INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Polychlorinated biphenyls (PCBs) are pollutants that persist in river, lake, and harbor sediments wherever they were manufactured or used. Concerns about their bioaccumulation and potential toxicity to humans and wildlife have spurred interest in novel approaches for remediation. Industrial PCB mixtures, such as Aroclors, provide a daunting challenge for microbial dechlorination and degradation, because they each contain 60 to 90 different molecular species known as congeners. Each congener consists of a biphenyl skeleton substituted with 1 to 10 chlorines (Fig. 1). Fortunately, microorganisms in many PCB-contaminated sediments already have the ability to dechlorinate many of the PCB congeners (10, 12-14, 22, 24, 29, 36; see reference 3 for a review). Microbial dechlorination can play an important role in natural restoration because it decreases the toxicity of PCBs and increases their degradability (5, 25, 26). Accordingly, one goal of our research is to discover innovative approaches for enhancing or accelerating microbial PCB dehalogenation in situ.

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Structure of halogenated biphenyls, showing the numbering scheme (A) and the dechlorination of a major heptachlorobiphenyl by Process N, Processes N plus LP, and Process P (B). Each chlorinated or brominated biphenyl congener can have from 1 to 10 halogens (chlorines or bromines) at the positions shown in panel A. The halogens at positions 2 and 6 on either ring are designated ortho, those at carbons 3 and 5 are designated meta, and those at carbon 4 are designated para. Panel B shows one of the major components of Aroclor 1260 and the terminal dechlorination products produced from this congener by Process N, Processes N plus LP, and Process P. The most extensive chlorine removal results from sequential dechlorination by Processes N and LP (27).

The upper Housatonic River, including Woods Pond, is contaminated with partially dechlorinated Aroclor 1260 (2, 4), and previous studies have indicated that its sediments harbor several discrete populations of PCB-dechlorinating microorganisms (1, 8, 32, 37). Despite rigorous attempts, none of these microorganisms has been isolated or obtained in a sediment-free culture (38). Hence, we are limited to describing these populations in terms of their PCB dechlorination activity. Three major PCB dechlorination activities, known as Dechlorination Processes N, P, and LP, have been observed in Woods Pond sediment (1, 8, 32, 37). (A microbial dechlorination process is a set or series of dechlorination reactions that determines the substrate range, the specific chlorines targeted, and the sequence of dechlorination [3].) Several lines of evidence indicate that three discrete microbial populations are responsible for Dechlorination Processes N, P, and LP, as discussed below.

(i) The dehalogenation specificities of these three dechlorination processes are distinctly different. Process N primarily removes flanked meta chlorines from chlorophenyl groups with the following substitution patterns: 2,3,4- (234-), 236-, 245-, 2345-, 2346-, and 23456- (3) (chlorines that are removed are underlined). Process P removes flanked para chlorines from 234-, 245-, and 2345-chlorophenyl groups (1), and Process LP primarily removes the unflanked para chlorines from 24- and 246-chlorophenyl groups (8). Figure 1B shows a heptachlorobiphenyl that is a major component of Aroclor 1260 and the terminal dechlorination products produced from it by Processes N, P, and the combination of N and LP.

(ii) Dechlorination Processes N, P, and LP have different temperature ranges. Process N occurs at 8 to 30°C, Process P occurs at 12 to 34°C, and Process LP occurs at 18 to 30°C (37).

(iii) These three dechlorination processes can be activated separately by specific primers. Process N is primed by 236-CB, 2346-CB, and 23456-CB (see below for an explanation of chemical abbreviations) (32, 37); Process P is primed by 25-34-CB, 24-34-CB, and 245-CB (1, 32); and Process LP is primed by 246-CB (8, 37).

We hypothesized that PCB primers stimulate and support the growth of PCB dechlorinators, perhaps by serving as electron acceptors. This hypothesis is strongly supported by the recent finding that adding a high concentration of PCBs to microcosms of anaerobic sediment resulted in a nearly 200-fold transient increase in the number of PCB dechlorinators concurrent with the dechlorination of the PCBs (21).

