Genome-Resolved Metagenomics and Metatranscriptomics Reveal Insights into the Ecology and Metabolism of Anaerobic Microbial Communities in PCB-Contaminated Sediments

Growth of organohalide-respiring bacteria such as Dehalococcoides mccartyi on halogenated organics (e.g., polychlorinated biphenyls (PCBs)) at contaminated sites or in enrichment culture requires interaction and support from other microbial community members. To evaluate naturally occurring interactions between Dehalococcoides and key supporting microorganisms (e.g., production of H2, acetate, and corrinoids) in PCB-contaminated sediments, metagenomic and metatranscriptomic sequencing was conducted on DNA and RNA extracted from sediment microcosms, showing evidence of both Dehalococcoides growth and PCB dechlorination. Using a genome-resolved approach, 160 metagenome-assembled genomes (MAGs), including three Dehalococcoides MAGs, were recovered. A novel reductive dehalogenase gene, distantly related to the chlorophenol dehalogenase gene cprA (pairwise amino acid identity: 23.75%), was significantly expressed. Using MAG gene expression data, 112 MAGs were assigned functional roles (e.g., corrinoid producers, acetate/H2 producers, etc.). A network coexpression analysis of all 160 MAGs revealed correlations between 39 MAGs and the Dehalococcoides MAGs. The network analysis also showed that MAGs assigned with functional roles that support Dehalococcoides growth (e.g., corrinoid assembly, and production of intermediates required for corrinoid synthesis) displayed significant coexpression correlations with Dehalococcoides MAGs. This work demonstrates the power of genome-resolved metagenomic and metatranscriptomic analyses, which unify taxonomy and function, in investigating the ecology of dehalogenating microbial communities.


■ INTRODUCTION
Organohalide-respiring bacteria (OHRB) "breathe" halogenated organics to obtain energy for growth.As organohalides are often undesirable groundwater, sediments, or soil contaminants, bioremediation strategies (e.g., biostimulation, bioaugmentation, and natural attenuation) that take advantage of the unique OHRB lifestyle are popular and feasible for achieving contaminated site cleanup, especially for the chlorinated ethenes. 1−5 However, OHRB-focused bioremediation strategies for polychlorinated biphenyls (PCBs) lag behind those for chlorinated ethenes. 6,7n OHRB genus of particular interest and relevance, Dehalococcoides, contains members that can grow with PCBs as an electron acceptor, hydrogen as an electron donor, and acetate as a carbon source. 7,8The key enzymes for organohalide respiration by Dehalococcoides are reductive dehalogenases (RDases) which contain a corrinoid cofactor involved in catalysis. 9−14 Because Dehalococcoides lead specialized lifestyles, their genomes are streamlined and lack many complete pathways to produce required substrates and cofactors (e.g., corrinoids) and metabolize harmful intermediates (e.g., carbon monoxide (CO)).−18 Sequencing technology advances have facilitated studies of interactions between Dehalococcoides and other microbes in PCB-contaminated marine sediments and enrichment cultures. 3,12,14,19,20Further deciphering the interactions between Dehalococcoides and indigenous microorganisms is important to improve and optimize halogenated pollutant bioremediation strategies, such as maintaining growth and reductive dechlorination activity of Dehalococcoides at contaminated sites.
In previous studies, we found 16S rRNA and reductive dehalogenase genes from Dehalococcoides in PCB-contaminated wastewater lagoon sediments showing dechlorinating activity without any historical report of PCE contamination. 21PCB dechlorination with concomitant growth of Dehalococcoides occurred in microcosms incubated under a natural attenuation scenario using sediments from the site. 22Nucleic acids (DNA and RNA) were extracted from a different set of less enriched site sediment microcosms with different PCB concentrations and subjected to metagenomic and metatranscriptomic sequencing. 23he purpose of this study was to recover additional metagenome-assembled genomes (MAGs), including Dehalococcoides, and perform a genome-resolved gene expression analysis of the sediment microbial community.We also aimed to assign functional roles to MAGs and perform genome-resolved network analysis to investigate relationships between Dehalococcoides and key members of the sediment microbial community.A genome-resolved metatranscriptomics approach could pave the way for future ecological analysis of dehalogenating microbial communities that reveals relationships between Dehalococcoides and the indigenous microbial community members and provides insights that could promote increased dehalogenation rates.
