Mucus is fascinating stuff. Stringy yet slippery, permeable but tough. These properties also make it difficult to solubilise and give it a notable tendency to clog chromatographic columns and stay at the origin of electrophoretic gels. As a consequence, muchof the literature has been descriptive rather than molecular. This is changing rapidly. Nine genes have so far been identified which code for the protein cores of the mucins.'These have multiple tandem repeats which exhibit genetic polymorphisms, raising the possibility that they may account for the genetic components of diseases which affect mucosae. They have been shown to have tissue specificity, with MUC2 being the major secreted mucin in the normal colon2 but with a change to MUC5AC in colon cancer. 3 So far, no association between MUC gene products and benign mucosal disease has been demonstrated. It is the sugar coating, however, rather than the protein core whichgives the mucus itsremarkable properties and which accounts for about 60-80% of the mass of mucins. The structure of the mucin sugar (oligosaccharide) chains and the nature of the genes which determine them is proving an even harderproblem to crack.

Mucin oligosaccharide chains are almostentirely initiated by 0-linkage of N-acetylgalactosamine to serine or threonine in the protein core, incontrast to the oligosaccharides of most otherglycoproteins which are more commonly initiated by N-acetylglucosamine, N-linked to asparagine. The 0-linked structures have blood group antigenicity which immediately raises the possibility, or even probability, of genetic variation. They are also highly complex, with potential variations in their sequence, linkage, branching, and substitution by ester sulphation or 0-acetylation, givingthem a many-fold greater structuraldiversity than a similarly sized peptide. In intestinal mucosal diseases, such as ulcerative colitis or Crohn's disease, the hypothesis that there might be an underlying genetic abnormality inmucus seems par-ticularly attractive, 4 and there is direct evidence that the mucus layer is structurally altered, thin in ulcerative colitis and thick in Crohn's disease. 5 There is already evidence that the oligosaccharide chains in inflammatory bowel disease tend to be about half the normal length (of seven to 10 residues) and show increased expression of oncofetal carbohydrate antigens that are also found in colonic polyps and cancer, 6 and also demonstrate aberrant expression of blood group antigens. 7 These changes seem likely to be the result of alterations in the expression of specific glycosyl-and sulphotransferases. Most of the relevant enzymes have yet to be sequenced and cloned, and as 12 different N-acetylgalactosamine transferases responsible for the initiation of 0-linked mucin oligosaccharide chains have already been identified, the identification and sequencing of all the relevant enzymes is going to be a formidable task. Colonic mucins are most readily distinguished from mucins elsewhere in the intestine by their high degree of sulphation but, whereas the structure of the neutral mucin