INFECTION AND IMMUNITY, Dec. 1984, p. 839-8440019-9567/84/120839-06$02.00/0Copyright © 1984, American Society for Microbiology

Difference in Susceptibility to Gram-Negative Urinary TractInfection Between C3H/HeJ and C3H/HeN Mice

L. HAGBERG,1 R. HULL,2 S. HULL,2 J. R. McGHEE,3 S. M. MICHALEK,3 AND C. SVANBORG EDENl*

Department of Clinical Immunology, Guldhedsgatan 10, S413 46 Goteborg, Sweden1; Department of Microbiology,Baylor School of Medicine, Houston, Texas 770302; and Department of Microbiology, University ofAlabama,

Birmingham, Alabama 352943

Received 12 July 1984/Accepted 17 September 1984

The difference in susceptibility to urinary tract infection between C3HIHeJ and C3H/HeN mice was tested forwith gram-negative strains differing in lipopolysaccharide composition. Recently, impaired clearance ofEscherichia coli from the kidney of C3H/HeJ compared to C3H/HeN mice was shown to be correlated with theLPS low responsiveness. In this study, a difference in clearance from the kidneys of C3H/HeJ and C3H/HeNmice was found only with lipopolysaccharide-containing bacteria. Gram-positive bacteria, e.g., Staphylococcussaprophyticus and Streptococcus agalactiae, were recovered in essentially equal numbers from the kidneys ofmice of both strains. In contrast, of the lipopolysaccharide-containing strains used, all persisted in highernumbers in the kidneys of C3H/HeJ mice than in the kidneys of C3H/HeN mice. Variations in the 0 side chaindid not eliminate this difference. E. coli Hu734 075+K5+ and the rjb- mutant 075-K5+ remained in similarnumbers in C3H/HeJ mice, although 075-K5+ was eliminated more rapidly in C3H/HeN mice. The corestructure did not affect the differential persistence in the two mouse strains. The rjb mutants with R1-R4 coreswere eliminated after 24 h from the C3H/HeN mice, but remained in significant numbers in the kidneys of C3H/HeJ mice. Even the Re mutant of Salmonella minnesota persisted in low numbers in C3H/HeJ mice. Therelative bacterial recovery from either mouse strain was related to the overall virulence of the infectingbacterial strain, but the difference between C3H/HeJ and C3H/HeN mice was associated with responsiveness toparts of lipopolysaccharide common to the bacterial strains tested.

Stimulation of antibacterial defense mechanisms throughlipopolysaccharide (LPS) is important for resistance togram-negative bacterial infections (2, 9, 21). C3H/HeJ miceare hyporesponsive to LPS; i.e., they do not respond toconcentrations of LPS which induce mitogenic responses inother mice, e.g., the syngeneic C3H/HeN strain (28, 34). Theinability of C3H/HeJ mice to respond to LPS has beenattributed to a defective Lps gene (35). C3H/HeJ mice alsoexhibit increased susceptibility to parenteral Salmonellatyphimurium infection (18, 20, 33). We recently demonstrat-ed impaired clearance of ascending Escherichia coli infec-tion from the kidneys of C3H/HeJ mice as compared to C3H/HeN mice (29; C. Svanborg Eden, submitted for publica-tion). Back-cross analysis revealed a correlation between E.coli persistence in the kidneys and responsiveness to LPS.In the present study, a complementary approach was used toevaluate the role of LPS-induced host effector functions forthe clearance of bacteria from the urinary tract, i.e., infec-tion with bacterial strains differing in LPS composition inC3H/HeJ and C3H/HeN mice.The LPS of gram-negative bacteria is a complex of mole-

cules composed of the repeating 0 antigenic oligosaccha-rides, a core region, and lipid A (12, 27). Lipid A hasmitogenic and toxic activities; the oligosaccharides enhancethe virulence (13, 31). Although structural variants of lipid Aare known (32), there are no genetic techniques known foraltering the lipid A moiety in E. coli while leaving other traitsintact. In contrast, distinct loci on the E. coli chromosomemay be mutated to vary the length of the oligosaccharidechain (12). Rfb- mutants are unable to synthesize the 0 sidechain, but they retain a complete core. Rfa- mutants havemore or less incomplete core structures, and they fail to

* Cotresponding author.

anchor the synthesized 0 polysaccharides (19). A limitednumber of E. coli core oligosaccharide structures have beendescribed, R1 through R4 and K-12 (23). In this report,strains differing in LPS content, oligosaccharide chainlength, and core composition were used to study the role ofLPS in the susceptibility of C3H/HeJ mice. The results showa difference in clearance from the kidneys of C3H/HeJ andC3H/HeN mice only for LPS-containing bacteria. A domi-nant role of lipid A is suggested by the higher susceptibilityof C3H/HeJ mice to all lipid A-containing bacterial strains,including Salmonella minnesota Re.

