis, a gram-negative enteric bacterium, is a

frequent cause of nosocomial and catheter-associated uri-nary tract infections (25, 32, 43). Urease (urea amidohydro-lase; EC 3.5.1.5) produced by this species catalyzes thehydrolysis of urea to carbonic acid and ammonia (24).Alkaline pH promotes precipitation of calcium or magne-sium ions in urine, which results in the formation of urinaryor renal stones composed of carbonate-apatite [Ca1O(PO4CO3OH)6(OH)] or struvite (MgNH4PO4 6H20) (8).

P. mirabilis urease plays a significant role in infection.Besides initiating the formation of urinary stones, urease hasbeen implicated as a contributing factor in the pathogenesisof pyelonephritis, hepatic coma, hyperammonemia, comple-ment inactivation, and urinary catheter obstruction (24).These complications result from the toxicity of ammonia forepithelium and the subsequent rise in urine pH. Using an

isogenic strain of P. mirabilis HI4320 in which a specificdeletion within the urease structural subunit had been con-

structed, our laboratory reported that the lack of ureaseaccounts for a reduction in infectivity in CBA mice transure-thrally challenged with the mutant as compared with theparent strain (13). Furthermore, in vitro assays showed thaturease appeared to contribute to the cytotoxicity for cul-tured human renal proximal tubular epithelial cells in con-

junction with hemolysin activity of P. mirabilis (23).The urease of P. mirabilis, a nickel-containing protein

estimated to be 212 to 250 kDa, is composed of three subunitpolypeptides, UreA, UreB, and UreC, with predicted mo-

lecular sizes of 11, 12.2, and 61 kDa in the ratio of 2:2:1,respectively (15, 16). The enzyme localizes to the cytoplasmof the cell (15) and is induced solely by urea (28, 29).

* Corresponding author.

The urease from jack bean (Canavalia ensiformis) hasbeen well characterized and the amino acid sequence hasbeen determined by Edman degradation (20, 35). Biochem-ical and biophysical studies of the active site of the jack beanurease suggested that a cysteine residue (Cys-592) serves asa general acid catalyst in the hydrolysis of urea. The histi-dine residues in this region (residues 479 and 607) may bindnickel (35), a function critical for catalysis (7).

In contrast to the well-studied jack bean urease, less isknown about catalytic activities of microbial ureases. Com-parison of the predicted amino acid sequence of P. mirabilisurease subunits with that of jack bean showed very highsimilarity (58% exact match of amino acid sequence), espe-cially between the amino acids that correspond to theputative active site of jack bean and a region of the UreCsubunit of P. mirabilis urease (16). From these data, weinferred the location of the active site of P. mirabilis urease.We postulated that changing a key amino acid residue withinthe predicted active site would drastically affect enzymeactivity. In this report, we describe the effect of changing thecodon for His-320 to that of Leu-320 within ureC whichencodes the large subunit of the P. mirabilis urease.

MATERIALS AND METHODS

Bacterial strains and plasmids. P. mirabilis H14320 (ureasepositive, MR/P and MR/K fimbriated, and hemolysin posi-tive) was isolated from a patient with urinary catheter-associated bacteriuria (14). Escherichia coli HB101 (F-hsdS20 supE44 proA2 leuB6 rpsL20 recA13 lacY] galK2thi-1 ara-14 mcrB xyl-5 mtl-i), E. coli DH5S( [(supE44AlacU169) (680 lacZAM15) hsdRl7 recAl endAl gyrA96thi-1 rel1A], E. coli XL-1 blue (supE44 hsdRl 7 recAl endA1gyrA46 thi relA1 Lac- F'[proAB+ lacIq lacZAM15 TnlO

2570

Vol. 61, No. 6

P. MIRABILIS UREASE: ROLE OF His-320 IN UreC 2571

(Tetr)], E. coli SDM [hsdRl7 mcrAB recAl supE44 TetrA(Lac-proAB) (F' traD36proAB lacIq ZAM15)], and E. coliSE5000 [araD139 A(arg F- lac) U169 rpsL150 (Strr) reLAIflbBS301 deoCl ptsF25 rbsR recA56] were used as recipientsfor transformation (1, 21).

P. mirabilis HI4320 was maintained on Luria agar modi-fied to contain 2% agar and 0.5% glycerol to prevent swarm-ing (3). E. coli strains were maintained on Luria agar. Forlong-term storage, all strains were stored at -70°C in Tryp-ticase soy broth (BBL) supplemented with 20% (vol/vol)glycerol.

Plasmid pMID1003, encoding urease, was constructedfrom a gene bank cosmid clone of P. mirabilis HI4320chromosomal DNA (15). Plasmids pUC19 (31) and pBlue-script (Stratagene) were used for subcloning.

Modified urea segregation agar. Urease-positive recombi-nant clones were differentiated from urease-negative cloneson modified urea segregation agar as previously described(12).DNA isolation. Plasmid DNA was isolated from cultures

by alkaline sodium dodecyl sulfate (SDS) extraction (4) andpurified by centrifugation in cesium chloride-ethidium bro-mide density gradients (21). Single-stranded phagemid DNAwas isolated from cultures of E. coli(pMID1701) infectedwith a VCS-M13 helper phage (4 x 108 PFU) as described bySambrook et al. (31).

Oligonucleotide synthesis. Oligonucleotides were synthe-sized by the phosphoramidite method on an Applied Biosys-tems automated DNA synthesizer (model 380 B).

