The species Shigellajiexneri comprises a serologically heterogeneous group of dysentery bacilli whose 0 antigens are polysaccharide-lipid Apolypeptide-lipid B complexes strictly comparable in their gross structural and biological properties to the analogous antigens present in all gram-negative bacilli of the family Enterobacteriaceae (Simmons, 1971a). As shown in fig. 1, the lipopolysaccharide (LPS) component of these antigens comprises two distinct regions-a common basal structure shared by all S . JIexneri serotypes (except serotype 6) and 0-specific side-chains that determine the serological specificity and cross-reactivity of the whole antigen (Simmons, 1966). The common basal sugars 2-keto-3-deoxy-octonate, L-glycero-D-manno-heptose phosphate, D-glucose, D-galactose and Nacetyl-D-glucosamine are incorporated into the growing basal structure in that order and defects in the genes controlling this biosynthetic pathway result in an incomplete basal structure of rough (R) specificity (Johnston et al., 1967). The 0-specific side-chains (each consisting of six to eight repeating pentasaccharide units) may be further subdivided into two regions-a primary unbranched chain of L-rhamnose and N-acetyl-D-glucosamine identical to the variant Y antigen, and secondary side-chains of a-D-glucosyl and 0-acetyl residues which are substituted on the primary chain in a position that is unique for each serotype (Simmons, 1969). The specificity of the linkage of the glucose secondary side-chain is governed by specific UDP-glucose transferases that are phage-dependent and map near the lac locus (Manson et al., 1970). Loss of the phage attachment site in recombination experiments with Escherichia coli HfrC (lac+) strains usually leads to loss of the glucose secondary sidechains and, thus, to the production of lactosefermenting hybrids of variant Y-type specificity (Manson et al., 1970). However, hybrids with other specificities have also been described (Romanowska and Lachowicz, 1970 ;Katzenellenbogen et al., 1973;Lugowski et al., 1975).In an earlier review (Simmons, 197 la), structural and genetic aspects of the biosynthesis of the S. JIexneri 0 antigens were considered in detail. The conclusions drawn at that time about the structural changes involved in smooth to rough (S-+R) mutation, in the production of X and Y variants, and in the modification of type-specificity by lysogenic conversion remain valid today. However, some of the structures themselves require to be revised, largely as a result of methylation studies performed during the past decade by Lindberg and his colleagues (Lindberg et al., 1973;Kenne et al., 1977a and b ;Kenne et al., 1978). This review summarises the evidence for these currently accepted structures and indicates how these studies have elucidated the biology of the S. J-lexneri 0 antigens. Structure of the basal region of S.J-lexneri lipopolysaccharidesThe biosynthesis of the S. jiexneri LPS basal region proceeds from the 2-keto-3-deoxy-octonatelipid A core of chemotype Re by the sequential addition of...

