Four sediment cores were collected from Lake Hartwell, SC, in 1987 and 1998 and analyzed for polychlorinated biphenyl (PCB) congeners. Total PCBs ranged from -0 to 58 microg/ g. Positive matrix factorization (PMF) was applied to the data sets to determine PCB source profiles. Two factors were determined for each data set. One factor resembled the original estimated PCB mixture of 80% Aroclor 1016 and 20% Aroclor 1254 and the other factor was a dechlorinated version of the mixture. Evidence of a dechlorination plateau is apparent from the PMF loading solutions because the dechlorinated congener profiles do not change from 1987 to 1998, butthe contribution to the profile from the dechlorinated factor increases from 73% (1987) to 87% (1998). PMF source contributions and plots of PCB concentration versus congener for individual samples provide evidence of enhanced dechlorination at high concentrations. After source apportionment an anaerobic dechlorination model was applied to the dechlorinated source profiles to quantify possible dechlorination pathways. It was found that dechlorination process M, extended to target biphenyl rings with up to six chlorines, provided the best fit for an individual process, and M + Q provides the best fit for combined processes, although M + LP also provides a similarfit. Process LP targets the higher chlorinated congeners and appears to dechlorinate PCBs in the sediments initially.

The enantioselectivity of microbial reductive dechlorination of chiral PCBs in sediments from Lake Hartwell, SC, was determined by microcosm studies and enantiomer-specific GC analysis. Sediments from two locations in the vicinity of the highest levels of PCB contamination were used as inocula. Dechlorination activity was monitored by concentration decreases in the spiked chiral PCBs and formation of dechlorination products using both achiral and chiral chromatography. Live microcosms spiked with PCB132 (234-236) exhibited dechlorination of PCB132 to PCB91 (236-24) and PCB51 (24-26). Meta dechlorination was the dominant mechanism. Microcosms spiked with PCB149 (245-236) exhibited preferential para dechlorination of PCB149 to PCB95 (236-25), followed by meta dechlorination to PCB53 (25-26) and subsequently PCB19 (26-2). Dechlorination of chiral PCB132 and PCB149 was not enantioselective. In Aroclor 1254-spiked microcosms, reductive dechlorination of PCB149 also was nonenantioselective. These results suggest that dechlorinating enzymes responsible for the dehalogenation of the chiral PCB132 and PCB149 congeners bind the two enantiomers equally. Reductive dechlorination of PCB91 and PCB95, however, occurred in an enantioselective manner, indicating that the dechlorinating enzymes for these PCBs are enantiomer-specific. The chlorine substitution pattern on the biphenyl ring appears to influence whether reductive dechlorination of chiral PCB congeners is enantioselective. Enantioselective PCB dechlorination by the microbial population of Lake Hartwell sediments occurs for select chiral PCBs; thus, certain chiral PCBs might be useful as markers for in situ reductive dechlorination.