A mixture of organic acids (acetate, propionate, butyrate, and lactate) was added into the bottles with an initial concentration of 2.5 mM each. 25 Site location, sediment sampling protocols, and microcosm preparation details exist elsewhere. 21,22,25roclor 1248 is the likely contaminating PCB mixture. 21No exogenous PCB congeners or microorganisms were added to the microcosms.The two sediment samples had different PCB concentrations.The high-PCB microcosms (HPCBM) contained an average of 28.04 ± 2.89 μg/mL PCBs, while the low-PCB microcosms (LPCBM) contained an average of 4.28 ± 1.05 μg/mL (p < 0.0001; Figure S1A). 23,26fter 200 days of incubation, there were no significant changes in total PCB concentrations in either the HPCBM or the LPCBM (p > 0.05; Figure S1B).This is expected, as possible aerobic PCB hydroxylation activity was minimal in the anaerobic microcosms.
PCB Extraction and Quantification.A liquid−liquid PCB extraction method was applied to microcosm slurry samples (2 mL), followed by addition of surrogate and internal standards as described previously. 22,25Each extracted batch of samples contained at least one laboratory blank comprised of hexane spiked with surrogate standards.PCBs in slurry extracts were quantified with a modified US EPA method 1668C. 27,28rrogate standard recoveries and method blanks allowed us to evaluate extraction procedure efficiencies and quantify background PCB concentrations. 29The PCB congener data set is deposited in Iowa Research Online (DOI: 10.25820/ data.006156). 26Statistical differences among PCB congener data were analyzed with an independent two-sided t-test with α = 0.05.
DNA and RNA Extractions, High-Throughput Sequencing, and qPCR.DNA was extracted periodically during the experiment for the qPCR analysis of Dehalococcoides 16S rRNA gene abundance.DNA and RNA were extracted for highthroughput sequencing at day 200 of the experiment.Extraction and high-throughput sequencing of DNA and RNA from microcosms and qPCR analysis followed procedures described previously, 22,23 with more detail in Section S1.
The 220 MAGs obtained from individual assembly along with the 62 coassembled MAGs were dereplicated to 158 nonredundant MAGs with dRep (version 3.4.2) using 95% and 99% average nucleotide identity (ANI) for primary and secondary clustering, respectively. 39The two dereplicated individually assembled MAGs classified as D. mccartyi were included in subsequent analysis along with the coassembled D. mccartyi MAG.Additional details about taxonomic classification, phylogeny, reads per kilobase per million mapped reads (RPKM) and covered fraction calculations, and MAG functional annotation are in Section S2.
A Dehalococcoides pangenomic analysis, including the Dehalococcoides MAGs recovered here, was conducted with the Anvi'o pangenome workflow 40 as described in Section S2.
Comparisons between reductive dehalogenase (RDase) genes in Dehalococcoides MAGs and genomes were performed and an

Environmental Science & Technology
RDase amino acid phylogenetic tree was constructed as described in Section S3.
Metatranscriptomics Analysis and Coexpression Network.Nucleotide sequences from all MAG coding regions were concatenated and used to build an index with kallisto (version 0.46.1). 41Prokka-annotated 42 rRNA and tRNA genes in MAGs were included.Trimmed metatranscriptomic reads were mapped to the index to quantify estimated counts and effective length of each gene using the quant mode of kallisto (version 0.46.1). 41Differential expression (DE) analysis between samples with high (treatment) and low (control) PCB concentrations was performed with the estimated counts using package limma (version 3.54.2) 43 in R (version 4.2.2) 44 as it performed well under different conditions. 45n previous work, 3,19,20,46−48 co-occurrence networks were constructed using 16S amplicons to describe the relative abundance of microorganisms in the system (e.g., ASVs/ OTUs).For metatranscriptomic data, the transcripts per million (TPM) value was used as an analogous representation for the overall gene expression of a MAG.TPM values were calculated for each of the 160 MAGs and used to build the coexpression network.To calculate the TPM of MAGs, estimated transcript counts and effective gene length mapped to each coding gene of an MAG were summed, while the mapping results for noncoding

Environmental Science & Technology
RNAs (i.e., rRNA and tRNA) were excluded.The TPM of a MAG was calculated as follows where q i is the sum of estimated counts of each gene from an MAG, l i is the sum of the effective length of each gene in an MAG.Estimated counts and effective length of each gene were kallisto outputs.∑ j (q j /l j ) corresponds to the sum of q i /l i for each MAG.Spearman's rank correlation coefficients for TPM values of enriched genes (except rRNA and tRNA) in HPCBM from DE analysis and the TPM values of the MAGs were computed with the R package Hmisc (version 5.0−1). 49The p values were adjusted with the "BH" method.The coexpression network was constructed using the R package igraph (version 1.4.1). 50MAGs with correlation coefficients >0.8 and an adjusted p-value <0.05 to other MAGs were considered significantly correlated in the analysis.Functional roles were not considered when constructing the network.The network was then visualized in Gephi (version 0.10) with a force atlas layout. 51RESULTS AND DISCUSSION

PCB Congener Profile Changes in Microcosms.