MATERIALS AND METHODSMouse strains. C3H/HeJ mice were purchased from the

original breeder, Jackson Laboratory, Bar Harbor, Maine.C3H/HeN mice were purchased from Charles River U.K.Ltd, Margate, Kent, England. Female mice older than 6weeks were used and matched by age for individual experi-ments.

Bacteria. Bacteria used in the study are listed in Table 1.The gram-positive strains used were one recent urinaryisolate of Staphylococcus saprophyticus and a strain ofStreptococcus agalactiae originating from the RockefellerUniversity collection (no. 090R). E. coli 414 was isolatedfrom the stools of a healthy child. E. coli strains Hu734,Hu973, and Hu972 were chemically induced mutants ofpyelonephritis E. coli isolate GR12 (5). Hu734, a lacmutant, was isolated after treatment with nitrous acid (15)but retained other properties from GR12. Hu973 and Hu972were constructed in the following way: GR-12 was mutagen-ized with nitrous acid as described by Miller (15), andnonreverting histidine auxotrophs (Hu748 his-) were identi-fied after two cycles of ampicillin-D-cycloserine enrichment(3). A spontaneous nalidixic acid-resistant derivative ofHu748 (Hu969) was selected on agar containing nalidixic

839

Vol. 46, No. 3

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

840 HAGBERG ET AL.

acid (100 pg/ml). Hu969 was mated with an E. coli K-12 Hfrdonor, PK191 (10) (Genetic Stock Center, Yale University,New Haven, Conn.), for 1 h on filters as previously de-scribed (8). Histidine-independent exconjugants were testedfor coinheritance of the rfb locus of E. coli K-12 (1), by usinga commercially prepared anti-075 serum (Difco Labora-tories, Detroit, Mich.) in a slide agglutination assay. Hu972was 075-, and Hu971 was 075+. Hu971, the his+075+exconjugant, was plated on agar containing rifampin (200 ,ug/ml) to select a spontaneous rifampin-resistant mutant(Hu973). All strains were tested for hemagglutination prop-erties as previously described (5). The construction of the E.coli mutants R, through R4 has been described by Schmidtand co-workers for F614 and F653 (24), F470 and F576 (25),and F2513 (23). Salmonella typhimurium SR-11 (26) was amouse virulent strain obtained from D. Briles, Departmentof Microbiology, University of Alabama, Birmingham.

Inoculum. E. coli and Salmonella strains were grown ontryptic soy agar, TSA (BBL Microbiology Systems, Cock-eysville, Md.), and Streptococcus group B and Staphylococ-cus saprophyticus were grown on blood agar (Difco) over-night. E. coli Hu973 Rif' was maintained on TSA-rifampicinplates (100 ,ug/ml). For infection, it was passaged once onantibiotic-free TSA. After harvest, the bacterial cell densitywas adjusted by absorbance at 597 nm to a concentration of109 bacteria per ml by dilution in phosphate-buffered saline(300 mosm/liter [pH 7.2]).

In the mixed-infection experiments, equal numbers ofbacteria from each strain (109 bacteria per ml) were mixed inthe same inoculum. The concentration of each strain in theE. coli Hu734-414 mixture was easily determined on lactose-bromthymol blue agar plates, since E. coli 414 fermentedlactose and formed yellow colonies, whereas E. coli Hu734colonies remained blue. The exact concentration of eachstrain in the E. coli Hu973-972 mixture was determined by

serial dilutions on TSA and TSA-rifampicin plates. Hu973Rif was quantitated as the number of colonies on the TSA-rifampicin plates, and Hu973 was quantitated by subtractionof the colonies on TSA-rifampicin from those on TSA.Control experiments with passage of Hu973 Rif' on antibiot-ic-free medium or in mice did not show segregation of theantibiotic marker.

Virulence-associated traits. The E. coli strains were testedfor 0 and K antigens, hemolysin production, and resistanceto serum killing (30). Adhesins were tested by agglutinationof human and guinea pig erythrocytes (5) and attachment tohuman and mouse uroepithelial cells (5). E. coli strainsHu734, 972, and 973 retained from GR12 adhesins specificfor mannosides and globoseries glycolipid receptors. The075- strain showed, however, increased susceptibility tokilling by serum (manuscript in preparation). F 2513 carriedadhesins specific for mannosides. The remaining rfa, rfb,and Re mutants did not express adhesins.