Site-directed mutagenesis. The T7-Gen in vitro mutagene-sis system (U.S. Biochemicals, Cleveland, Ohio), describedby Vandeyar et al. (41), was used for oligonucleotide-directed mutagenesis. To convert His-320 to Leu-320, theoligonucleotide 5' ATGGTCTGTCTTCATCTCGAT 3' wasused for priming single-stranded template (the wild-typesequence contains an A in place of the underlined T). E. coliSDM was used as the recipient for transformation withmutated single-stranded DNA.DNA sequencing. Dideoxy sequencing was performed with

a-35S-dATP (1,000 Ci/mmol; Amersham) and Sequenase(U.S. Biochemical) (34). An oligonucleotide was synthesizedfor use as a primer for sequencing the region of ureC targetedfor mutation: 5' ACGAATACGAGAITCAGC 3' is thereverse complement of nucleotides 2920 to 2937 of ureasegene sequences reported in reference 16. Labeled DNAreaction mixtures were separated by electrophoresis on6.0% polyacrylamide containing 7 M urea. Gels were fixedwith 10% acetic acid-12% methanol for 1 h and then driedunder vacuum and autoradiographed.

Preparation of antisera. Two New Zealand White-SPFrabbits (Hazelton Dutchland, Inc.) were injected subcutane-ously with column-purified urease isolated from Morganellamorganii (cross-reacts antigenically with the large subunit ofP. mirabilis urease [12]) in complete Freund adjuvant at aconcentration of 100 ,ug/ml. Rabbits were boosted at 3weeks, and blood was tested again at 5 weeks. Blood wastaken at 6 weeks by cardiac puncture.Western blot. Soluble protein (30 ,ug) from whole-cell

French press lysates of strains containing recombinant ure-ase genes was denatured in SDS-gel sample buffer andelectrophoresed on an SDS-10 to 20% polyacrylamide gra-dient gel (9). As well, soluble protein was loaded directlyonto a nondenaturing 8% polyacrylamide gel and electro-phoresed. Protein was transferred to nitrocellulose with aHoeffer Transphor Power-Lid (Model TE 50) at 100 V and500 mA for 4 h. Antisera raised against purified M. morganii

urease were used at a suitable dilution for the developmentof Western blots (immunoblots) by the method of Towbin etal. (40).

Spectrophotometric urease assay. Rates of urea hydrolysisby soluble protein in cell lysates were measured by thephenol red assay of Hamilton-Miller and Gargan (10) asdescribed previously (14). Protein concentrations were de-termined by the method of Lowry et al. (19) with bovineserum albumin as a standard.

Purification of urease. Bacterial cells (33 to 40 g [wet wt]),harvested from 4 liters of Luria broth after 24 h of growth at37°C with aeration, were suspended in 1 ml of 20mM sodiumphosphate, pH 6.8, per g of cells and lysed by passagethrough a precooled French pressure cell at 20,000 lb/in2.Insoluble material including membrane was removed bycentrifugation (two spins: 10,000 x g, 15 min, 4°C; and100,000 x g, 60 min, 4°C). Supernatant was used for purifi-cation according to the method used in our laboratory for M.morganii urease (14). Four chromatography resins (DEAE-Sepharose, Phenyl-Sepharose, Mono-Q, and Superose 6)were used (12) in conjunction with an fast-protein liquidchromatography system (Pharmacia). Urease activity wasmeasured for each fraction by the phenol red spectrophoto-metric assay. Inactive urease, synthesized by the mutant,was detected by enzyme-linked immunosorbent assay(ELISA) using rabbit anti-M. morganii urease antibody.ELISA. Column fractions were diluted 1:1,000, and sam-

ples (100 ,Il) of each dilution were assayed by ELISA usingrabbit anti-M. morganii urease polyclonal serum. The stan-dard procedure of Voller et al. (42) was used for develop-ment of the assay.

Nickel ion transport. E. coli SE5000 containing each con-struct was grown in M9 minimal medium (containing, perliter, 6 g of Na2HPO4, 3 g of KH2PO4, 0.5 g of NaCl, 1 g ofNH4CI, 10 ml of 0.01 M CaCI2, 1 ml of 1 M MgSO4, 10 ml of20% glucose, and 1 ml of 1% thiamine [21]) containing noamino acids at 37°C for 18 h. A sample (0.2 ml) from eachculture was transferred into 20 ml of fresh minimal mediumcontaining ampicillin (200 ,ug/ml), incubated with shaking(200 rpm) at 37°C until the optical density at 550 nm (OD550)reached 0.2. At this OD, protein concentrations were 0.35,0.37, and 0.31 mg/ml for E. coli transformed with pBlue-script, pMID1701, and pMID1702, respectively. Then63NiC12 (13.9 Ci/g; Amersham) was added to final concentra-tion of 0.5 ,uM and samples (1 ml) of each culture wereremoved after 0.25, 0.5, 1, 2, 5, and 10 min of furtherincubation and filtered through nitrocellulose filters (0.45-Rmpore diameter). Filters were washed three times with 3 ml ofphosphate-buffered saline (PBS), pH 7.2, and counted byliquid scintillation. Assays were done on three differentdays.

