Anaerobic microcosms were set up in duplicate with the marine sediment collected from the Gulf of Taranto (Mar Piccolo, Ionian Sea, Italy – sampling station “1l” 40° 28′ 46 N, 17° 15′ 38 E) (Supplementary Figure S1). The marine sediment was collected just below (~1 cm) the water/sediment interface using polycarbonate sample tubes as described in Franzo et al. (2016).

Sediment samples were immediately transferred to the laboratory for the microcosms set-up. The 90 g of dry weight sediment were anaerobically incubated in sterile 250-mL serum bottles with 70 mL of synthetic marine water as previously described in Matturro et al. (2016b). The bottles were sealed with Teflon-faced butyl rubber stoppers and fluxed with a mixture of N₂/CO₂. All microcosms were incubated at 20°C under rotation. After 350 days of anaerobic incubation, lactate was added as fermentable carbon source (0.7 mM). A sterile control microcosm was also prepared with the autoclaved marine sediment and no PCB dechlorination was observed. Chemical and biological analyses were conducted during the anaerobic incubation before (350 days of incubation) and after the lactate addition (420 days of incubation).

Four gams of slurry from each microcosm were collected in 30 mL glass tubes and stored at -20°C until further processing. PCBs were extracted from the slurry collected in each microcosm and then quantified following the procedure reported in Matturro et al. (2016b). Quantification was based on a three-point calibration curve and data were expressed as ng g^-1 dry sediment.

Data reported in the present study are referred to the PCB quantification performed before and after lactate addition. The detailed composition of the original marine sediment used to set up the microcosm study is reported in Matturro et al. (2016b).

The sediment slurry (1 g dry weight) was anaerobically collected before and after lactate addition from the serum bottles with sterile spatulas under a N₂ flux. Samples were immediately fixed in formaldehyde (2% vol/vol final concentration) and then processed to extract cells from sediment particles as previously described (Barra Caracciolo et al., 2005), using Optiprep^® (Sigma, Italy) as density gradient medium instead of Nicodenz (Sigma, Italy). Extracted cells were filtered through a 0.2 μm polycarbonate membrane (Millipore, 25 mm diameter) by gentle vacuum (<0.2 bar) and used in CARD-FISH assay as previously described (Matturro et al., 2016b). After the hybridization assay, total cells were stained with Vectashield Mounting Medium^® with DAPI (Vector Labs, Italy). Cell counting was performed through microscopic analysis on at least 20 randomly selected microscopic fields for each sample. Cell abundances were expressed as cells per dry weight of marine sediment (cells g^-1). Means and standard deviations were calculated with Microsoft Excel^®.

DNA was extracted from 0.25 g of dry weight sediment in the original sample and from the microcosm before and after lactate addition. The extraction was performed with PowerSoil DNA Isolation kit (MoBio, Italy) within 48 h from each sampling according to the manufacturer’s instructions. DNA was eluted in 100 μL of sterile water and the concentration and purity were determined by NanoDrop 2000c spectrophotometer (Thermo Scientific, USA). Aliquots were stored at -80°C for a few days and then used for Real time PCR quantification (qPCR) and Next Generation Sequencing (NGS).

DNA was used for qPCR absolute quantification assays targeting D. mccartyi 16S rRNA and reductive dehalogenase genes tceA, bvcA, vcrA, pcbA1, pcbA4, pcbA5. qPCR reactions targeting 16S rRNA, tceA, bvcA, vcrA genes were performed with TaqMan^® chemistry in 20 μL total volume of SsoAdvanced^TM Universal Probes Supermix (Biorad, Italy), including 3 μL of DNA as template, 300 nM of each primer and 300 nM of TaqMan^® probe composed by 6-carboxyfluoresceine (FAM) as the 5′ end reporter fluorophore and N,N,N,N,-tetramethyl-6-carboxyrhodamine (TAMRA) as the 3′ end quencher.

qPCR reactions targeting pceA, pcbA1, pcbA2, pcbA3 dehalogenase genes were performed with SybrGreen^® chemistry in 20 μL total volume of SsoAdvanced^® Universal SYBR^® Green Supermix (Biorad, Italy) including 3 μL of DNA as template and 300 nM of each primer. Primers and probes used for each reaction are listed in Supplementary Table S1. Standard curves for the absolute quantification were constructed by using the long amplicons method previously reported in Matturro et al. (2013). Each reaction was performed in triplicate with CFX96 Touch^TM Real-Time PCR Detection System (Biorad, Italy). Quantitative data were expressed as gene copy numbers g^-1 sediment, and error bars were calculated with Microsoft Excel^® on triplicate reactions for each sample.