Comparing PCB congener profiles at day 0 and at final sampling day 200 revealed greater differences in the HPCBM congener profile after 200 days 25 than in the LPCBM, where few congeners diverged from the initial profile (Figure 1).In a previous microcosm study with these sediments over a 430-day incubation period, the mass fractions of tetrachlorinated PCB 66 and PCB 61/70/74/66 decreased, while mass fractions of trichlorinated dechlorination products PCB25 and PCB26/29 increased. 22Here, we observed significant decreases in PCB 66

Environmental Science & Technology
(p = 0.0182) and PCB 61/70/74/66 (p = 0.0235) in the HPCBM but not in the LPCBM (Figures 1 and S2).Although mass fractions of expected dechlorination products PCB25 and PCB26/29 increased in the HPCBM but did not increase in the LPCBM (Figure S2), the changes were not statistically significant (p > 0.05).It is likely the shorter incubation time in this study compared to previous work (200 days vs 430 days) 22 was insufficient to allow expected PCB dechlorination products to accumulate to statistically significant levels.
Mass fractions of trichlorinated PCB 31 and PCB 20/28 and dichlorinated PCB 8 also decreased after 200 days in the HPCBM compared to the LPCBM.Changes in mass fractions of these known PCB dechlorination products of higher chlorinated congeners, such as PCB 66, 52 are further evidence of active PCB dechlorination in the HPCBM.PCB 4 is often considered a dechlorination end-product. 53Volatilization from the aqueous phase 54 likely explains the greater abundance of PCB 4 at day 0 compared to day 200 in both the HPCBM and LPCBM (Figure 1).
Dehalococcoides Growth in Microcosms.We measured Dehalococcoides (Dhc) growth in the microcosms by quantifying the Dhc 16S rRNA gene abundance over time.Initially, little variability existed between Dhc 16S rRNA gene abundance in the HPCBM and LPCBM.The Dhc 16S rRNA gene abundance in all microcosms decreased from 2 × 10 4 to 1.5 × 10 3 copies/ mL by day 21 (Figure S3).We observed a similar Dhc 16S rRNA gene abundance pattern with time in a previous microcosm study. 22The reason for the initial temporal decrease in Dhc 16S rRNA gene abundance is unknown but could possibly be explained by Dhc cells adapting to new incubation conditions in the microcosms.
Dhc growth in the HPCBM outpaced that of the LPCBM between days 83 and 211 (Figure S3).There was a significant increase in Dhc 16S rRNA gene abundance in the HPCBM between day 83 and day 211 (p = 0.014).The decrease in Dhc 16S rRNA gene abundance in the LPCBM between those time points was not significant (p = 0.119).On day 211, Dhc 16S rRNA gene abundance averaged 1.5 × 10 4 gene copies/mL in HPCBM after 130 days of continuous increase.Conversely, the 16S rRNA gene abundance in the LPCBM averaged 1.2 × 10 3 copies/mL at day 211.The mean Dhc 16S rRNA gene abundance significantly differed between the HPCBM and LPCBM at days 180 (p = 0.008) and 211 (p = 0.0003).DNA and RNA were extracted for sequencing at day 200, after Dhc 16S rRNA gene abundance in the HPCBM had significantly increased relative to that in the LPCBM.
In a previous microcosm study with the same sediments, the Dhc 16S rRNA gene abundance was 3−4 orders of magnitude greater 22 (after 430 days incubation) than in this study (after 200 days incubation).Thus, we would expect additional Dhc growth in the HPCBM had we provided an additional incubation time.
Recovery and Diversity of MAGs from Sediment Metagenomes.A combined individual and coassembly approach has previously been suggested for effective recovery of MAGs. 31The maximum MAG coverage fraction across metagenome samples ranges from 90 to 100%, with 132/160 MAGs covered by at least two metagenomic samples (out bar of Figure 2A,2B).Individual assembly contributed 90/160 dereplicated MAGs, followed by coassembly (38/160) and refined medium-quality MAGs (32/160).The analysis suggests that combining coassembly, individual assembly, and refining medium-quality bins was effective for recovering high-quality MAGs from metagenomes.