Experimental urinary tract infection. The precise detailsconcerning the infection model have been previously report-ed (5). In brief, the animals were anesthetized by etherinhalation. The bladder of each animal was emptied bygentle compression of the abdomen. Immediately thereafter,a soft polyethylene catheter (outer diameter, 0.61 mm; KeboGrave, Sweden) adapted to a needle (0.4 by 20 mm) on a 1 mltuberculin syringe (Asik, Denmark) was transurethrally in-serted into the bladder. A 0.05-ml amount of the appropriatemicrobial inoculum was injected. The catheter was with-drawn immediately, and no further manipulations wereperformed. The ether anesthesia lasted for about 30 s, afterwhich time the mice were allowed food and drink ad libitum.

Bacterial recovery from tissues. After 24 h or 4 days,animals were sacrificed by cervical dislocation. Kidneys andbladders were removed aseptically. The tissues were homog-enized in 5 ml of phosphate-buffered saline in disposable

TABLE 1. Bacterial strains used for experimental urinary tract infectionLipopolysaccharide structure

Strain Origin Other virulence propeities0 antigen Core

E. coli GR12 Acute pyelonephritis isolate (reference 5) 075+ Complete core Capsule (K5), resistant to theE. coli Hu734 lac- mutant of GR12 (reference 5) 075+ Complete core bactericidal effect of serum,E. coli Hu973 Recombinant of GR12(his-) 075+ Complete core ColV+, hemolysin negative.E. coli Hu972 Recombinant of GR12 (his-) 075- Complete core Carry adhesins specific for

the globoseries glycolipidreceptors and"mannosides"

E. coli 414 Fecal strain (reference 5) 019, 023 Complete core Capsule (nontypable),sensitive to the bactericidaleffect of serum, hemolysinnegative, nonadhering

Salmonella typhimurium (reference 26) Mouse virulentSR-11

E. coli F614 rfa Mutant of 08:K27 (reference 24) R1 complete core Capsule (nontypable),E. coli F470 rfb Mutant of 08:K27 (reference 25) R1 complete core sensitive to the bactericidalE. coli F576 rfb Mutant of 08:K42 (reference 25) R2 complete core effect of serum, hemolysinE. coli F653 rfb Mutant of 0111 (reference 24) R3 complete core negative, R1-R3, Re595 areE. coli F2513 rfb Mutant of 014:K7 (reference 23) R4 complete core nonadhesive. R4 cariesSalmonella sp. Re595 Mutant of S. minnesota (KDO)2 adhesins specific for

mannosides

Streptococcus agalactiae Rockefeller University collection no.O90R

Staphylococcus saprophyticus Human urine

INFECT. IMMUN.

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

URINARY TRACT INFECTION, C3H/HeJ AND C3H/HeN MICE 841

TABLE 2. Ascending urinary tract infection with E. coli Hu734 and Salmonella typhimurium SR-11Bacterial recovery, geometric mean (SEF)a

Bacteria Mouse No. ofstrain animals Kidney C3H/He b Bladder C3H/HeNb

E. coli Hu734 C3H/HeN 20 247 (3.3) 174C 187 (3.4) 1.4C3H/HeJ 20 42,768 (1.8) 269 (3.2)

Salmonella typhimurium SR-li C3H/HeN 10 515 (2.7) 136C 3,668 (6.5) 0.68C3H/HeJ 10 70,195 (1.8) 2,478 (8.0)

a Bacterial recovery 24 h after infection. Geometric mean with standard error factor (SEF), the number by which he mean must be multiplied and divided to getthe upper and lower standard error, respectively.

b Ratio of bacteria recovered from C3H/HeJ to that recovered from C3H/HeN.c Significantly more bacteria were recovered in C3H/HeJ kidneys than in C3H/HeN kidneys (P < 0.001).

plastic bags with a Stomacher 80 homogenizer. Serial dilu-tions of the homogenate were performed on appropriate agarplates. The bacterial recovery was calculated as the log ofCFU in a 0.1-ml homogenate. In the mixed-infection proto-col, each strain was calculated as R = (concentration in thetissue)/(concentration in the inoculum). The relative recov-ery of bacteria from each animal of strains a and b in theinoculum mixture was calculated as the ratio Ra/Rb.