Morris and colleagues demonstrated that microorganisms from two PCB-contaminated sediments with no history of brominated biphenyl contamination could partially dehalogenate hexabrominated biphenyls, whereas microorganisms from a sediment that was not contaminated with halogenated biphenyls could not (23). Based on their findings, these authors proposed that PCB-dechlorinating microorganisms might be able to dehalogenate brominated biphenyls. Likewise, we recently found that the anaerobic microorganisms in Woods Pond sediment can dehalogenate many mono-, di-, and tribrominated biphenyls even though the upper Housatonic River has no history of contamination with brominated biphenyls (6). Our data suggested that PCB-dehalogenating microorganisms in Woods Pond might exhibit relaxed specificity for brominated biphenyl dehalogenation. Since many of the brominated biphenyls are completely dehalogenated to biphenyl in a relatively short period of time (6), we reasoned that they might be good candidates for priming PCB dechlorination.

In this paper we report the results obtained when various mono-, di-, and tribrominated biphenyls were tested for the ability to prime microbial dechlorination of the partially dechlorinated Aroclor 1260 in Woods Pond sediment. Most of the brominated biphenyls primed Process N or Process P dechlorination. Our results demonstrate that compounds other than PCBs can activate extensive PCB dechlorination in sediment.

(A preliminary report of these findings was presented at a conference on microbial dehalogenation held in 1992 [7].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. In this paper each PCB and brominated biphenyl congener is identified by listing the substituted position(s) on each ring, separated by a hyphen, and followed by the designation -CB (chlorinated biphenyl) or -BB (brominated biphenyl) (Fig. 1). For example, 2,6-dibromobiphenyl is abbreviated 26-BB, and 2,3',4',5-tetrachlorobiphenyl is abbreviated 25-34-CB. All commercially available mono-, di-, and tribrominated biphenyls (97 to 99% pure as determined by gas chromatography [GC] with a flame ionization detector [FID]) were purchased from Accu-Standard (New Haven, Conn.) or from Ultra Scientific (North Kingstown, R.I.). Chemists at GE Corporate Research and Development synthesized highly purified 26-BB (99.93% pure as determined by GC-FID) for use in our quantitative experiments. L-Malic acid (cell culture reagent quality; catalog no. M-7387) was purchased from Sigma Chemical Co. (St. Louis, Mo.), and the pH was adjusted to 7 with sodium hydroxide.

Slurry preparation and incubation. Multiple core samples (depth, 45 cm) of sediment were collected from the west side of Woods Pond (Lenox, Mass.) (2), a shallow impoundment on the Housatonic River. The core samples were pooled in glass jars, filled to the top with site water, and stored at 4°C until they were used. On a dry weight basis, each gram of sediment contained 30 µg of partially dechlorinated Aroclor 1260 and approximately 7,000 µg of weathered hydrocarbon oil. Slurries were prepared under an atmosphere of 95 to 97% nitrogen and 3 to 5% hydrogen in an anaerobic chamber by mixing wet sediment (2 volumes) with glass-distilled water (3 volumes). Bromobiphenyl congeners were added from concentrated stock solutions prepared in acetone. The final concentration of acetone was 0.5% (vol/vol). Unless indicated otherwise, the final concentration of each brominated biphenyl was 350 µmol per liter. This concentration is well above the aqueous solubility of brominated biphenyls; hence, these compounds likely partition into the organic material in the sediment, including the excess oil. Therefore, when referring to halogenated biphenyls, we use the term micromolar (µM) to mean micromoles per liter of slurry. In most cases, disodium malate was also added to a final concentration of 10 mM because previous experiments showed that malate accelerated dehalogenation (6). No other nutrients were added. The serum bottles were crimp sealed with Teflon-lined butyl rubber septa. Sterile controls were prepared by pasteurization (75°C, 20 min), followed by incubation (23 to 25°C, 24 h) and autoclaving (121°C, 3 h). Live controls were incubated with malate and acetone at final concentrations of 10 mM and 0.5%, respectively. Experimental samples and controls were incubated in the dark at 23 to 25°C. All live incubations became methanogenic within 1 to 2 weeks. The quantitative data shown for 26-BB (see Fig. 2 through 4) were derived from triplicate samples. All other congeners were tested in duplicate.