The 160 MAGs were classified into 24 phyla according to the Genome Taxonomy Database (GTDB) 55 (Figure 2).The most abundant phyla in both the HPCBM and LPCBM were Caldisericota, Chlorof lexota, Methanobacteriota, Bacteroidota, Desulfobacterota, Spirochaetota, and Firmicutes.Two other phyla, Halobacteriota and Cloacimonadota were only abundant in LPCBM.Phylum-level classifications of the MAGs were compared with those from a previous read-based taxonomic analysis, 23 with the GTDB phylum shown as their corresponding NCBI phylum (Figure S4).There were 46 MAGs classified at a taxonomic level higher than that of the genus and 49 MAGs with no corresponding NCBI genus classification (Table S4), which suggests that they are potentially novel taxa.Many of the abundant MAGs have not been previously reported in PCBdegrading microbial communities harboring Dehalococcoides.Additional taxonomic analysis is provided in Section S4.
The genome-resolved approach used here contrasts with those of many previous metagenomic studies of organohaliderespiring microbial communities.Using read-based methods 2,56−58 makes it challenging to establish clear links between phylogeny and function.−64 Here, our genome-resolved analysis yielded 160 MAGs in an effort to represent the sediment microbial community.This approach provides new insights about relationships between MAGs and guidelines on MAG recovery from organohaliderespiring microbial communities.
Although many high-quality MAGs were recovered from organohalide-respiring microbial communities in this study by analysis of short-read sequences (i.e., Illumina platform), reconstructing contigs and MAGs from short-read sequences is complicated, time-consuming, 65,66 and potentially unreliable because of read mapping uncertainties. 67More recently developed long-read sequencing platforms can now routinely generate reads >10 Kb. 65−67 Long reads are more suitable for de novo assembly and provide more complete MAGs or draft genomes for microbial identification. 65,67Long-read sequencing should be considered more frequently for future research involving recovery of MAGs from microbial communities.
Dehalococcoides Pangenomic Analysis.We compared the individually assembled Dehalococcoides MAGs from this study with 46 of 51 Dehalococcoides genomes and assemblies found on NCBI in April 2023 (including the coassembled Dehalococcoides MAG 25 ) (Figure S5).Dehalococcoides MAG sizes ranged from 1.2 to 1.4 Mb, which are similar to complete Dehalococcoides genome sizes in NCBI (Figure S5).These MAGs contained between 1190 and 1448 coding genes, which suggests that these Dehalococcoides have streamlined genomes. 58,68ierarchical clustering of single-copy genes (SCGs) in genomes/assemblies revealed the three distinct clades known as Pinellas, Victoria, and Cornell, 69 which was also supported by the ANI between genomes/assemblies.The Dehalococcoides MAGs were placed in the Cornell subgroup (Figure S5).The ANI among the three Dehalococcoides MAGs was >99.8% and was >98% with PCB-dechlorinating strain CG4.A partial 16S rRNA gene was recovered from the coassembled Dehalococcoides Environmental Science & Technology MAG 25 but not from the individually assembled Dehalococcoides MAGs.
A total of 27 RDase genes with >90% identity to each other were annotated in the Dehalococcoides MAGs (Figure S6).The individually assembled Dehalococcoides MAGs do not harbor any previously identified PCB dehalogenase genes (i.e., pcbA1, pcbA4, pcbA5, 10 and mbrA 13 ) and as previously reported for the coassembled Dehalococcoides MAG. 25 Compared to a previous pangenomic analysis which compared the genomic and evolutionary characteristics of Dehalococcoidia, 70 here, the pangenomic analysis focuses on previously identified PCB dehalogenase genes in other Dehalococcoides genomes/assemblies (further described in Section S5 and Table S5).
Gene Expression in Dehalococcoides MAGs and Identification of a Novel, Expressed Reductive Dehalogenase Gene.We examined gene expression in the Dehalococcoides MAGs in metatranscriptomic samples (Figure 3 and Table S6).Two of the 27 annotated RDase genes (JHMAAFGB_00007 and JHMAAFGB_00005) were expressed in the HPCBM.JHMAAFGB_00007 was expressed in both HPCBM bottles (Figure 3), but JHMAAFGB_00005 was only expressed in one HPCBM bottle (Table S6).JHMAAFGB_00007 was uniquely present in one Dehalococcoides MAG (5_bin.153;labeled with rdhA in Figure S5).JHMAAFGB_00007 was not present in any other Dehalococcoides genomes/assemblies on NCBI, although it is in a variable region of the Dehalococcoides pangenome (Figure S5).Conserved RDase domains, including Pfam13486 (Fe−S binding domain) and TIGR02486 (corrin and Fe−S clustercontaining domain), were identified in these expressed RDase genes (Figure S7).