Statistics. Differences in bacterial recovery were evaluatedby the Student t test. Due to a large positive skewing of therecovery values, a transformation to logarithmic scale wasundertaken, and the antilog of the mean, the geometricmean, was used to present the results. The standard error ofthe mean was also calculated from the log values and theantilog; the standard error factor was used in the presenta-tion of results. The standard error factor is the number bywhich the mean must be multiplied and divided to get theupper and lower standard error, respectively.

RESULTSPersistence of bacteria in relation to their LPS content. The

susceptibilities ofC3H/HeJ and C3H/HeN mice to ascendingurinary tract infection were compared by using E. coli strainHu734, which earlier had been used to develop the mousemodel in CBA/CA mice. C3H/HeJ mice had ca. 170 timeshigher bacterial recovery from kidneys than did C3H/HeN(Table 2). No significant difference was found in recovery ofbacteria from the bladders of mice of the two strains. Thefirst line of evidence that the susceptibility of C3H/HeJ mice

was linked to the LPS moiety was obtained with anothergram-negative species, Salmonella typhimurium SR-11 (Ta-ble 2). This strain was recovered in ca. 140 times greaternumbers from the kidneys of C3H/HeJ mice than from thoseof C3H/HeN mice 24 h after infection.The second line of evidence was the comparison of

susceptibilities to the gram-positive bacteria, Streptococcusagalactiae and Staphylococcus saprophyticus. The bacterialrecovery from kidneys and bladders is shown in Table 3,Both strains persisted for at least 4 days in the kidneys ofC3H/HeJ and C3H/HeN mice, yet no significant differencein bacterial recovery from kidneys or bladders was foundbetween the two mouse strains.

Persistence of E. coli strains differing in clinical origin. Thesusceptibilities of C3H/HeJ and C3H/HeN mice were thencompared by using E. coli strains of different clinical origin.The mice were infected with mixtures of the acute pyelone-phritis strain Hu734 and the fecal strain 414 (Table 4). Thetotal bacterial recovery from kidneys of C3H/HeJ mice wasca. 30 times higher than that from those of C3H/HeN mice.E. coli Hu734 was recovered in significantly higher numbersthan E. coli 414 was from both kidneys and bladders of C3H/HeJ and C3H/HeN mice (P < 0.05).

Persistence of E. coli differing in 0 antigen. The role of the0-specific oligosaccharide side chain in the persistence inthe mouse urinary tract was analyzed by infection withmixtures of the homogeneic mutants Hu973 (Rif) 075+ andHu972 075- (Table 5). In both the kidneys and bladders ofthe C3H/HeN mice 1 and 4 days after infection, the 075+

TABLE 3. Lack of difference between LPS low responder mice in the persistence in the urinary tract of non-LPS-containing bacteria,Streptococcus agalactiae and Staphylococcus saprophyticus

Time Bacterial recovery, (SEF)'Mouse No. ofBacteria strain afer animals Kidney C3H/HeJ to Bladder C3H/HeJ toinfection Kidey C3H/HeNb Badr C3H/HeNto

Streptococcus agalactiae C3H/HeN 1 Day 10 1,412 (1.4) 1.1 1,023 (2.1) 0.8C3H/HeJ 8C 1,513 (1.4) 794 (3.1)

Staphylococcus saprophyticus C3H/HeN 10 1,438 (1.3) 2.2 22.2 (1.4) 1.0C3H/HeJ 10 3,224 (1.7) 2 21.2 (1.6)

Streptococcus agalactiae C3H/HeN 4 days 10 1,995 (1.2) 1.4 5.6 (1.5) 0.6C3H/HeJ 71 2,818 (1.2) 3.2 (1.8)

Streptococcus saprophyticus C3H/HeN 9C 474 (2.3) o 5 5.8 (2.3) 0.3C3H/HeJ 9 225 (3.0) 0 2.0 (1.4)a Mean and SEF.b See footnote b of Table 2; no significant differences were found between C3H/HeJ and C3H/HeN mice.c Two animals in each group were contaminated with Proteus spp. and were excluded.d One animal was contaminated with Proteus spp. and was excluded.

VOL. 46, 1984

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

842 HAGBERG ET AL.

TABLE 4. Ascending urinary tract infection with a mixedinoculum of E. coli strains Hu734 and 414

Bacteria recovered (SEF)'Organ, mouse strain

Total Relativeb

KidneyC3H/HeJ 6761 (1.4) 20.9 (3.2)C3H/HeN 224 (1.5) 14.8 (1.5)C3H/HeJ to C3H/HeN 30.2

BladderC3H/HeJ 316 36.3 (3.1)C3H/HeN 42.6 (3.6) 42.7 (3.6)C3H/HeJ to C3H/HeN 7.4a Mean and SEF. Eight mice per group.bRatio of Hu734 to 414 recovered from each animal.c See footnote b of Table 2.

mutant persisted in significantly higher numbers. In theC3H/HeJ mice, no significant difference was observed be-tween the 075+ and 075- mutants. The total bacterialrecovery from C3H/HeJ kidneys was ca. 270 times greaterafter 24 h and ca. 870 times greater after 4 days than thatfrom C3H/HeN kidneys.