Nickel ion incorporation. E. coli SE5000 containingpMID1701, pMID1702, or pBluescript KS IIV- were cul-tured for 18 h at 37°C with aeration in 20 ml of M9 minimalmedium (see above) containing 0.025 ,uM 63NiCl2. Bacterialcells were harvested, resuspended in 20 mM sodium phos-phate, pH 6.8, and ruptured in a precooled French pressurecell at 20,000 lb/in2. The cell lysate was centrifuged (10,000x g, 10 min, 4°C) to remove cell debris, and then thesupematant was centrifuged (100,000 x g, 90 min, 4°C) toremove membrane. Protein concentrations ranged from 2.94to 4.15 mg/ml. Recombinant P. mirabilis urease was immu-noprecipitated by incubation of soluble protein (diluted to500 pg in 750 ,ul) with 1:10 and 1:20 dilutions of anti-M.morganella urease for 18 h at 4°C in 750 p.l of 2% TritonX-100-50 mM Tris-0.15 mM NaCI-0.1 mM EDTA (pH 8.0)

VOL. 61, 1993

2572 SRIWANTHANA AND MOBLEY

'aIs

pMID1701 L_

m

P. mirabilisJack bean

P. vulgarisH. pylori

U. urealyticum

c

,/

.1

I E l Ffl G |

/308 TVDEHLDHLMVCHHLDPSIP 327

581 :I::::::::::::::RR:: 600

308 327

310 :RA::X::::::::::K:: 329

314 :IA::::::::::::N:KV: 333

- 200

- 92.5

- 46K.aerogenes 310 :I:::::::::::::::D:A 329

FIG. 1. Highly conserved amino acid sequences within UreC ofP. mirabilis urease. For bacterial ureases, numbers refer to theresiduies number within the large subunit. For jack bean urease,numbers refer to the residue number within the unique subunit.Amino acids identical to the P. mirabilis sequence are indicated (:).See text for references.

- 30

i*MWAMMW- 14.3

(33). A 10% (wet wt/vol) suspension of nonviable Staphylo-coccus aureus cells (Cowan strain) in 40 mM sodium phos-phate buffer, pH 7.2-0.15 M NaCl containing 0.005% sodiumazide (Sigma) was added to the mixture, and it was incubatedfor 1 h at 4°C. After centrifugation in a microcentrifuge for 5min, the pellet was washed twice in PBS (containing, perliter, 8 g of sodium chloride, 0.2 g of potassium chloride, 1.15g of disodium hydrogen phosphate, and 0.2 g of potassiumdihydrogen phosphate, pH 7.4) containing 0.1% SDS and 1%Triton X-100 and once in 10 mM Tris, pH 8.0. The immu-noprecipitate was extracted from the pellet with 200 ptl of 1%SDS in 10 mM Tris-1 mM EDTA by heating for 10 min at

TABLE 1. Urease activity of wild type, recombinant, andmutant ureases

Sp act (,umol ofNH3/min/mg ofprotein) of cell

Strain and plasmid Comments lysates

M9 Luriamediuma broth

P. mirabilis HI4320 Wild type uninducedb 1.2 5.7(None) Inducedc 3.8 27.5

E. coli DH5ot Recombinant urease 61.4 _d(pJMK7)

E. coli SE5000pBluescript Vector control 0.12 <0.01pMID1701 Phagemid 149.0 95.4pMID1702 His-320---Leu-320 <0.01 0.02pMID1703 His-533--Leu-533 0.5 0.1pMID1704 His-472--Leu-472 11.7 3.9pMID1705 His-527--Leu-527 24.7 54.1pMID1706 His-312-- Leu-312 4.2pMID1707 His-321- yLeu-321 <0.01pMID17O8 Cys-319--Ala-319 <0.01a M9 medium c6ntaining 0.4% D-glucose; no amino acid supplement.b No urea in medium.c 50 mM urea in medium.d -,not tested.

FIG. 2. Western blot of cell lysates from wild-type and recom-binant urease strains. Soluble protein derived from bacterial celllysates was denatured and electrophoresed on SDS-polyacrylamidegradient gels (10 to 20% polyacrylamide) and transferred to nitro-cellulose by electroblotting. Western blots were developed by using1:100 anti-M. morganii urease antiserum. Lanes: Recombinant, E.coli(pMID1701); His320 Mutant, E. coli(pMID1702) in which His-320of UreC has been changed to Leu-320, P. mirabilis HI4320, parentalstrain and source of urease genes; and M. morganii TA43, source ofpurified urease against which antiserum was raised.

100°C and centrifuging for 10 min in a microcentrifuge (2).The supernatant (100 ,ul) was counted by liquid scintillation.

RESULTS

Construction of phagemid for site-directed mutagenesis.Plasmid pJMK7 was constructed by subcloning a 6.0-kbHpaI fragment, generated by partial digestion of pMID1003,into the SmaI site of pUC19 (Fig. 1). This plasmid encodesan active urease, produced at constitutive levels. Regulatorysequences encoded by ureR (28) are not present in thisclone. Phagemid pMID1701 was constructed by cloning the6.2-kb EcoRI-SalI fragment encoding urease genes frompJMK7 into pBluescript KS II/- cut with the same enzymes.Recombinant clones (white colonies) were selected on Luriaagar containing ampicillin (200 ,ug/ml) and 5-bromo-4-chloro-3-indolyl-13-D-galactosidase (20 ,uglml). Clones were ureasepositive on modified urea segregation agar (12) containingampicillin (200 ,ug/ml). Phagemid pMID1701 was used toprepare single-stranded template for site-directed mutagen-esis.