In other PCB dehalogenation studies using transcriptomics, the Dehalococcoides cell density was several orders of magnitude higher than that in our study.Thus, a wide range of RDase gene expression from Dehalococcoides genomes was detected, but only those that were significantly expressed were identified as PCB dehalogenase genes. 10,12,13,71Here, RNA was extracted from microcosms that were far less enriched for Dehalococcoides.It is possible that other RDase genes were expressed in HPCBM at levels that were undetectable by metatranscriptomic sequencing despite the robust sequencing coverage (45−71 million sequences per sample).
Temporal analysis of RDase gene expression in some Dehalococcoides strains (e.g., CBDB1) show a peak in expression during early growth stages that declines in later stages, 72−75 while in other strains, RDase gene expression remains at constant levels until the organohalide substrate is depleted. 76ere, we evaluated RDase gene expression at one time point in the HPCBM, likely in the early growth phase, according to qPCR data.A time-series metatranscriptomics analysis could have revealed JHMAAFGB_00007 and JHMAAFGB_00005 expression dynamics during growth phases and whether these were the only RDases significantly expressed temporally, as seen in other PCB-dechlorinating Dehalococcoides strains. 76 phylogenetic tree of RDase genes from Dehalococcoides MAGs and selected genes from the Reductive Dehalogenase Database 69   (JHMAAFGB_00380 and JHMAAFGB_00392, respectively), and RDase genes cloned from the same site (Figure S9).21,22 T h e R D a s e g e n e s ( J H M A A F G B _ 0 0 3 8 0 a n d JHMAAFGB_00392) were >90% identical to RDase genes retrieved from the clone library (Figure S6).
The expressed RDase gene (JHMAAFGB_00007) was in a small cluster away from most RDase genes in the RDD.An RDase gene from 1,2,3,4-tetrachlorodibenzo-p-dioxin-dechlorinating Dehalococcoides strain H1−3−2.001 61(PKH47860.1,OG342) grouped nearest to JHMAAFGB_00007 (Figure S9).BLAST analysis indicated that the pairwise amino acid identity between JHMAAFGB_00007 and PKH47860.1 was 48.18%.JHMAAFGB_00007 was annotated as 3-chloro-4-hydroxyphenylacetate reductive dehalogenase (cprA) with the Prokka default database, although the pairwise amino acid identity was 23.75%.The cprA gene product from Desulfitobacterium hafniense transforms 3-chloro-4-hydroxyphenylacetate to 4hydroxyphenylacetate and dehalogenates chlorinated phenols. 77,78nalysis of genes on the contig that harbored the significantly expressed rdhA (JHMAAFGB_00007) revealed that the second expressed rdhA (JHMAAFGB_00005) was encoded nearby (Figure 4A).Genes that encode dehalogenase anchoring protein (rdhB) were also located as expected near JHMAAFGB_00007 and JHMAAFGB_00005.None of the genes on this contig had annotated functions related to horizontal gene transfer.The predicted tertiary structure of the JHMAAFGB_00007 product (Figure 4B) was compared with the predicted tertiary structures of cprA, pcbA4, and PKH47860.1 products (Figure 4C,D).The predicted dehalogenase domain of the JHMAAFGB_00007 product was distinct from the other predicted protein structures.However, between positions 157 and 289 (within the dehalogenase domain), the predicted structures of the JHMAAFGB_00007 and PKH47860.1 products were similar, as shown by the small root-mean-square error at each atomic position values (Cα RMSE).This suggests that the JHMAAFGB_00007 product catalyzes the removal of chlorine from chlorinated aromatics (e.g., PCBs).Consequently, the lines of evidence presented thus far suggest that the significantly expressed RDase gene (JHMAAFGB_00007) is novel and has a high potential to encode a PCB dehalogenase.However, further experimental evidence is needed to validate PCB transformation by the product of this rdhA.