Persistence in relation to core structure. Mice were infectedwith E. coli core mutants R, through R4 or Salmonellaminnesota Re to determine whether changes in the LPS corecomposition affected the difference in susceptibility (Table6). All the mutants except F 2513 persisted in lower numbersthan the wild-type E. coli or Salmonella species in thekidneys of C3H/HeJ and C3H/HeN mice. The core composi-tion did not influence the ability of bacteria to survive inmuch larger numbers in C3H/HeJ kidneys than in C3H/HeNkidneys. The E. coli F 2513 R4 mutant differed from E. colimutants R, through R3 in adhesive properties (Table 1).

In the urinary bladders, low numbers of bacteria persistedin both mouse strains, except for E. coli F 614, which wasrecovered in significantly higher numbers in C3H/HeJ mice,and F 2513, which persisted significantly better in the C3H/HeN mice.

DISCUSSIONBacteria differing in the content or composition of LPS

were used to study the difference in the susceptibilities ofC3H/HeJ and C3H/HeN mice to ascending urinary tract

infection. The deficient clearance of E. coli from the kidneysof C3H/HeJ mice was recently shown to be linked to theLpsd phenotype (29). The importance of LPS in triggeringhost defense mechanisms in C3H/HeN but not in C3H/HeJmice was supported by the results of the present study.The elimination from the kidneys and bladders of bacteria

not containing LPS, e.g., Streptococcus agalactiae andStaphylococcus saprophyticus, did not significantly differbetween the two mouse strains. In contrast, all E. colistrains, Salmonella typhimurium SR-11, and Salmonellaminnesota Re595, regardless of LPS side-chain composition,persisted in higher numbers in the kidneys of C3H/HeJ thanin the kidneys of C3H/HeN mice. Variations in the 0 sidechain or core structure did not eliminate this difference.The E. coli strains of this study were selected to differ in

two ways: in virulence-associated traits or in LPS composi-tion. The overall virulence of the bacterial strains deter-mined their ranking within each mouse strain but did notaffect the relative difference between their levels of virulencefor C3H/HeJ and C3H/HeN mice. The pyelonephritis strainE. coli Hu734 persisted ca. 15 to 42 times more than the fecalstrain 414 did in both mouse strains. A similar level ofdifference in recovery of the two E. coli strains was previ-ously noted in CBA mice (5). This suggests that the virulencefactors of Hu734 which were lacking in 414 contributedequally in C3H/HeJ and C3H/HeN mice and that hostresponses activated by those bacterial properties were simi-lar in the two mouse strains.The variation in LPS composition may include the 0 side

chain, the core, or lipid A. In wild-type isolates, loss of the 0antigen often coincides with loss of other virulence traits,e.g., adhesion, capsule, resistance to serum killing, etc. (14).The mutants of GR12, Hu734, 972, and 973 were isogenic,i.e., did not differ in other measurable virulence-associatedtraits, core structure, or lipid A composition. The 0 sidechain did not seem to affect bacterial persistence in the C3H/HeJ mouse. Possibly, the greater susceptibility of the C3H/HeJ mice may have made it difficult to distinguish a smallereffect due to the 075 antigen. Alternatively the C3H/HeJmice did not have defense mechanisms inactivated by the075 antigen. This was in contrast to the C3H/HeN mice, inwhich the 075+ mutant persisted significantly better thanthe 075- mutant. A difference in virulence related to thelength of the 0 side chain was previously shown, forexample, in experimental Salmonella infection (13, 16, 31).