Extraction and analysis of PCBs and brominated biphenyls. (i) Extraction. Aliquots (1 ml) of the 30-ml slurries were sampled periodically and extracted in vials with Teflon-lined foam-backed screw caps by vigorous shaking (24 h) with anhydrous diethyl ether (5 volumes). We used elemental mercury (0.25 volume) or acid-reduced copper filings (~0.3 g) to remove sulfur (17). Quantitative comparisons of the mole percent (mol%) congener distributions for samples extracted by this simple procedure and by a rigorous Soxhlet procedure (16) demonstrated no significant differences (28); hence, we routinely used the simple one-step extraction procedure.

(ii) GC analysis. Dehalogenation of the brominated biphenyls was monitored by GC with a mass spectrometer operated in the selected ion mode as previously described (6). Debromination products were identified as described previously (6). PCB dechlorination was monitored by high-resolution capillary GC analysis with a Ni⁶³ electron capture detector (ECD). We used a Hewlett-Packard model 5890 GC-ECD operated in splitless mode and equipped with a DB-1 (polydimethylsiloxane) capillary column (length, 30 m; inside diameter, 0.25 mm; phase thickness, 0.25 µm; J & W Scientific, Inc., Folsom, Calif.). The injector and detector temperatures were 250 and 300°C, respectively. We used the following multistage temperature program. The initial temperature was 40°C; the temperature was increased at a rate of 20°C/min to 160°C, kept at 160°C for 3 min, increased at a rate of 2°C/min to 200°C and then at a rate of 8°C/min to 260°C, and kept at 260°C for 15 min. We used helium as the carrier gas at a column flow rate of 1.5 ml/min and nitrogen as the makeup gas at a flow rate of 6 ml/min. The septum purge flow rate was 3 ml/min, and the splitless vent flow rate was 57 ml/min. We are able to resolve 118 PCB peaks using these conditions (19).

(iii) Quantitative congener-specific PCB analysis. Previous GC-mass spectrometry analyses of the partially dechlorinated PCBs in Woods Pond sediment (2) and of the extensively dechlorinated PCBs resulting from priming with 26-BB (28) showed that many of the PCB congeners produced by dechlorination either are not present in Aroclors or are present only in minor quantities. Furthermore, some of these PCB dechlorination products coelute with less chlorinated PCB congeners that are more prominent in the Aroclors. Hence, it was apparent that we would not be able to accurately quantify extensively dechlorinated PCBs by using any single Aroclor or any mixture of Aroclors as a standard. We therefore developed a quantitative standard that included known amounts of Aroclor 1260 and of 43 additional PCB congeners that were previously identified as dechlorination products of Aroclor 1260 in Woods Pond sediment (2, 28). In addition, several mono-, di-, and trichlorobiphenyls were included because they are frequently observed as PCB dechlorination products in other systems (12, 14, 24). PCB congener assignments were made as previously described (19) and reported for Woods Pond sediment (1, 32). The concentrations of the individual components of Aroclor 1260 in our standard were calculated from previously determined weight percent distributions for the congeners in Aroclor 1260 (19, 28). Our customized standard permits quantitation of 84 PCB peaks, including all significant peaks detected in our samples (28).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Quantitation of the PCB dechlorination primed by 26-BB. Dehalogenation of 26-BB to 2-BB and then biphenyl began within the first week of incubation and was essentially complete within 60 days (6). Most (~70 mol%) of the PCBs in Woods Pond sediment have six or more chlorines; hence, the mole percent fraction of hexa- through nonachlorobiphenyls provides a very effective way to assess the extent of dechlorination of the Aroclor 1260 residue. No significant PCB dechlorination occurred in either live (Fig. 2A) or autoclaved controls, but PCB congeners with six or more chlorines were rapidly dechlorinated in samples primed with 26-BB (450 µM) (Fig. 2A). The hexachlorobiphenyls decreased most rapidly, followed by the hepta- and then the octachlorobiphenyls (Fig. 2B). The pentachlorobiphenyls and the trichlorobiphenyls increased slightly. However, the primary products were tetrachlorobiphenyls, which increased to 60 mol% of the total PCBs (Fig. 2B). Most of the dechlorination occurred in the first 2 months, but dechlorination continued at a slower rate throughout the rest of the incubation period even though the 26-BB had been depleted. After 179 days, the dechlorination had decreased the fraction of PCBs having six or more chlorines by 81%.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Time course of dechlorination of Aroclor 1260 primed by 26-BB (450 µM) in Woods Pond sediment. (A) Change in the total hexa- through nonachlorobiphenyls (Hexa-Nona-CBs) as a function of time. Symbols: , live controls; , 26-BB-primed samples. (B) Change in PCB homolog distribution in 26-BB-primed samples as a function of time. Symbols: , nonachlorobiphenyls; , octachlorobiphenyls; , heptachlorobiphenyls; , hexachlorobiphenyls; , pentachlorobiphenyls; , tetrachlorobiphenyls; , trichlorobiphenyls. The data shown are averages for triplicate samples ± standard deviations (shown as error bars); if no bar is evident, the deviation was smaller than the size of the symbol.