Correlations between Overall MAG Gene Expression within the PCB-Dehalogenating Community.The coexpression network analysis revealed that 154 of 160 MAGs were significantly correlated to others (Figure 5).Of these, 39 MAGs were significantly correlated to Dehalococcoides MAGs.We The growth of Dehalococcoides requires hydrogen, acetate, and corrinoid cofactor 8 that are produced by other microbes in the community.MAGs that express a hydrogenase gene and any short chain fatty acid transformation gene could be considered acetate and hydrogen producers.Microbes that salvage, assemble, and transport portions of the corrinoid cofactor are also beneficial to Dehalococcoides growth.Important functions include cobyrinate a,c-diamide production (expression of cbiA), 5,6-dimethylbenzimidazole (DMB) production (expression of DMB synthase), corrinoid salvaging (expression of cobS/cobV), and corrinoid transport (expression of btuB and either btuC or btuD).The Dehalococcoides MAGs contain an incomplete Wood−Ljungdahl pathway (Figure S10), which could lead to carbon monoxide (CO) accumulation and impact growth. 15hus, Dehalococcoides could benefit from the activity of CO metabolizers (as identified by the expression of a CO dehydrogenase gene).A gene encoding resuscitation-promoting (RP) factor (rpf B) reportedly accelerated the enrichment of an anaerobic PCB-dechlorinating culture, 79 suggesting that RPproducing microbes, i.e., those expressing rpf B, can enhance Dehalococcoides growth.More details of functional roles associated with Dehalococcoides growth are described in Section S6.
We examined expression of selected genes involved in the metabolic processes described above within each of the 160 MAGs (Section S6, Figure S11, and Table S7).MAGs were assigned a functional role if they contained and expressed one or more key functional genes in a minimum of two samples.Among the 160 MAGs, 112 were assigned at least one functional role (Figure 5 and Table S7).Of these 112, 31 were significantly correlated to Dehalococcoides MAGs.A higher proportion of MAGs correlated with Dehalococcoides had a functional role (31/39) compared to that of the MAGs overall (112/160).This is anticipated as those functions were specifically chosen for promoting the growth of Dehalococcoides.This is the first time that metabolic functions known to benefit the activity of Dehalococcoides were verified in a microbial community with coexpression analysis.
The percentage of the total gene expression related to Dehalococcoides MAGs for cobyrinate a,c-diamide (72.2%) and DMB producers (62.4%) were elevated compared to the other roles (13.8−42.4%)(Figure S12), which highlights the importance of these intermediates in supporting growth of Dehalococcoides.In contrast to cobyrinate a,c-diamide and DMB producers, gene expression by corrinoid assemblers correlated to Dehalococcoides MAGs was 13.8%.The relatively low representation and activity of corrinoid assemblers and DMB producers in this community suggest that corrinoid synthesis processes could limit growth of Dehalococcoides and slow PCB dechlorination.
Although RP was reported to accelerate enrichment of a PCBdechlorinating culture, 79 RP producers appear to be rare in sediment microbial communities.The relatively high percentage of total gene expression attributed to RP producers correlating to Dehalococcoides MAGs (42.4%) suggests that adding the exogenous resuscitation-promoting factor (Rpf) could be an alternative strategy for developing PCB-dechlorinating enrichment cultures.
Specific CO metabolizers and acetate/H 2 producers appear to be less critical roles supporting Dehalococcoides growth because a smaller percentage of their total expression (29.0% for CO metabolizers and 26.1% for acetate/H 2 producers) was related to Dehalococcoides MAGs and their widespread presence (86/  154 MAGs).This suggests that Dehalococcoides growth in these microcosms was less affected by CO accumulation or carbon source and electron donor limitations.
I n t e r e s t i n g l y , t w o M A G s ( G C A _ 0 2 1 3 7 2 3 3 5 . 1 _ A S M 2 1 3 7 2 3 3 v 1 a n d GCA_021372315.1_ASM2137231v1) classified as methanogens (g_Methanomassiliicoccus) and identified as cobyrinate a,cdiamide producers and corrinoid assemblers, were also correlated with Dehalococcoides MAGs (Table S8), although this genus would compete with Dehalococcoides for H 2 . 80Indeed, nine MAGs classified as methanogens (g_Methanoregula, s_Methanoculleus sp002501655, g_Methanofastidiosum, f_Methanobacteriaceae, and g_Methanobacterium) were not significantly coexpressed with Dehalococcoides MAGs (i.e., TPM values were 1−2 orders of magnitude higher than Dehalococcoides MAG TPMs).−85 This suggests that network coexpression analysis can reveal potential competition for H 2 between methanogens and OHRB.Competition for H 2 could also contribute to slow Dehalococcoides growth despite the widespread presence and activity of acetate/H 2 producers in the community. 86nother MAG, classified as g_Rhodococcus, harbored several aromatic-ring cleavage dioxygenases and a complete aerobic corrinoid biosynthesis pathway, making it a potential aerobic PCB degrader and corrinoid producer.However, gene expression from this MAG was low (Table S8) and its corrinoid biosynthesis genes were not expressed.Only three of its expressed genes correlated with Dehalococcoides (Table S9).A PCB-degrading Rhodococcus strain was previously found in a viable but nonculturable (VBNC) state under anaerobic conditions, and its ability to degrade PCBs recovered after exposure to oxygen. 87We speculate that this Rhodococcus MAG represents an aerobic PCB degrader in a VBNC state under the

Environmental Science & Technology
anaerobic conditions in these microcosms; however, further tests would be needed to validate its VBNC state.