16S rRNA gene of the microbial community in the original marine sediment and in the marine sediment after lactate addition, was amplified using primers 27f (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492r (5′-TACGGYTACCTTGTTACGACTT-3′) for the Bacteria domain using Hot Start Taq98 (Lucigen, Italy). PCR reactions were performed with the following cycle: 2 min at 98°C, 30 s at 98°C + 30 s at 58°C + 1 min at 72°C for 38 cycles and 15 min at 72°C. PCR products were purified using the QIAquick^® PCR purification kit (Qiagen, Milan, Italy). Cloning of PCR products was carried out using pGEM-T Easy Vector System (Promega, Italy) into Escherichia coli JM109 competent cells (Promega, Italy) according to the manufacturer’s instructions. Positive inserts were amplified from recombinant plasmids obtained from white colonies by PCR using the sequencing primers T7f (5′-TAATACGACTCACTATAGGG-3′) and M13r (5′-TCACACAGGAAACAGCTATGAC-3′) and amplicons of 1465 bp length were purified using the QIAquick PCR purification kit (Qiagen, Milan, Italy). 16S rRNA gene sequences of the clone inserts were obtained using the following primers: 530f (5′-GTGCCAGCMGCCGCCG-3′), 926f (5′-AAACTYAAAKGAATTGACGG-3′), 907r (5′-CCGTCAATTCMTTTRAGTTT-3′), 519r (5′-GWATTACCGCGGCKGCTG-3′).

The 16S rRNA gene sequences were analyzed with the ARB software (Ludwig et al., 2004) using the SILVA 16S rRNA SSU Reference database release 102 (Pruesse et al., 2007). Sequences were analyzed for chimera formation using the chimera-checking tool QIIME Software. The phylogenetic tree was constructed using the maximum likelihood method RAxML (Stamatakis et al., 2008). For the construction of the phylogenetic tree with sequences from the original marine sediment a total of 18 16S rRNA gene sequences were used and Chlamidyae was chosen as outgroup (Supplementary Table S3).

For the construction of the phylogenetic tree with sequences from marine sediment after lactate addition a total of 20 rRNA gene sequences were used and Acidobacteria was chosen as outgroup (Supplementary Table S4).

The 16S rRNA gene sequences were deposited in the GenBank database under the PopSet accession number 969532403 (sequences obtained from the original marine sediment) and the PopSet accession number 969532350 (sequences obtained after lactate addition).

The sediment was taken from the Station 1l, one of the most polluted areas of the Mar Piccolo heavily contaminated by esa-CBs, penta-CBs, and hepta-CBs (Cardellicchio et al., 2016). In particular, esa-CB 153 and the mixture of PCBs 163+138 were mainly found at concentration >600 ng g^-1 dry sediment (Figure 1A). As demonstrated in a previous treatability study (Matturro et al., 2016b), PCB RD was sustained by the sediment organic carbon under controlled anaerobic conditions with an overall decrement of the most high-chlorinated PCBs up to 50% and the simultaneous increment of low-chlorinated congeners usually found as byproducts of the PCB anaerobic RD (i.e., congeners 18, 28+31) (Bedard, 2008).

In order to further enhance the PCBs dechlorination, in this study we have evaluated the effectiveness of the addition of a fermentable organic substrate. As reported in Figure 1A, PCB dechlorination was efficiently promoted and a further decrement of at least 20% up to 70% of the main congeners was observed after lactate addition. Marked decrements were reported for the highest chlorinated congeners 183 (70% of decrement), for the mixture of PCBs 163+138 (63% of decrement) and for the congener 153 (50% of decrement). Additionally, tetra-CB 77 was completely depleted and likely transformed into less chlorinated by-products. In particular, among the screened congeners, tetra-CB 47 strongly increased (from 0.06 to 50 ng g^-1 dry marine sediment) at the end of the treatment (Figure 1B).

Next Generation Sequencing analysis of 16S rRNA gene fragments was performed on the original marine sediment, during the anaerobic incubation and after the lactate addition. Sample preparation using the V1–3 bacterial primers were successful for all sampling times and yielded between 16.040 and 28.419 reads after QC and bioinformatic processing. A total of 851 OTUs including 21 OTUs affiliated to unknown phyla and 830 OTUs related to 35 already described phyla was obtained.

