A number of oxygenated monoterpenes present at low concentrations in plant oils have anticarcinogenic properties. One of the most promising compounds in this respect is (؊)-perillyl alcohol. Since this natural product is present only at low levels in a few plant oils, an alternative, synthetic source is desirable. Screening of 1,800 bacterial strains showed that many alkane degraders were able to specifically hydroxylate L-limonene in the 7 position to produce enantiopure (؊)-perillyl alcohol. The oxygenase responsible for this was purified from the best-performing wild-type strain, Mycobacterium sp. strain HXN-1500. By using N-terminal sequence information, a 6.2-kb ApaI fragment was cloned, which encoded a cytochrome P450, a ferredoxin, and a ferredoxin reductase. The three genes were successfully coexpressed in Pseudomonas putida by using the broad-host-range vector pCom8, and the recombinant converted limonene to perillyl alcohol with a specific activity of 3 U/g (dry weight) of cells. The construct was subsequently used in a 2-liter bioreactor to produce perillyl alcohol on a scale of several grams.The production of (Ϫ)-perillyl alcohol from L-limonene is of interest because of the limited availability of (Ϫ)-perillyl alcohol in nature and its proven anticarcinogenic properties; phase II trials to evaluate perillyl alcohol for the treatment of breast, pancreatic, and colorectal cancer are in progress (37). The only microbial enzyme system described thus far that transforms limonene to perillyl alcohol was found in Bacillus stearothermophilus BR388. However, this enzyme system is not sufficiently regiospecific; significant quantities of carveol, carvone, and terpineol are also produced (25). Removal of these side products is difficult as their boiling points and hydrophobicities are almost identical, and expensive purification methods (for example, chromatography) would be required to obtain sufficiently pure perillyl alcohol. Therefore, the industrial production of perillyl alcohol with this Bacillus enzyme system is not attractive. The conversion of limonene to perillic acid by a Pseudomonas putida strain expressing a cymene monooxygenase was described by Mars et al. (21) and could be interesting as perillyl alcohol is likely to be an intermediate in the production of perillic acid. Other literature concerning limonene biotransformations was reviewed recently (5).The approach that we used to find strains capable of regiospecific hydroxylation of limonene consisted of screening a collection of 1,800 bacterial strains grown on a range of relatively reduced substrates, such as toluene, naphthalene, and various alkanes. Using this approach, we anticipated that we would find oxygenases involved in catabolic pathways that would accept L-limonene as a substrate. Previous work has demonstrated that many catabolic oxygenases accept a wide range of unnatural substrates. Toluene dioxygenases, for example, have been shown to oxygenate more than 100 substrates (reference 11 and references therein). Another example is ...