Structural analysis of a representative group of Shigella flexneri lipopolysaccharides has shown that the 0-specific side chains of these polymers consist of a primary unbranched chain of N-acetylglucosamine and rhamnose substituted with secondary side chains of glucose. Sh. flexneri Y variants which lack the glucose secondary side chains and are thus structurally identical with the primary side chains of the different serotypes, belong to one of two types namely, Yl which consists of N-acetylglucosaminyl-( 1 +2)-rhamnosyl- ( 14) ) -rhamnose and a-glucosyl-(1 +2)-rhamnose respectively. The group antigen 3,4 in which rhamnose is the immunodominant sugar, seems to be related to the internal rhamnosyl-( 1+4)-rhamnose sequence. The serotype 6 antigen (VI) used in this study is so different structurally from its analogues that this organism is no longer regarded as a true Sh. flexneri species. From the structural relationship of the different 0-antigens, X and Y variants are the consequence of enzyme defects that interrupt the biosynthesis of the complete smooth lipopolysaccharides. Yl variants probably result from defects in the specific UDP-glucose transferases of serotypes l a and 2a while Y, variants probably result from defects in the analogous UDPglucose transferases of serotypes 4a, 3a, 5 and variant X. Variant X is an intermediate stage in the biosynthesis of serotypes 3a and 5 from the precursor Yz structure. Further evidence for the role of UDP-glucose transferases in conferring type specificity on the cryptic Y structures comes from two sources. Firstly, the type specific or T-loci of Petrovskaya that map near the lac locus are those in which glucose is the immunodominant sugar and loss of this locus by genetic recombination results in a glucose-deficient Y-type hybrid. Secondly, the structures proposed show that all phage conversions of Sh. flexneri reported to date, can be explained in terms of single enzyme changes involving specific UDP-glucose transferases. From the limited number of antigens studied, it can be predicted that the expression of type specific antigens will depend on the presence of group factors, as the former are always end group determinants which cannot be incorporated into the growing molecule before biosynthesis of the more central regions which determine group antigenicity.The Shigella flexneri are a serologically heterogeneous group of dysentery bacilli whose 0-antigens are lipid A-polysaccharide-lipid B-polypeptide complexes comparable in their gross structural and biological properties [I -41 to the analogous substances first isolated from Sh. shigae [5] and now known to be present in many other gram-negative bacilli of the family Enterobacteriaceae [6]. I n addition to their unequivocal importance in medicine, these 0-antigens are of special relevance to the biology of gram-negative bacilli in that comparative structural analysis with other genera will almost certainly reveal similarities and differences of the kind that will help t o elucidate the taxonomy, biosynthesis, gene...

1. Smooth to rough mutation has the same biochemical basis in Shigella as in Salmonella. It is the result of enzyme defects blocking the incorporation of the O-specific side chains that characterize the smooth lipopolysaccharide with the consequent exposure of the underlying basal structures that determine ;rough'-specificity. 2. The Shigella flexneri basal structure resembles its Salmonella analogue in that it has the same qualitative sugar composition, and enzyme defects in its biosynthetic pathway give rise to ;rough'-lipopolysaccharides that are indistinguishable from those of Salmonella chemotypes Ra, Rb, Rc and Rd. However, the Salmonella and Shigella basal structures are not identical as judged by quantitative analysis and the absence of serological cross-reaction. 3. The Sh. flexneri basal structure side chain has been isolated and characterized as an alpha-N-acetylglucosaminyl-(1-->4)-galactosyl-(1-->3)-glucose sequence with alpha-glucosyl radicals substituted on the 3- and 4-positions of the galactose and glucose respectively. The different sugar types in this side chain are incorporated into the growing molecule in the same order as in Salmonella, which explains why the enzyme defects associated with smooth to rough mutation produce the same series of R-chemotypes from both genera. The terminal alpha-glucosyl and alpha-N-acetylglucosaminyl-(1-->4)-galactosyl residues of the Sh. flexneri basal structure are sufficiently different from the terminal alpha-galactosyl and alpha-N-acetylglucosaminylglucosyl residues of the Salmonella analogue that they offer an explanation for the absence of serological cross-reaction between these two basal structures.

1. A series of oligosaccharides was isolated from Salmonella milwaukee lipopolysaccharide by partial acid hydrolysis. 2. Structural studies on these oligosaccharides indicated that the O-specific side chain of this lipopolysaccharide has a repeating pentasaccharide unit that is probably alpha-d-galactosyl-(1-->3)-beta-d-galactosyl- (1-->3)-N-acetylgalactosaminyl-(1-->3)-N-acetyl- d-glucosaminyl-(1-->4)-l-fucose. 3. Another oligosaccharide, which is not structurally related to the repeating pentasaccharide unit, has also been isolated and it is indistinguishable from an oligosaccharide obtained from Salmonella ;rough' (R) lipopolysaccharides. The isolation of this and similar core oligosaccharides from all chemotype VI lipopolysaccharides supports the view that Salmonella S-lipopolysaccharides have a common core that is probably identical with RII lipopolysaccharide.