Fatty acid synthetase of chicken liver is rapidly and reversibly inactivated by 5,5'-dithiobis-(2-nitrobenzoic acid) (DTNB) at a rate (k2 = 132 mM-1 S-1 in 3 mM EDTA, 1% (v/v) glycerol, pH 7.0, at 25 degrees C) up to 2200 times higher than the reaction of this reagent with simple thiol compounds. The inactivation is caused by the reaction of the phosphopantetheine SH group, since it is protected competitively by either acetyl- or malonyl-CoA, and since the inactivated enzyme is unreactive with the phosphopantetheine label chloroacetyl-CoA but reactive with the cysteine reagent 1,3-dibromopropanone. Moreover, chloroacetyl-CoA prevents the modification of the rapidly reacting essential SH group by DTNB. The number of SH groups involved in inactivation was determined by correlating activity loss with the extent of reaction and by stopped-flow analysis of substrate (or chloroacetyl-CoA) protection. Values between 0.91 and 1.15 SH groups/dimer were obtained, indicating the presence of substoichiometric amounts of the prosthetic group in the fatty acid synthetase preparations used in this study. Inactivation of the synthetase by DTNB is strongly inhibited by increasing salt concentration and protected noncompetitively by NADP+ and NADPH. Treatment of the enzyme inactivated at low salt by salt, NADP+, or NADPH also effectively reduced cross-linking between enzyme subunits. The parallel effects of these treatments on the reaction with DTNB and subsequent dimerization are consistent with a minimum model of two discreet conformation states for fatty acid synthetase. In the low salt conformer, the phosphopantetheine and cysteine SH groups are juxtaposed, and the DTNB reaction (k2 approximately 132 mM-1 S-1) and dimerization are both facilitated. Transition to the high salt conformer by the above treatments is accompanied by an approximately 20-fold reduction of reactivity with DTNB (k2 = 6.8 mM-1 S-1) and reduced dimerization, due to spatial separation of the SH groups. During palmitate synthesis, the enzyme may oscillate between these conformation states to permit the reaction of intermediates at different active sites. Results obtained by studies on the effect of pH on DTNB inactivation implicate a pK of 5.9-6.1 for the essential SH group independent of salt concentration. This value is 1.5-1.8 pH units lower than the pK of 7.6-7.7 for CoA and may explain the 23-fold increase of the rate constant from a value of 0.3 mM-1 S-1 for CoA to that of the high salt conformer.