TABLE 5. Role of the 0 side chain in the persistence of E. coli in the urinary tract of LPS responder and low-responder miceBacteria recovered (SEF)'

Organ, mouse strain Day 1 Day 4Total Relativeb Total Relativeb

KidneyC3H/HeJ 10,000 (1.7) 0.9 (2.3)C 91,201 (1.3) 1.7 (1.4)CC3H/HeN 37.2 (2.5) 11.0 (2.3)d 105 (1.9) 4.7 (1.4)dC3H/HeJ to C3H/HeJC 269 869

BladderC3H/HeJ 191 (1.9) 2.3 (1.5)' 251 (1.6) 3.6 (2.0)'C3H/HeN 40.7 (3.4) 24.5 (2.6)d 11.5 (1.3) 5.6 (1.4)'C3H/HeJ to C3H/HeN' 4.7 21.8a Bacterial recovery after infection with a mixture of the isogenic E. coli strains Hu973 (075') and Hu972 (075-). For day 1, groups consisted of 8 mice; for day

4, groups consisted of 10 mice.b Recovery of 075+ strain versus that of 075- strain from each mouse.c Recovery of 075+ and 075- strains not significantly different.d p < 0.05.e See footnote b of Table 2.fp < 0.01.

INFECT. IMMUN.

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

URINARY TRACT INFECTION, C3H/HeJ AND C3H/HeN MICE 843

TABLE 6. Influence of the LPS core structure on the persistence of E. cOli in the urinary tract of LPS responder and LPS low-respondermice

Bacteria recovered (SEF)"

Bacteria Kidney Bladder

C3H/HeJ C3H/HeN C3H/HeN C3H/HeJ C3H/HeN C3H/HeN

E. coli F614 R1 2,138 (1.9) 3.6 (2.1) 594 159 (2.3) 2.6 (1.6) 61'F470 R1 3,084 (2.6) 1.7 (1.7) 632 33.9 (1.2) 1.6 (1.3) 18'F576 R2 1,899 (2.6) 1.2 (1.1) 985 7.5 (2.1) 1.1 (1.1) 3.2F653 R3 1,106 (1.6) 3.8 (2.2) 410 4.5 (1.5) 7.4 (2.4) 1.1F2513 R4 8,860 (1.5) 26.2 (2.3) 340 2.0 (1.6) 199 (3.6) 0.009C

Salmonella minnesota Re 83.2 (1.3) 2.3 (1.6) 36 1.8 (3.2) 1 (1) 1.8a Bacterial recovery 24 h after infection (mean and SEF); 10 mice per group.b See footnote b of Table 2. Significant difference (P < 0.01) was found between C3H/HeJ and C3H/HeN mice for recovery of all bacteria from kidneys.' Significant difference between C3H/HeJ and C3H/HeN mice ( P < 0.01).

The removal of the 075 antigen increased the hydrophobic-ity and susceptibility to serum killing of the 075- mutant.This may facilitate the interaction with phagocytic cells (11)and survival in vivo (22). The exact mechanisms of clearancerelated to the 075 antigen still remain undefined.The core mutants R1 through R4 varied both in virulence-

associated traits and LPS composition, compared both toeach other and to the isogenic GR12 mutants. The relativelylow recovery of the core mutants from both C3H/HeJ andC3H/HeN mice, compared to that from the isogenic GR12mutants, was consistent with their less virulent phenotypes.Indeed, the E. coli R4 strain, which was recovered in highernumbers than mutants R, through R3, expressed mannose-sensitive adhesins, which have been shown to contribute topersistence in the mouse bladder (6). The lowest recoverywas found with Salmonella minnesota Re in both mousestrains. The ratio of recovery of the core mutants from C3H/HeJ and C3H/HeN was, however, not affected by the corecomposition and was largely of the same magnitude as in themore virulent strains. Taken together, these data stronglypoint to lipid A as the determinant of the increased suscepti-bility of C3H/HeJ mice. Mutants of pyelonephritis strainsdiffering in lipid A but retaining the core and 0 side chainoligosaccharide structures remain to be developed.C3H/HeJ and C3H/HeN mice have previously been shown

to differ in susceptibility to lethal effects of Salmonellatyphimurium (18, 20, 33). In this infection model, E. coliremained localized in the urinary tract for at least 1 monthand did not kill the mice or cause detectable disease (L.Hagberg, D. Briles, C. Svanborg Eden, submitted for publi-cation). The role of the Lpsd phenotype for susceptibility toSalmonella typhimurium infection was recently questionedby the results showing equal susceptibility of high- and low-responder mice (4, 17). In contrast, thefersistence of E. coliin the kidney correlated with the Lps phenotype in C3H/HeJ, C3HeB/FeJ, and C3H/HeN mice (Hagberg et al.,submitted).The high susceptibility of the C3H/HeJ mouse to E. coli

infection in the urinary tract makes it a suitable tool forstudying the virulence of genetically manipulated bacteria, inparticular, E. coli K-12 as a host for recombinant DNA. K-12derivates are eliminated from the urinary tract in normalmice, including C3H/HeN, but are retained in C3H/HeJ mice(C. Svanborg Eden, unpublished data). Single virulenceproperties like the presence of adhesins on E. coli have beenfound to enhance persistence in the mouse urinary tract instudies with chemically induced mutant strains and recombi-

nant DNA transformants differing only in adhesins (6). Thecontribution of other factors or virulence determinants, e.g.,hemolysin, serum resistance, capsule, or combinationsthereof, is presently being analyzed.

