Clostridium thermocellum ATCC 27405 produces an extremely complicated multi-component cellulase aggregate (cellulosome) highly active on crystalline cellulose. From the cellulosome, two subunits, CelS (or Ss; M(r) = 82,000) and CelL (or SL, CipA; M(r) = 250,000), have been identified as essential for crystalline cellulose degradation [Wu et al. (1988) Biochemistry 27:1703]. We have determined the DNA sequence of the celS gene from four cloned DNA fragments encompassing this gene [Wang et al. (1993) J Bacteriol 175:1293]. To express the entire celS gene in Escherichia coli, the celS structural gene was amplified by the polymerase chain reaction (PCR) employing the PCR primers corresponding to sequences flanking the desired gene. This PCR product (2.1 x 10(3) bases; 2.1 kb) was cloned into an E. coli expression vector pRSET B. Subsequent expression of the cloned gene resulted in a fusion protein (rCelS; M(r) = 86,000) as inclusion bodies. The rCelS protein was recognized specifically by an anti-CelS antiserum in a Western blot analysis. The inclusion bodies were purified and solubilized in 5 M urea. The refolded rCelS produced very little reducing sugar from carboxymethylcellulose. However, it showed a higher activity on the crystalline cellulose (Avicel) and an even higher activity on phosphoric-acid-swollen Avicel. These results indicate that the CelS is an exoglucanase.