The gene encoding fumarase (fum) from Thermus thermophilus was expressed in yeast Saccharomyces cerevisiae. The recombinant cells were heated at 70°C to inactivate indigenous enzymes and used for the bioconversion of fumaric acid to L-malic acid. By heating the host cells at 70°C, substrate is able to go across the heat-damaged membrane of the microorganism and a desired product can be formed. This new concept, called Synthetic Metabolic Engineering (SME), has already been applied successfully in Escherichia coli. Unfortunately, E. coli membrane is weakened too much during the heat treatment and enzyme leakage appears. The surface structure of yeast is more rigid than that of E. coli and this might be taken as an advantage for application of SME. When continuous or repeated-batch reaction is carried out, enzyme leakage becomes a major drawback. It is assumed that yeast cells could overcome this problem by retaining more enzymes in the cell during and after the heat treatment. In order to prove this hypothesis, a thermophilic fumarase (FUM) was over-expressed in two hosts, S. cerevisiae as well as E. coli. fum was first modified to be over-expressed in yeast cells and FUM was successfully produced in yeasts. Optimization of SME techniques was carried out for yeast cells. Then, enzyme activity and enzyme leakage was investigated for both strains. E. coli showed high level of FUM expression, though considerable amount of enzyme leaked to supernatant. On the other hand, even though the level of FUM expression in S. cerevisiae was low, yeast cells overcome leakage problem and are re-usable. This study showed the first trial of SME in yeast cells and possibility of utilization of yeast as a host strain for SME.