All of the OAg–ADH preparations were characterized by a sugar recovery greater than 80%, with more Lapatinib in vivo than 80% of OAg chains activated and < 2% (in moles) of free to linked hydrazide groups. We confirmed the absence of dimer and aggregate formation with the reaction condition used by analysing the OAg–ADH using HPLC-SEC. This showed the presence of one peak with the same kd value as the underivatised OAg. The OAgoxADH preparation was characterized by a total sugar recovery of 73%,

with 20% activation (molar % linked ADH to Rha) and < 2% free to linked hydrazide groups. In theory, the presence of more than one ADH linker per OAg chain in OAgoxADH could favour the OAg binding to the NHS-Sepharose. However the binding capacity was found to be 3.7 mg of OAgoxADH and 4.3 mg of OAg–ADH per ml of resin. The prepared affinity columns were tested with a commercially-available preparation of purified polyclonal

rabbit anti-Salmonella Typhimurium O:4,5 antibodies to determine if the hydrolysis and activation of OAg with ADH had impaired the antigenic structure of the OAg. 3.7 and 4.3 mg of OAgoxADH and OAg–ADH respectively were linked to NHS-Sepharose columns and 300 μl of O:4,5 antibodies (with an antibody concentration corresponding to 1666 ELISA units) were applied to each column. 92% of the antibodies bound to the OAg–ADH column ( Fig. 2A) and 96% bound to the OAgoxADH column ( Fig. 2B) with the remaining applied antibodies detected in the flow through and subsequent wash fractions. 89% and 90% of bound antibodies were eluted with 0.1 M glycine, Rapamycin 0.1 M NaCl pH 3 buffer from the OAg–ADH and OAgoxADH columns respectively. Following the previous result confirming the functional antigenic integrity

of both forms of derivatised OAg bound to NHS-Sepharose, we applied a protein preparation concentrated from human serum containing polyclonal anti-Salmonella antibodies to both columns. The proteins had been precipitated from human serum using ammonium sulphate in order to reduce the presence of contaminants that could interfere with the interactions between OAg on the columns and corresponding antibodies, and to concentrate the antibodies. 300 μl of resulting protein solution (with an antibody concentration corresponding to 1000 ELISA units) were applied to each 1 ml column. A high proportion (> 75%) of the antibodies applied to Acetophenone the columns in the serum protein solution bound to the column as shown by the low signal in the ELISA for OAg antibodies in the flow through and wash fractions ( Fig. 2C and D). For the OAg–ADH column ( Fig. 2C), elution with 0.1 M glycine, 0.1 M NaCl pH 3 and immediate neutralisation with 2 M Tris pH 9, resulted in a recovery of only 14% of the bound antibodies. For the OAgoxADH column, only 2% of the bound antibodies were eluted under the same conditions ( Fig. 2D). The same results were obtained whether the ELISA was performed coating the plates with purified OAg from S.