Several MAGs with less characterized taxonomy and without an assigned functional role were also correlated to Dehalococcoides MAGs (Figure 5B,C).6][47][48]88 Because the functional potential of the OTUs/ASVs from less-studied taxonomies is typically poorly understood, this could lead to overemphasis on better-known taxonomies while overlooking other possibly important taxa. A enome-resolved approach can overcome these phylogeny/function issues.For example, MAG 7_bin.119classified as o_Syntrophales represents a versatile potential CO metabolizer and producer of DMB, corrinoids, H 2 , and acetate that was also significantly correlated with the coassembled Dehalococcoides MAGs (Table S8).We also note that the network analysis was performed for one time point, which was 200 days after stimulating the microcosms with an electron donor.A network based on a temporal metatranscriptomics analysis might have revealed additional microbial community interactions that occurred either prior to or after the onset of Dehalococcoides growth in the HPCBM.
To further explore possible functional connections between MAGs without assigned functional roles that correlated with the Dehalococcoides MAGs (Figure 5C), an additional network was constructed with expressed genes from MAGs without assigned functional roles.An MAG (classified as class Sumerlaeia) had a high TPM, which led to correlations between 1011 expressed genes in that MAG to those in the Dehalococcoides MAGs (Figure S13).More than half (550) of the genes were annotated as hypothetical proteins.The remaining 461 genes with KO identifiers were used to reconstruct the metabolism pathways.Although many expressed genes were categorized under fundamental cellular processes (e.g., carbon metabolism (27)  and biosynthesis of secondary metabolites (73) and amino acids (24)), we also noted 35 expressed genes involved in biofilm formation and 15 expressed bacterial secretion system genes for cell wall-degrading enzymes (Table S9).These results are consistent with the observed expression of a cyclic di-GMP phosphodiesterase gene, which controls bacterial cell adhesion 89 by Dehalococcoides in the HPCBM (Figure 3).Overall, these results suggest that biofilm formation was potentially beneficial to Dehalococcoides in the HPCBM.

■ ENVIRONMENTAL IMPLICATIONS
Genome-resolved metagenomics and metatranscriptomics yielded a relatively large database (160) of representative reference MAGs from indigenous PCB-dechlorinating sediment communities compared to previous genome-resolved analyses of dechlorinating communities 62−64 and further enabled identification of active novel taxa and expressed functions and potentially new interactions in sediment microbial communities that dechlorinate PCBs.Genome-resolved metatranscriptomics facilitated the detection and analysis of an expressed (and previously unknown) RDase gene in the Dehalococcoides MAGs.The evidence presented strongly suggests that Dehalococcoides are responsible for anaerobic PCB dechlorination processes in the HPCBM and also likely contribute to natural attenuation of PCBs at the site. 21Although 112 MAGs were assigned functions that could support Dehalococcoides growth, only 31 were strongly coexpressed with Dehalococcoides MAGs.The low proportion (13.8%) of corrinoid producers that correlated to Dehalococcoides MAGs could explain the slow growth of Dehalococcoides in the microcosms.Future work could focus on shaping the desired microbial profile by adjusting the physicochemical parameters of the microbial habitat to increase the dechlorination rate.Overall, genome-resolved metagenomic and metatranscriptomic analyses provides valuable insights into the ecology and interactions of microbial communities containing OHRB (e.g., Dehalococcoides) and could facilitate new research aimed at enhancing reductive dechlorination rates in bioaugmentation cultures and at contaminated sites via biostimulation.

Data Availability Statement
Raw metagenomic reads (SRA accession numbers SRX11347095 to SRX11347098) and metatranscriptomic sequencing data (SRA accession numbers SRX11347095 to SRX11347098) and high-quality MAGs are available under BioProject accession number PRJNA743546 in NCBI.The scripts used for the metagenomic and metatranscriptomic analysis are accessible via GitHub: github.com/danghongyu/Workflow_for_genome_resolved_analysis.