Specificity of the PCB dechlorination primed by 26-BB. A comparison of the PCB congener distribution profiles in the live controls and in samples primed with 26-BB revealed the full impact of the dechlorination (Fig. 3). Many of the major hexa- and heptachlorobiphenyls were almost completely removed, and there were corresponding large increases in four tetrachlorobiphenyls, 24-24-CB, 24-26-CB, 24-25-CB, and 25-26-CB, and smaller increases in 26-4-CB, 24-4-CB, and 26-26-CB. There were also decreases in the amounts of certain pentachlorobiphenyls (e.g., 236-25-CB, 245-24-CB, and 236-34-CB), but several other pentachlorobiphenyl peaks increased, e.g., 246-24-CB (Fig. 4A).

View larger version (29K): [in this window] [in a new window]   FIG. 3.   PCB congener distribution of the Aroclor 1260 residue in Woods Pond sediment after 100 days in live controls (A) and in samples primed with 26-BB (450 µM) (B). The PCB congener designations indicate the positions of the chlorine atoms on each phenyl ring, and the hyphen represents separation of the rings. The data shown are averages for triplicate samples.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   Absolute differences in the PCB congener distributions of the Aroclor 1260 residues (controls minus experimental samples) as a result of priming with 26-BB (450 µM) (A) and priming with 4-4-BB (350 µM) (B). The PCB congener designations indicate the positions of the chlorine atoms on each phenyl ring, and the hyphen represents separation of the rings. The data shown are averages for duplicate samples primed with 4-4-BB (obtained at 120 days) and for triplicate samples primed with 26-BB (obtained at 100 days). The scale is the same for both panels.

PCB dechlorination primed by 26-BB: reproducibility and concentration effect. We repeated our experiments with 26-BB many times and determined that the results were highly reproducible. We also tested the effect of 26-BB at concentrations ranging from 50 to 1,000 µM (Fig. 5). The concentration of 26-BB affected the rate and final extent of PCB dechlorination; the optimal concentration was 200 to 500 µM. Higher concentrations did not significantly increase dechlorination.

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Effect of the concentration of 26-BB used as primer on the time course and extent of dechlorination. The samples contained 26-BB at the following concentrations: 0 µM (live control) (), 50 µM (), 100 µM (), 200 µM (), 500 µM (), and 1,000 µM (). No malate was added. The data shown are means for duplicate samples. Hexa- through Nona-CBs, hexa-through nonachlorobiphenyls.