Highly conserved amino acid sequence within UreC of P.mirabilis urease. The complete amino acid sequence of jack

INFECT. IMMUN.

P. MIRABILIS UREASE: ROLE OF His-320 IN UreC 2573

03

0.4

0.3

A

0.2

0.1

0

0.35

0.3

0.25

g 0.2

aO0.15

0.1

0.05

0

1

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40Volume (ml)

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40Volume (ml)

FIG. 3. Purification of urease from E. coli(pMID1701) or E.coli(pMID1702) by column chromatography. (A) Purification ofcatalytically active urease encoded by pMID1701. Urease waschromatographed on DEAE-Sepharose, Phenyl-Sepharose, andMono-Q resins. Fractions containing peak urease activity were

pooled and applied to Superose 6. Fractions were monitored forprotein at ODm and assayed for urease activity by the phenol redspectrophotometric assay. (B) Purification of inactive mutant ureaseencoded by pMID1702. Urease was purified as described aboveexcept that catalytically inactive urease protein was monitored withan ELISA read at OD405. Pooled fractions from MonoQ containingpeak urease protein were applied to Superose 6.

bean urease has been determined by Edman degradation ofthe single-subunit polypeptide (35). The nucleotide anddeduced amino acid sequences of the ureases of five bacte-rial species, including P. mirabilis (16), Proteus vulgaris(26), Helicobacterpylori (6), Kiebsiella aerogenes (27), andUreaplasma urealyticum (5), have also been reported. Whenamino acid sequences were aligned, we demonstrated thatthe deduced amino acid sequences of UreA, UreB, andUreC of P. mirabilis (16) and the ureases of other specieshad a high similarity to the amino acid sequence ofjack beanurease (Fig. 1) (16). One region of amino acid sequence,corresponding to amino acid residues 581 to 600 ofjack beanurease and residues 308 to 327 of P. mirabilis UreC (largesubunit), is very highly conserved among all species (Fig. 1).Indeed, the amino acid sequence that aligns with residues314 to 322 of P. mirabilis UreC is invariant among all ureasesequences. This region contains Cys-592 of jack bean, a

residue postulated to play a central role in catalysis. It waspostulated that the corresponding amino acid sequencewithin UreC resides within the active site of P. mirabilisurease.

Alteration of His-320 to Leu-320 within UreC by site-directed mutagenesis. Following mutagenesis with the oligo-nucleotide as described in Materials and Methods, 40colonies were picked and screened for urease activity onmodified urea segregation agar. One transformant, pMID1702, showed no urease activity and was selected for furtheranalysis. E. coli SDM (pMID1702) produced an undetectablelevel of urease activity as compared with uninduced andurea-induced P. mirabilis H14320 and nonmutated recombi-nant enzyme synthesized by E. coli XL-1 blue (pMID1701)(Table 1).

Nucleotide sequencing. To confirm that pMID1702 con-tained the mutant nucleotide, phagemid DNA was used as atemplate for nucleotide sequencing analysis. The autoradio-graph demonstrated that CAT, a codon for His-320, hadbeen changed to CTI7, a codon for leucine (data not shown).No other changes were noted between wild-type (pMID1701) and mutant (pMID1702) sequences for 750 bp up-stream and 750 bp downstream of the putative active site.One of the 39 urease-positive clones was subjected tonucleotide sequencing and revealed no alteration in wild-type sequence within 750 bp in either direction from theintended site of mutation.

Synthesis of UreC detected by immunoblotting. The cata-lytically inactive urease encoded by pMID1702 was tested todetermine whether the alteration of one nucleotide causedany effect on expression of the large urease structuralsubunit, UreC. Soluble protein from whole-cell lysates waselectrophoresed by SDS-polyacrylamide gel electrophoresisand subjected to Western blotting with antisera to purifiedM. morganii urease (Fig. 2). UreC, synthesized by E. coliXL-1 blue (pMID1701) and mutant (pMID1702) had identicalelectrophoretic mobilities and were synthesized in qualita-tively similar amounts. The recombinant ureases both ap-peared as doublets which were not observed in the wild-typeparental strain.

Protein purification. To further investigate whether thechange from His-320 to Leu-320 resulted in any subtlealterations in charge, hydrophobicity, or native size, therecombinant and mutant ureases were subjected to a purifi-cation scheme used for ureases of other species. Fractionscontaining urease were detected by measuring enzyme ac-tivity for the active recombinant (pMID1701) enzyme or byan ELISA for the inactive mutant (pMID1702) enzyme.Both active and mutant ureases were eluted from DEAE-

Sepharose, Phenyl-Sepharose, and Mono-Q column resinsunder identical conditions and in the same fractions. Ureaseswere eluted on DEAE-Sepharose at 200 mM KCI, on Phe-nyl-Sepharose at 0 mM KCl, and on Mono-Q at 100 mMKCl. On the last column, Superose 6, a molecular sieve,both active and mutant proteins eluted at a peak 15.0 ml afterloading (Fig. 3), corresponding to an apparent molecular sizeof 250 kDa.To further determine whether the mutation affected the

native structure of the enzyme, urease protein that waseluted from the Superose 6 column was electrophoresed onnondenaturing gels, tested for urease activity (Fig. 4A), andsubjected to Western blotting (Fig. 4B). No difference inelectrophoretic mobility between recombinant and mutantenzyme was observed, again suggesting that enzyme assem-bly was not affected by the substitution of Leu-320 forHis-320.