Additional information about nucleic acid extraction, qPCR settings, sequencing, and bioinformatics analysis; PCB concentration in microcosms; results for the diversity of recovered MAGs, RDases, and functional role assignment; taxonomic classification, phylogeny, RPKM, covered fraction and functional annotation of the MAGs; oligonucleotide primer information; pertinent qPCR parameters in accordance with MIQE guidelines; comparison of individual PCB congeners as measured in high-and low-PCB-contaminated sediment microcosms; and Dehalococcides pangenomic analysis with selected genomes/assemblies from NCBI and Dehalococcoides MAGs with a heatmap showing the average nucleotide identity (ANI) (PDF) List of MAGs with a completion (comp.)≥80; RPKM and GTDB-tk taxonomy classification of dereplicated MAGs and the correlated NCBI phylum and genus; pairwise BLAST results of genes in the same cluster with identified PCBs dechlorinating genes from Dehalococcoides pangenomic analysis; all expressed genes from individually assembled and coassembled Dehalococcoides MAGs; functional role assignment for MAGs based on the expression of genes involved in corrinoid synthesis, lactate and H 2 production, CO metabolism, and RP production; correlation to Dehalococcoides MAGs and functional role assignment for all MAGs presented in the network, together with log 10 (TPM), lowest taxonomy assignment, and phylum; and metabolism reconstruction of significantly expressed genes from MAG 5_191 and genes from the MAG of Rhodococcus (XLSX)

Figure 1 .
Figure1.Profile comparisons of all 209 PCB congener mass fractions on days 0 and 200 in the HPCBM (top) and LPCBM (bottom).PCBi = PCB congener numbers from 1 to 209.The relationship between the PCB congener number and the IUPAC name can be found at www.epa.gov/sites/default/files/2015-09/documents/congenertable.pdf.When a PCB is labeled with multiple numbers (e.g., PCB 61/70/74/76), this indicates that these congeners coelute during gas chromatography separation of all 209 congeners.The black diagonal line on each plot shows where values would appear if the PCB congener mass fraction at day 0 equals the mass fraction at day 200.Congeners falling above or below the black line indicate a higher or lower mass fraction from day 0 to day 200 in the HPCBM or LPCBM, respectively.Data points represent the average PCB congener mass fraction measured in duplicate bottles.PCB congener number labels not shown in the plots were omitted for clarity.Error bars represent the variability in measurements collected from biological duplicates.The vertical error bar is the variability tested on day 200 and the horizontal error bar is the variability tested on day 0. The HPCBM data was adapted from ref 25. with the permission of Oxford University Press, copyright 2022.

Figure 2 .
Figure 2. (A) Maximum likelihood phylogenetic tree of MAGs, taxonomy, genome length, log 10 (RPKM) and coverage fraction in metagenomic data.Blue circles at the end of tree branches depict coassembled MAGs, no circles depict individually assembled MAGs, and red circles depict MAGs with a refined individual assembly.The shading of the branches indicates the phylum.The prefix s_, g_, f_, o_, c_, and p_ in the taxa name represents the taxonomy rank at species, genus, family, order, class, and phylum levels.The innermost bar chart depicts the overall contig length in each MAG (gray bars).The heatmap shows the log 10 (RPKM) of MAGs in the HPCBM and LPCBM metagenomes.The bar charts in the outer circle show MAG coverage fractions in duplicate samples with a high (blue; HPCBM) or low (red; LPCBM) PCB concentration.(B) Distribution MAGs with more than 90% coverage fraction by the number of samples.
(RDD) revealed that RDase genes in Dehalococcoides MAGs grouped within 26 different ortholog groups (OGs) (Figure S9).Two OGs (OG14 and OG15) contained RDase g e n e s r e t r i e v e d f r o m D e h a l o c o c c o i d e s M A G s

Figure 3 .
Figure 3. Log 10 of transcripts per million (log 10 TPM) values for genes from Dehalococcoides MAGs expressed in at least two metatranscriptomics samples with high (red) or low (blue) PCB concentration.

Figure 4 .
Figure 4. (A) Genes encoded on the contig containing the expressed rdhA (JHMAAFGB_00007).(B) The predicted tertiary structure of the JHMAAFGB_00007 product with the dehalogenase domain is highlighted in pink and the Fe−S domain in dark blue.(C from left to right) Comparison of the predicted JHMAAFGB_00007 product structure to cprA, pcbA4, and PKH47860 (D) The root-mean-square error at each atomic position (Cα RMSE) is between the expressed rdhA (JHMAAFGB_00007) and the other three predicted structures with highlighted dehalogenase (blue) and Fe−S (orange) domains.Discontinuities represent missing positions for structure matching.The blue dashed line represents Cα RMSE = 5, and the red dashed line represents Cα RMSE = 1.