Plasmids were constructed which contained a large portion of each of the four nonstructural genes of Sindbis virus fused to the N-terminal two-thirds of the trpE gene of Escherichia coli. The large quantity of fusion protein induced from cells containing these plasmids was subsequently used as an antigen to generate polyclonal antisera in rabbits. Each antiserum was specific for the corresponding nonstructural protein and allowed ready identification of each nonstructural protein and of precursors containing the sequences of two or more nonstructural proteins. These antisera were used to determine the stability of the mature nonstructural proteins and to examine the kinetics of processing of the nonstructural proteins from their respective precursors in vivo. Pulse-chase experiments showed that the precursor P123 is cleaved with a half-life of approximately 19 min to produce P12 and nsP3; P12 is then cleaved with a half-life of approximately 9 min to produce nsP1 and nsP2. Thus, although the rate of cleavage between nsP1 and nsP2 is faster than that between nsP2 and nsP3, the latter cleavage must occur first and is therefore the rate-limiting step. The rate at which P34 is chased suggests that the cleavage between nsP3 and nsP4 is the last to occur; however the regulation of nsP4 function in Sindbis virus-infected cells may be even more complex than was previously thought. The products nsP1 and nsP2 (and nsP4) are relatively stable; nsP3, however, is unstable, with a half-life of about 1 h, and appears to be modified to produce heterodisperse, higher-molecular-mass forms. In general, the processing schemes used by Sindbis virus and Semliki Forest virus appear very similar, the major difference being that most nsP3 in Sindbis virus results from termination at an opal condon, whereas in Semliki Forest virus cleavage of the P34 precursor is required.