ACKNOWLEDGMENTSThe rfb mutants R1 through R4 were originally constructed and

kindly provided by G. Schmidt, B. Jann, and K. Jann, Max PlanckInstitut for Immunbiologie, Freiburg, Federal Republic of Germany.Salmonella typhimurium SR-11 was kindly provided by D. Briles,Department of Microbiology, University of Alabama, Birmingham,who also read and criticized the manuscript. The excellent technicalassistance of Inga Engberg and the typing aid of Anne-Bell Ek andAnn-Charlotte Malmefeldt are greatly appreciated.

This study was supported by grants from the Swedish MedicalResearch council (No. 215), the Medical Faculty, University ofGoteborg, The Ellen, Walter and Lennart Hesselman Foundationfor Scientific Research, and U.S. Public Health Service grant Al18462.

LITERATURE CITED1. Bachmann, B. J. 1983. Linkage map of Escherichia coli K-12,

Edition 7. Microbiological Rev. 47:180-230.2. Chedid, L., M. Parant, F. Parant, and F. Boyer. 1968. A

proposed mechanism for natural immunity to enterobacterialpathogens. J. Immunol. 100:292-301.

3. Curtiss, R. III, L. J. Charamella, C. M. Berg, and P. E. Harris.1965. Kinetic and genetic analyses of D-cycloserine inhibitionand resistance in Escherichia coli. J. Bacteriol. 90:1238-1250.

4. Eisenstein, T., L. W. Deakins, L. Killar, P. H. Saluk, and B. M.Sultzer. 1982. Dissociation of innate susceptibility to Salmonellainfection and endotoxin responsiveness in C3HeB/FeJ mice andother strains in the C3H lineage. Infect. Immun. 36:696-703.

5. Hagberg, L., I. Engberg, R. Freter, J. Lam, S. Oiling, and C.Svanborg Eden. 1983. Ascending, unobstructed urinary tractinfection in mice caused by pyelonephritogenic Escherichia coliof human origin. Infect. Immun. 40:273-283.

6. Hagberg, L., R. Hull, S. Hull, S. Falkow, R. Freter, and C.Svanborg Eden. 1983. Contribution of adhesion to bacterialpersistence in the mouse urinary tract. Infect. Immun. 40:265-272.

7. Hull, R. A., R. E. Gill, P. Hsu, B. H. Minshew, and S. Falkow.1981. Construction and expression of recombinant plasmidsencoding type 1 or D-mannose-resistant pili from a urinary tractinfection Escherichia coli isolate. Infect. Immun. 33:933-938.

8. Hull, S. I., R. A. Hull, B. H. Minshew, and S. Falkow. 1982.Genetics of hemolysin of Escherichia coli. J. Bacteriol.151:1006-1012.

9. Landy, M. 1956. Increase in resistance following administrationof bacterial lipopolysaccharides. Ann. N.Y. Acad. Sci. 66:292-303.

VOL. 46, 1984

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

844 HAGBERG ET AL.

10. Low, K. B. 1973. Escherichia coli K-12 F-prime factors, old andnew. Bacteriol. Rev. 36:587-607.

11. Magnusson, K. E., J. Davies, T. Grundstrom, E. Kihistrom, andS. Normark. 1980. Surface charge and hydrophobicity of Salmo-nella, E. coli, Gonococci in relation to their tendency toassociate with animal cells. Scand. J. Infect. Dis. 24:135-140.

12. Makela, P. H., and B. A. D. Stocker. 1969. Genetics of polysac-charide biosynthesis. Ann. Rev. Genet. 3:291-322.

13. Makela, P. H., V. V. Valtonen, and M. Valtonen. 1973. Role of0-antigen (lipopolysaccharide) factors in the virulence of Sal-monella. J. Infect. Dis. 128(suppl.):81-85.

14. Mattsby-Baltzer, I., L. A. Hanson, B. Kaijser, P. Larsson, S.Oiling, and C. Svanborg Eden. 1982. Experimental Escherichiacoli ascending pyelonephritis in rats: changes in bacterial prop-erties and the immune response to surface antigens. Infect.Immun. 35:639-646.