His320 (Active)Superose 6

OD280 nm, Urease

A

VOL. 61, 1993

2574 SRIWANTHANA AND MOBLEY

911

BA

FIG. 4. Western blot of purified native urease from E. coli(pMID1701) and E. coli(pMID1702). (A) Superose 6 fractions withpeak urease activity (pMID1701) or peak ELISA values (pMID1702)were electrophoresed on an 8% nondenaturing polyacrylamide geland tested for urease activity in the gel. (B) Protein was transferredto a nylon membrane. For development of the Western blot,antiserum to purified M. morganii urease (1:100) was used as theprimary antibody and goat anti-rabbit immunoglobulin G conjugatedwith alkaline phosphatase was used as the secondary antibody. Thearrow identifies active urease in panel A and the correspondingregion in panel B.

Nickel ion requirement. E. coli (pMID1701) was grownwith shaking (200 rpm) at 37°C in M9 minimal medium withand without NiCl2 (0.25 ,ug/ml). Urease activity of celllysates was increased more than 10-fold from a specificactivity of 11 ,umol of NH3 per min per mg of protein in theabsence of NiCl2 to 121 ,umol of NH3 per min per mg ofprotein in the presence of NiCl2. This suggested that exog-

enously added Ni2+ ions could be taken into the cell,resulting in elevated urease activity, an observation consis-tent with previous studies of the wild-type enzyme (30).

Transport and accumulation of nickel by E. coli containingurease gene sequences. To determine whether urease genesequences facilitated Ni2+ transport and accumulation, we

measured the amount of Ni2+ ion taken up by E. colicarrying vector alone or cloned urease genes. Initial rates oftransport (measured in the first 1 min) for a standardizedsuspension of E. coli transformed with pBluescript,pMID1701, and pMID1702 were not significantly different(P > 0.1) at 2,572, 1,583, and 3,334 cpm of Ni2+ per min,respectively. When monitored for 60 min, E. coli containingpMID1701 or pMID1702 both tended to accumulate 63NiC12in higher amounts than E. coli containing vector alone (Fig.5). However, a significant difference (P = 0.048) in theamount of 63Ni2' accumulated was reached only after 60min. There were no significant differences (P > 0.1) betweenlevels of incorporation of 63Ni2+ within the first 60 min bycells harboring recombinant and mutant urease gene se-

quences (although the mutant tended to be lower at 60 min).

Incorporation of 63Ni2+ into urease protein. Although ini-tial rates of 63Ni2+ uptake were not significantly differentbetween urease clones and vector, the urease clones incor-porated more label at 60 min than the vector control. Thissuggested that Ni2+ ions were being incorporated into cel-lular proteins, most likely urease. To measure this, weallowed incorporation to continue for 18 h, at which timeurease was specifically immunoprecipitated from 500 ptg ofsoluble protein, derived from French press cell lysates, withantiserum (1:10 dilution) to purified M. morganii urease (Fig.6). Negligible 63Ni2+ isotope (<80 cpm) was precipitatedfrom soluble protein derived from cultures of E. coli con-taining pBluescript vector. On the other hand, antiserumbrought down 10,166 + 841 cpm from E. coli (pMID1701)expressing the catalytically active urease. Interestingly, E.coli (pMID1702) containing urease genes with the site-directed His-320 to Leu-320 alteration was able to incorpo-rate only 5,896 ± 1,179 cpm (58% compared to pMID1701) of63Ni2+, significantly less (P = 0.007) than the active recom-binant clone. Similar results were obtained with a 1:20dilution of antiserum. The urease encoded by pMID1702(Leu-320) incorporated only 59% (4,598 ± 1,215 cpm versus7,743 ± 1,260 cpm; P = 0.036) of the 63Ni2+ bound to theactive urease encoded by pMID1701 (His-320). In an inde-pendent experiment, counts per minute of 63Ni2+, precipi-tated from pMID1702, were 40 and 32% of counts per minuteprecipitated from pMID1701 with 1:10 and 1:20 dilutions ofantiserum, respectively.To determine the relative amounts of urease protein in the

soluble protein, we used an ELISA to measure proteinrecognized by the antiserum directed against M. morganiiurease. At a dilution of 1:500, OD405 readings for E. coli,transformed with pBluescript, pMID1701, and pMID1702,were 0.272, 0.324, and 0.386, respectively. These data indi-cate that the lower number of counts per minute immuno-precipitated from the pMID1702 sample was not due to thepresence of less urease protein than that produced bypMID1701. Indeed, more of the enzyme was present in thepMID1702 sample.

Additional histidine mutations. Six other site-directed mu-tations in residues of UreC were made. Three mutationsreduced, but did not eliminate, urease activity (Table 1).Plasmids pMID1704 (His-472-+Leu-472), pMID1705 (His-527- Leu-527), and pMID1706 (His-312-+Leu-312) ex-pressed 8, 17, and 3%, respectively, of the activity ofpMID1701. Plasmid pMID1703 (His-533--3Leu-533), on theother hand, expressed negligible urease activity and ap-peared to be catalytically inactive. Plasmids pMID1707(His-321-*Leu-321) and pMID1708 (Cys-319--3Ala-319) ex-pressed no detectable urease activity. E. coli transformedwith these plasmids produced comparable amounts of ureaseas observed on Western blots of nondenaturing polyacryl-amide gels (data not shown). Plasmids pMID1703 andpMID1704 produced ureases that migrated identically to thepMID1701 control protein, suggesting that the enzymes wereassembled and did not have significant alterations in quater-nary structure. Plasmid pMID1705, in contrast, produced aurease that migrated more slowly, suggesting an alteration instructure. This alteration, however, was apparently notsufficient to abolish catalytic activity.

