Guillain-Barré syndrome (GBS) is often associated with serum antibodies to glycoconjugates such as GM1 and GQ1b. The pathogenic role of these antibodies and other serum factors has not yet been clarified. We have investigated the effect of serum, plasma filtrate, and highly purified IgG and IgM from 10 patients with typical GBS on motor nerve terminals in the mouse hemidiaphragm. Quantal endplate currents were recorded by means of a perfused macro-patch-clamp electrode. The plasma filtrate of all GBS patients led to a 5- to 20-fold reduction of evoked quantal release within 7 to 15 minutes of continuous superfusion. In 4 patients, the amplitudes of single quanta were clearly reduced (by 10-66% of control values), indicating an additional postsynaptic action. Blocking effects could be reversed to a variable degree within 15 to 18 minutes after washout. Purified IgG was as effective as native serum, whereas a purified GBS IgM fraction did not block transmission. Sera from convalescent patients and IgG from healthy subjects were without blocking effect. The effects were complement independent and there was no link to the presence (in 6 patients) or absence (in 4 patients) of detectable antibodies to GM1 or GQ1b. In GBS, antibodies to an undetermined antigen depress the presynaptic transmitter release and, in some cases, the activation of postsynaptic channels. We suggest that weakness in the acute stage of GBS may be caused in part by circulating antibodies.