15. Miller, J. H. 1972. Experiments in molecular genetics, p. 135-139. Cold Spring Harbor, Cold Spring Harbor, N.Y.

16. Nakano, M., and K. Saito. 1969. Chemical components in thecell wall of Salmonella typhimurium affecting its virulence andimmunogenicity in mice. Nature (London) 222:1085-1086.

17. O'Brien, A. D., and D. L. Rosenstreich. 1983. Genetic control ofthe susceptibility of C3HeB/FeJ mice to Salmonella typhimur-ium is regulated by a locus distinct from known salmonellaresponse genes. J. Immun. 131:2613-2615.

18. O'Brien, A. D., D. L. Rosenstreich, I. Scher, G. Campbell, R. P.MacDermott, and S. B. Formal. 1980. Genetic control of suscep-tibility to Salmonella typhimurium in mice: role of the Lps gene.J. Immunol. 124:20-24.

19. 0rskov, I., F. 0rskov, B. Jann, and K. Jann. 1977. Serology,chemistry, and genetics of0 and K antigens of Escherichia coli.Bacteriol. Rev. 41:667-710.

20. Robson, H. G., and S. I. Vas. 1972. Resistance of inbred mice toSalmonella typhimurium. J. Infect. Dis. 126:378-386.

21. Rowley, D. 1955. Stimulation of natural immunity to Escherichiacoli infections. Lancet i:232-234.

22. Rowley, D. 1968. Sensitivity of rough gram-negative bacteria tothe bactericidal action of serum. J. Bacteriol. 95:1647-1650.

23. Schmidt, G., B. Jann, and K. Jann. 1974. Genetic and immuno-chemical studies on Escherichia coli 014:K7:H-. Eur. J. Bio-

chem. 42:303-309.24. Schmidt, G., B. Jann, and K. Jann. 1970. Immunochemistry of

R lipopolysaccharides of Escherichia coli; studies on R. mutantswith an incomplete core, derived from E. coli 08:K27. Eur. J.Biochem. 16:382-392.

25. Schmidt, G., B. Jann, and K. Jann. 1969. Immunochemistry ofR lipopolysaccharides of Escherichia coli; different core regionsin the lipopolysaccharides of 0 group 8. Eur. J. Biochem.10:501-510.

26. Schneider, H. A., and N. D. Zinder. 1956. Nutrition of the hostand natural resistance to infection. V. An improved assayemploying genetic markers in the double strain inoculation test.J. Exp. Med. 103:207-223.

27. Stocker, B. A. D., and H. P. Makela. 1971. In Microbial toxins,(G. Weinbaum, S. Kadis and S. J. Ajl, eds.), Vol. 4, 369-433.Acad. Press, N.Y.

28. Sultzer, B. M. 1968. Genetic control of leucocyte responses toendotoxin. Nature (London) 219:1253-1254.

29. Svanborg Eden, C., D. Briles, L. Hagberg, J. McGhee, and S.Michalek. 1984. Genetic factors in host resistance to urinarytract infection. Infection 12:118-123.

30. Svanborg Eden, C., L. Hagberg, L. A. Hanson, H. Leffler, and S.Olling. 1983. Adhesion of Escherichia coli in urinary tractinfection. CIBA Found. Symp. 80:161-187.

31. Valtonen, V. V. 1969. Virulence of salmonella strains with areduced amount of 0-antigen. J. Gen. Microbiol. 57:28-29.

32. Vogel, S. N., G. S. Madonna, L. M. Wahl, and P. D. Rick. 1984.In vitro stimulation of C3H/HeJ spleen cells and macrophagesby a lipid A precursor molecule derived from Salmonellatyphimurium. J. Immunol. 132:347-353.

33. von Jeney, N., E. Gunther, and K. Jann. 1977. Mitogenicstimulation of murine spleen cells: relation to susceptibility toSalmonella infection. Infect. Immun. 15:26-33.

34. Watson, J., and R. Riblet. 1974. Genetic control of responses tobacteria lipopolysaccharides in mice. I. Evidence for a singlegene that influences mitogenic and immunogenic responses tolipopolysaccharides. J. Exp. Med. 140:1147-1161.

35. Watson, J., R. Riblet, and B. A. Taylor. 1977. The response ofrecombinant inbred strains of mice to bacterial lipopolysaccha-rides. J. Immunol. 118:2088-2093.

INFECT. IMMUN.

on February 16, 2020 by guest

http://iai.asm.org/

Dow

nloaded from