Mouse monoclonal antibodies (MAbs) specific for the lipopolysaccharide (LPS) core types R1, R2, and R3 of Escherichia coliand a cross-reactive MAb that binds to the LPS core of almost all E. coli were used in ELISA to determine the frequency of cores resembling R1, R2, and R3 in strains of E. coli isolated from clinical samples (blood and urine specimens) and from the feces of asymptomatic individuals. Of the 180 wild-type isolates, 123 were assigned to R1 core type, 14 to R2, and 18 to R3. Twenty-five wild-type E. coliisolates could not be assigned to a particular core type and may have either an R4 or K12 core or a previously unrecognized core type. R1 core type was associated with 0 types 1,4,6,8, and 18 and with K1 or K5 capsules. R3 was associated with 015. 075 isolates could be of either R1 or R2 core type.The lipopolysaccharide (LPS) core region of Escherichia coli has a conserved overall structure, with an inner 2-keto-3-deoxy-octonate-heptose region and an outer hexose region. Most of the variation that does occur is in the outer part of the core, furthest from lipid A. In wild-type smooth E. coli. the core region can occur in at least four different forms, known as R 1, R2, R3, and R4 [1][2][3]. The naturally rough strain Kl2 has a different core region [4].The 0 antigen of smooth LPS may mask some of the epitopes and phage attachment sites in the core region. The chemical, serologic, and phage typing methods that have been used to define the different core types of E. coli are therefore not directly applicable to wild-type smooth strains. The frequency with which these different core types occur in wild-type E. coli is therefore not known.Little is known about the association between core types and 0 serotypes ofE. coli. except that the original R 1 and R2 rough mutants were both derived from 08 strains [I] and that the original R3 mutant was derived from an 0 III strain [2]. There is no published evidence on the relationship of LPS core types to capsular types or to virulence. Information about the distribution of 0 and K antigens has proved to be useful, for example in identifying and understanding enteropathogenic strains [5] and in suggesting novel therapeutic approaches to bacteremia [6]. Information about core types might be equally interesting.We therefore set out to produce a panel ofmouse monoclo-
Materials and MethodsBacteria and bacteriophages. E. coli RI (F470), R2 (F576), R3 (F653), R4 (2513) and KI2 (2131) and rough-specific phages FO, BrIO, C21, 6SR, and T4 were obtained from G. Schmidt (Forschunginstitut, Borstel, Germany). E. coli 0 I, 02, 04, 06, 012, 015, 018, and 075 were obtained from A. S. Cross (Walter Reed Army Institute, Washington, DC). Phages specific for E. coli capsular types KI (KIGS) and K5 (K5DG) were obtained from A. P. Roberts (Charing Cross and Westminster Medical School, London). Blood culture and urine isolates were obtained from the Clinical Bacteriology Laboratory, Department of Medical Microbiology, Edinburgh University, from routine clinical specimens. ...