Specificity of the PCB dechlorination primed by 4-4-BB. Dehalogenation of 4-4-BB to 4-BB and then biphenyl was first observed at 2 weeks and was essentially complete in 2 months (6). Figure 4 compares the PCB substrates and products of the dechlorination processes primed by 4-4-BB and 26-BB. This figure shows that the dechlorination primed by 4-4-BB had a narrower specificity and generated a different set of products than the dechlorination primed by 26-BB. Most of the major hexa- and heptachlorobiphenyls were substrates for the dechlorination primed by 4-4-BB, but they were not depleted as extensively as they were in the dechlorination primed by 26-BB. Many of the less prominent hexa- and heptachlorobiphenyls were not substrates for the dechlorination primed by 4-4-BB. The main products were 25-25-CB and 235-25-CB (Fig. 4B). Significant amounts of 23-25-CB, 24-25-CB, 25-3-CB, 24-3-CB, and 235-236-CB were also produced. A comparison of the substrates and products showed that this dechlorination process removed the flanked para chlorines on 34-, 234-, 245-, and 2345-chlorophenyl groups and converted them to 3-, 23-, 25-, and 235-chlorophenyl groups, respectively. This dechlorination matches Dechlorination Process P, which was previously primed in Woods Pond sediment by adding 25-34-CB (1).

Relative effectiveness of chlorinated and brominated biphenyl congeners for priming Dechlorination Processes P and N. Dechlorination Processes P and N can also be primed with certain PCB congeners, and 25-34-CB and 23456-CB are among the most effective (1, 32). We were not able to evaluate the brominated analogs of these PCB congeners because they are not commercially available. Furthermore, most of the chlorinated analogs of the brominated biphenyl congeners that are available, including 26-CB and 4-4-CB, are not dechlorinated in this sediment and do not prime PCB dechlorination (6, 32). Therefore, in order to evaluate the relative effectiveness of chlorinated and brominated biphenyls as primers, we compared the extent of PCB dechlorination primed by 25-34-CB and 23456-CB with that primed by 4-4-BB and 26-BB. We compared the extent of dechlorination primed by 25-34-CB with the extent of dechlorination primed by 4-4-BB by examining chromatograms for various time points in several different experiments performed with each congener. The dechlorination process primed by 25-34-CB first began at ~42 days and progressed steadily until 84 days. The data obtained at 111, 140, and 220 days reveal that the dechlorination progressed very little after 84 days. In contrast, the dechlorination primed by 4-4-BB was more extensive at 43 days than the dechlorination primed by 25-34-CB at 84, 140, or 220 days. The results were highly reproducible, both within and between experiments (for example, note the small variations in the data from duplicate incubations [1] [Tables 1 and 2]).

PCB dechlorination primed by other brominated biphenyls. Most of the other brominated biphenyls also primed PCB dechlorination (Table 3). The brominated biphenyls were completely dehalogenated to biphenyl by the pathways shown (6). Generally, meta and para bromines were removed before ortho bromines. In contrast, the chlorinated counterparts of nine of these congeners (2-CB, 3-CB, 4-CB, 24-CB, 25-CB, 26-CB, 2-2-CB, 4-4-CB, and 25-3-CB) were not dechlorinated and did not prime PCB dechlorination (summarized in reference 6).

Impact of primed dechlorination on PCB persistence in humans. Almost one-half of the total PCB content of Aroclor 1260 consists of congeners that have been reported to be highly persistent in humans; i.e., they have estimated half-lives in humans of more than 10 years (2, 11). The dechlorination that has occurred in situ at Woods Pond over the last few decades has reduced this fraction of PCBs to 32.2 mol%, but the dechlorination primed by 26-BB further reduced the fraction of highly persistent PCBs in the sediment to 7.7 mol% after 100 days and to 5.8 mol% after 179 days. These values correspond to impressive decreases of 76 and 82%, respectively. Likewise, the fraction of PCBs reported to be moderately persistent in humans (i.e., those with half-lives of 1 to 10 years in humans) decreased from 14.3 to 5.5 mol% in 100 days. Thus, the dechlorination primed by 26-BB markedly reduced the proportion of the PCBs that are most persistent in humans. After the primed dechlorination (100 days), 85 mol% of the residual PCBs had half-lives in humans of less than 1 year.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Correlation between the halogen configuration of the brominated biphenyl primer and the specificity of the PCB dechlorination primed. Our studies of PCB dechlorination in Woods Pond sediment demonstrated that the halogen configuration of a PCB congener determines whether the congener will be dechlorinated and which chlorine(s) will be removed (32). Some chlorophenyl rings (e.g., 2-, 3-, 4-, 23-, 25-, and 26-chlorophenyl rings) were not dehalogenated. At a given temperature, the halogen configuration also determines which PCB congeners will prime PCB dechlorination and which dechlorination process(es) they will prime (1, 8, 32, 37). Only PCBs substituted at carbons 2, 3, and at least one other site prime Process N (32), and congeners with 34- or 245-chlorophenyl rings prime Process P (1, 32). Only one congener, 246-CB, primes Process LP (8, 37).