DISCUSSION

Amino acid residues 308 to 327 of UreC, the large struc-tural subunit of the P. mirabilis urease (16), represent ahighly conserved amino acid sequence also found in ureases

INFECT. IMMUN.

01-0

'OV-1xps , le

t'.1-lill 4(6

P. MIRABILIS UREASE: ROLE OF His-320 IN UreC 2575

20000

p >.1 .091 .082 .048

15000

u

51000

0

0 10 20 30 40 50 60Minutes

pBluescript pMID 1701 pMID 1702FIG. 5. Nickel transport by E. coli containing cloned P. mirabilis urease determinants. 63NiC12 was added to exponentially growing

cultures of E. coli containing recombinant clones. Samples of the suspensions were vacuum filtered through nitrocellulose filters (0.45-pLmpore diameter) and washed with ice-cold buffer. Radiolabel retained on the filters was quantitated by liquid scintillation counting. Filterbackground was <100 cpm. Means of triplicate assays were compared by the t test. For pMID1701 versus pBluescript, P > 0.1 for all timepoints at 10 min or less; for 10 min and after, P values are indicated on the graph.

of P. vulgaris (26), H. pylon (6), U. urealyticum (5), and Kaerogenes (27), and the jack bean, Canavalia ensiformis (16)(Fig. 1). Indeed, UreC residues 314 to 322 are invariantamong all ureases for which sequence data are available.Within this sequence, Cys-319, His-320, and His-321 may bedirectly involved in the hydrolysis of urea. Cys-319 has beenpostulated to act as the general acid catalyst for the hydro-lysis (39); His-320 and His-321 are thought to coordinateNi2+ ions within the active site of this metalloenzyme (44).Changing His-320 to Leu-320 abolished urease activitywithin a clone that ordinarily expresses high levels of enzy-matically active urease (Table 1), supporting the postulatethat this residue is essential for catalysis.The supposition that His-320 resides in the active site is

well supported in the literature. The urease ofK aerogenesis made up of two independent catalytic units, each with a

stoichiometry of UreA2UreB2UreC (36); this has also beenpredicted for P. mirabilis. Using atomic absorption spectro-scopic analysis and titration of the active site with enzymeinhibitors, Todd and Hausinger (37) concluded that eachcatalytic unit possessed two Ni2+ ions, apparently in closeproximity to one another. This was consistent with earlierobservations of Dixon (7), working with the jack beanenzyme, who found that Ni2+ ions are coordinated byhistidine residues in the active site (11). One Ni2+ ion, inturn, binds a water molecule which can be displaced by thesubstrate. The other Ni2+ ion may bind a hydroxyl groupthat attacks the bound urea during hydrolysis. We find thatwhen one of the histidines (His-320) of UreC is changed toLeu-320, the inactive mutant enzyme is able to incorporate

only about half of the amount of 63Ni2" bound by the activerecombinant enzyme (Fig. 6). This would suggest that eachcatalytic unit of the mutant enzyme contains only a singleNi2+ ion.The essential cysteine (38, 39) must be in close proximity

to assist in catalysis. When this residue was altered withinUreC of the relatedK aerogenes urease by oligonucleotide-directed mutagenesis (22), enzyme activity was found to begreatly reduced but not abolished at the normal pH opti-mum. Activity could be restored at low pH, suggesting thatthe role of this residue may also be to coordinate Ni2+ ionsrather than act as an acid catalyst. Although Cys-319 appearsto play an important role in the catalytic reaction, it isapparently not essential for catalysis.Loss of activity in the mutant (Leu-320) P. mirabilis

urease is not due to lack of synthesis of structural subunits orsignificant disruption of quaternary structure of the nativeenzyme. E. coli containing pMID1701 (active urease) orpMID1702 (inactive urease) were both shown by Westernblot to synthesize UreC, in qualitatively equal amounts (Fig.2). These subunits are assembled into the native holoen-zyme. Chromatography profiles of recombinant and mutantenzymes run on DEAE-Sepharose, Phenyl Sepharose,MonoQ, and Superose 6 resins showed no differences inelution profiles, suggesting that the histidine replacement didnot significantly affect charge of the enzyme, hydrophobicproperties, or molecular size of the assembled protein (Fig.3).

It must be acknowledged, however, that alteration ofHis-320 of UreC, leading to loss of catalytic activity, does

VOL. 61, 1993

2576 SRIWANTHANA AND MOBLEY

a

CL

._

0

12000 -

10000 -

8000 -

6000 -

4000 -

2000 -

0 -

*pBluescript KSII/-pMID17O1

T *pIDi1702

T

1:10 1:20

Antibody Dilution

FIG. 6. Nickel incorporation into recombinant and mutant ure-ases. E. coli containing urease gene sequences was grown for 18 h inminimal medium containing 63NiC12. Soluble protein, derived fromFrench press lysates, was immunoprecipitated with dilutions ofantiserum raised against purified M. morganii urease. Radioactivityin the immunoprecipitate was quantitated by liquid scintillationcounting. Values are the means of triplicate determinations.

not unambiguously prove that this residue is directly in-volved in urea hydrolysis. This change could induce subtlealterations in protein conformation and perturb the activityof essential neighboring residues such as Cys-319 or His-321.It is also possible that the selected mutant, pMID1702, has asecond change in the sequence that was not detected be-cause it was farther than 750 bp from the intended mutation(i.e., beyond what was sequenced).Before nickel is incorporated into the apoenzyme, the ion

must gain entry to the cell. Uptake experiments were done todetermine whether urease gene sequences played a role inthis process. We concluded that P. mirabilis urease genesequences probably do not directly encode a nickel transportsystem, since initial rates of nickel uptake were not signifi-cantly different between E. coli containing urease clones andvector alone (Fig. 5). However, in cells expressing urease

gene sequences, there was a tendency towards higheramounts of accumulation of 63Ni2+. This suggested thaturease-related gene products facilitate uptake, perhaps byproviding a "sink" for nickel incorporation. It has beensuggested that accessory genes are necessary for insertion ofnickel into the urease apoprotein (18). To accomplish thistask, one could envision a model in which Ni2+ ions are

actively transported into the bacterium by a separatelyencoded metal ion transport system (17). Cytoplasmic Ni2l

would be bound by an accessory protein encoded by ureasegene sequences which would, in turn, insert the ions into theactive site of the newly synthesized apoenzyme possibly inan energy-dependent manner.

ACKNOWLEDGMENTS

We thank Julian Ketley for the construction of pJMK7.This work was supported by Public Health Service grants

AI23328, AG04393, and AI25567 from the National Institutes ofHealth.

REFERENCES1. Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. A.

Smith, J. G. Seidman, and K. Struhl. 1987. Current protocols inmolecular biology. Green Publishing Associates, New York.

2. Bassford, P. J., Jr., T. J. Silhavy, and J. R. Beckwith. 1979. Useof gene fusion to study secretion of maltose-binding protein intoEscherichia coli periplasm. J. Bacteriol. 139:19-31.

3. Belas, R., D. Erskine, and D. Flaherty. 1991. Transposonmutagenesis in Proteus mirabilis. J. Bacteriol. 173:6289-6293.

4. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extractionprocedure of screening recombinant plasmid DNA. NucleicAcids Res. 7:1513-1523.

5. Blanchard, A. 1990. Ureaplasma urealyticum urease genes: useof a UGA tryptophan codon. Mol. Microbiol. 4:669-676.

6. Clayton, C. L., M. J. Pallen, H. Kleanthous, B. W. Wren, and S.Tabaqchali. 1990. Nucleotide sequence of two genes fromHelicobacterpyloni encoding for urease subunits. Nucleic AcidsRes. 18:362.

7. Dixon, N. E., C. Gazzola, R. L. Blakeley, and B. Zerner. 1975.Jack bean urease. A metalloenzyme. A simple biological role fornickel? J. Am. Chem. Soc. 97:4131-4133.

8. Griffith, D. P., D. M. Musher, and C. Itin. 1976. Urease-theprimary cause of infection-induced urinary stones. Invest. Urol.13:346-350.

9. Hames, B. D. 1985. An introduction to polyacrylamide gelelectrophoresis, p. 1-86. In B. D. Hames and D. Rickwood(ed.), Gel electrophoresis of proteins-a practical approach.IRL Press, Washington, D.C.

10. Hamilton-Miller, J. M. P., and R. A. Gargan. 1979. Rapidscreening for urease inhibitors. Invest. Urol. 16:327-328.

11. Hasnain, S. S., and B. Piggott. 1983. An EXAFS study of jackbean urease, a nickel metalloenzyme. Biochem. Biophys. Res.Commun. 112:279-283.

12. Hu, L.-T., E. B. Nicholson, B. D. Jones, M. J. Lynch, andH. L. T. Mobley. 1990. Morganella morganii urease: purifica-tion, characterization, and isolation of gene sequences. J. Bac-teriol. 172:3073-3080.

13. Jones, B. D., C. V. Lockatell, D. E. Johnson, J. W. Warren, andH. L. T. Mobley. 1990. Construction of a urease-negativemutant of Proteus mirabilis: analysis of virulence in a mousemodel of ascending urinary tract infection. Infect. Immun.58:1120-1123.

14. Jones, B. D., and H. L. T. Mobley. 1987. Genetic and biochem-ical diversity of ureases of Proteus, Providencia, and Mor-ganella species isolated from urinary tract infection. Infect.Immun. 55:2198-2203.

15. Jones, B. D., and H. L. T. Mobley. 1988. Proteus mirabilisurease: genetic organization, regulation, and expression ofstructural genes. J. Bacteriol. 170:3342-3349.

16. Jones, B. D., and H. L. T. Mobley. 1989. Proteus mirabilisurease: nucleotide sequence determination and comparison withjack bean urease. J. Bacteriol. 171:6414-6422.

17. Kaltwasser, H., and W. Frings. 1980. Transport and metabolismof nickel in microorganisms, p. 463-491. In J. 0. Nriagu (ed.).Nickel in the environment. Wiley-Interscience, New York.

18. Lee, M. H., S. B. Mulrooney, M. J. Renner, Y. Markowicz, andR. P. Hausinger. 1992. Kiebsiella aerogenes urease gene cluster:sequence of ureD and demonstration that four accessory genes(ureD, ureE, ureF, and ureG) are involved in nickel metallo-center biosynthesis. J. Bacteriol. 174:4324-4330.

19. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.

INFECT. IMMUN.

P. MIRABILIS UREASE: ROLE OF His-320 IN UreC 2577

1957. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

20. Mamiya, G., K. Takishima, M. Masakuni, T. Kayumi, K.Ogawa, and T. Sekita. 1985. Complete amino acid sequence ofjack bean urease. Proc. Jpn. Acad. 61:395-398.

21. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

22. Martin, P. R., and R. P. Hausinger. 1992. Site-directed muta-genesis of the essential cysteine residue in Klebsiella aerogenesurease, p. 266. Abstr. Gen. Meet. Am. Soc. Microbiol. 1992,New Orleans, La. American Society for Microbiology, Wash-ington, D.C.

23. Mobley, H. L. T., G. R. Chippendale, K. G. Swihart, and R. A.Welch. 1991. Cytotoxicity of the HpmA hemolysin and urease ofProteus mirabilis and Proteus vulgaris against cultured humanrenal proximal tubular epithelial cells. Infect. Immun. 59:2036-2042.

24. Mobley, H. L. T., and R. P. Hausinger. 1989. Microbial ureases:significance, regulation, and molecular characterization. Micro-biol. Rev. 53:85-108.

25. Mobley, H. L. T., and J. W. Warren. 1987. Urease-positivebacteriuria and obstruction of long-term urinary catheters. J.Clin. Microbiol. 25:2216-2217.

26. Morsdorf, G., and H. Kaltwasser. 1990. Cloning of the genesencoding urease from Proteus vulganis and sequencing of thestructural genes. FEMS Microbiol. Lett. 66:67-74.

27. Mulrooney, S. B., and R. P. Hausinger. 1990. Sequence of theKlebsiella aerogenes urease gene and evidence for accessoryproteins facilitating nickel incorporation. J. Bacteriol. 172:5837-5843.

28. Nicholson, E. B., E. A. Concaugh, P. A. Foxall, M. D. Island,and H. L. T. Mobley. 1993. Proteus mirabilis urease: transcrip-tional regulation by UreR. J. Bacteriol. 175:465-473.

29. Nicholson, E. B., E. A. Concaugh, and H. L. T. Mobley. 1991.Proteus mirabilis urease: use of a UreA-LacZ fusion demon-strates that induction is highly specific for urea. Infect. Immun.59:3360-3365.

30. Rando, D., U. Steglitz, G. Morsdorf, and H. Kaltwasser. 1990.Nickel availability and urease expression in Proteus mirabilis.Arch. Microbiol. 154:428-432.

31. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory

Press, Cold Spring Harbor, N.Y.32. Setia, U., I. Serventi, and P. Lorenz. 1984. Bacteremia in a

long-term care facility: spectrum and mortality. Arch. Intern.Med. 144:1633-1635.

33. Stamm, L. V., T. C. Kerner, Jr., V. A. Bankaitis, and P. J.Bassford. 1983. Identification and preliminary characterizationof Treponema pallidum protein antigens expressed in Esche-richia coli. Infect. Immun. 41:709-721.

34. Tabor, S., and C. C. Richardson. 1987. DNA sequence analysiswith a modified bacteriophage T7 DNA polymerase. Proc. Natl.Acad. Sci. USA 84:4767-4771.

35. Takishima, K., T. Saga, and G. Mamiya. 1988. The structure ofjack bean urease. The complete amino acid sequence, limitedproteolysis and reactive cystein residues. Eur. J. Biochem.175:151-165.

36. Todd, M. J., and R. P. Hausinger. 1987. Purification andcharacterization of the nickel-containing, multicomponent ure-ase from Kiebsiella aerogenes. J. Biol. Chem. 262:5963-5967.

37. Todd, M. J., and R. P. Hausinger. 1989. Competitive inhibitorsof Klebsiella aerogenes urease: mechanisms of interaction withthe nickel active site. J. Biol. Chem. 264:15835-15842.

38. Todd, M. J., and R. P. Hausinger. 1991. Reactivity of theessential thiol of Klebsiella aerogenes urease: effect of pH andligands on thiol modification. J. Biol. Chem. 266:10260-10267.

39. Todd, M. J., and R. P. Hausinger. 1991. Identification of theessential cysteine residue of Kiebsiella aerogenes urease. J.Biol. Chem. 266:24327-24331.

40. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretictransfer of proteins from polyacrylamide gels to nitrocellulosesheets: procedure and some applications. Proc. Natl. Acad. Sci.USA 76:4350-4354.

41. Vandeyar, M., M. Weiner, C. Hutton, and C. Batt. 1988. Asimple and rapid method for the selection of oligodeoxynucle-otide directed mutants. Gene 65:129-133.

42. Voller, A., D. E. Bidwell, G. Huldt, and E. Engvall. 1974. Amicroplate method of ELISA and its application to malaria.Bull. W.H.O. 51:209-221.

43. Warren, J. W., D. Damron, J. H. Tenney, J. M. Hoopes, H. L.Muncie, and W. C. Anthony. 1987. Fever, bacteremia, and deathas complications of bacteriuria in women with long-term ure-thral catheters. J. Infect. Dis. 155:1151-1158.

44. Zerner, B. 1991. Recent advances in the chemistry of an oldenzyme, urease. Bioorgan. Chem. 19:116-131.

VOL. 61, 1993