Evidence that the microorganisms that dehalogenate PCBs also dehalogenate brominated biphenyls. Previously, we proposed that PCB-dechlorinating microorganisms in Woods Pond might also dehalogenate brominated biphenyls, albeit with relaxed specificity (6). The following four lines of evidence from the experiments reported here support this hypothesis. (i) The brominated biphenyls primed the same highly specific PCB dechlorination processes (N and P) that are primed by PCB congeners. (ii) The primed PCB dechlorination began shortly after the onset of brominated biphenyl dehalogenation. (iii) The specificity of the PCB dechlorination process that was activated correlated loosely with the halogen substitution on the brominated biphenyl primer (Table 3). (iv) The concentration of the brominated biphenyl primer affected the rate and extent of PCB dechlorination (Fig. 5). However, definitive proof that the same microorganisms dehalogenate PCBs and brominated biphenyls will require isolation of the microorganisms responsible.

Proposed explanations for the priming activity of halogenated biphenyls. We proposed previously that high concentrations of halogenated biphenyls prime PCB dechlorination because they are good substrates for dehalogenases and support the growth of PCB dechlorinators, perhaps by acting as electron acceptors (1, 3, 7, 8). Recently, Wu demonstrated that the concentrations of PCB-dechlorinating microorganisms increased nearly 1,000-fold after two additions of 26-BB (350 and 700 µmol per liter of slurry) (34). Proof that the halogenated biphenyls serve as electron acceptors will require experiments that demonstrate that the dehalogenation of these compounds is coupled to the oxidation of a specific electron donor(s), and the identification of appropriate electron donors will probably require sediment-free enrichment cultures. However, we propose that the following interpretation of our data can serve as a working hypothesis to guide research until pure or highly enriched cultures of the PCB-dechlorinating microorganisms in Woods Pond become available. (i) The first step in the dehalogenation of a brominated biphenyl primes PCB dechlorination and determines which dechlorination process (and consequently which of two different dehalogenating populations) will be activated. (ii) One population of microorganisms is predominantly meta dechlorinating. This population initiates ortho and meta dehalogenation of brominated biphenyls and Process N dechlorination of PCBs. Once it has been activated, this population can also remove para bromines, perhaps because the specificity for debromination is not very stringent. This is consistent with the observation that the congeners produced as intermediates do not accumulate significantly before dehalogenation. (iii) A second population is predominantly para dechlorinating. This population initiates para dehalogenation of congeners that have unflanked para bromines and Process P dechlorination of PCBs. Once it has been activated, this population can also remove ortho bromines (e.g., the 2-BB generated from dehalogenation of 24-BB).

Implications for bioremediation. Process N converts many of the PCBs in Woods Pond sediment to congeners containing 24- and 246-chlorophenyl groups. Process LP primarily targets the unflanked para chlorines on these two chlorophenyl groups (8, 37, 38). Consequently, when Process LP follows Process N, the combined action of these complementary dechlorination processes converts many of the hexa- through nonachlorobiphenyls in Aroclor 1260 to di- and trichlorobiphenyls (e.g., Fig. 1B) (8, 27, 37). None of the brominated biphenyls primed Process LP, but other halogenated compounds may be able to do so (27).