The composition of the core microbiome did not drastically change with the incubation under controlled anaerobic conditions. Similar taxonomic distinctness (TD) and evenness (E) indices were estimated on the marine sediment (TD = 1.66; E = 0.35) and after 350 days of anaerobic incubation (TD = 1.72; E = 0.39). Conversely, both indices appreciably decreased after lactate addition (TD = 1; 1 – D = 0.16).

In detail, in the original marine sediment 16S rRNA gene sequences were mainly affiliated to ε-proteobacteria (37%), γ-proteobacteria (12%), Chloroflexi (11%), δ-proteobacteria (9%), α-proteobacteria (9%), Firmicutes (6%) and Bacteroidetes (5%) (Figure 2A; Supplementary Table S2).

The relative abundances of the OTUs detected in the sediment remained quite constant during the incubation under controlled anaerobic conditions (Figure 2A). After the addition of lactate, ε-proteobacteria affiliated to Sulfurovum and Sulfurimonas genera (family Helicobacteraceae) increased up to 70% of total OTUs (Supplementary Table S2). Sulfurovum was the most abundant genus in the contaminated marine sediment (23.5%) and its relative abundance increased with the treatments reaching the highest value after lactate addition (63.6%). Conversely, Sulfurimonas decreased overtime (from 11.7 to 4.68%) (Supplementary Table S2).

OTUs affiliated to γ-proteobacteria included iron-oxidizing genus Acidiferrobacter and numerous sulfur-oxidizing bacteria such as Thiomicrospira genus (Piscirickettsiaceae), Thioal kalispira (Ectothiorhodospiraceae) and genera Sedimenticola, Thioalophilus, and Marinicella affiliated to unknown families.

Chloroflexi members were mainly affiliated to Dehalococcoidia class (≈ 60% of Chloroflexi) which includes numerous organohalide respiring microorganisms and, to a lesser extent, to the anaerobic and chemoorganotrophic Anaerolinea (≈ 30% of total Chloroflexi) and Ardenticatenia (<10% of total Chloroflexi), the latter reported to grow by dissimilatory iron- and nitrate-reduction (Kawaichi et al., 2013) (Figure 2B). Within Dehalococcoidia, Dehalobium was the only genus identified by NGS and represented only ≈ 6% of this class in the original marine sediment and <6% after the lactate addition. Among Chloroflexi, Anaerolinea and Dehalococcoidia increased over the anaerobic incubation (Supplementary Table S2).

Members of δ-proteobacteria were mainly affiliated to Desulfobacteraceae, Syntrophobactericeae and Desulfoarculaceae that are strictly anaerobic microorganisms able to use simple organic molecules as electron donors and sulfate or thiosulfate as electron acceptor (Sun et al., 2013). Relative abundance of sequences related to sulfate reducing bacteria increased under controlled anaerobic incubation and declined after lactate addition (Supplementary Table S2). Additionally, OTUs affiliated to α-proteobacteria comprised bacteria possessing the biochemical and ecological capacities to degrade organic pollutants and to be resistant to heavy metals such as Rhodobacterales, Rhodospirillales, and Rhizobiales, the latter recently considered a promising candidate for PCB degradation through oxidative pathways (Teng et al., 2015). In particular, facultative anaerobic photoheterotrophic genera such as Rhodobium were among the most abundant OTUs in the contaminated sediment (1.6%) and strongly increased over the anaerobic incubation (up to 4.8%).

Bacteroidetes was found at 4.8% in the contaminated sediment and mainly comprised of anaerobic, halophilic bacteria affiliated to Marinilabiaceae family and to uncultured Bacteroidetes BD2-2.

Firmicutes represented 3.3% of total OTUs and were mainly affiliated to Thermodesulfobiaceae, which includes several sulfate reducing bacteria (Mori et al., 2003).

Interestingly, evidences of moderately thermophilic and anaerobic chemoorganotrophs capable of fermenting proteinaceous substrates were also found in the marine sediment. In particular, sequences belonging to Coprothermobacter spp. (Firmicutes phylum) and to Caldithrix spp. (Deferribacteres phylum) were found and overall accounted for about 3% of total sequences (Supplementary Table S2). Relative abundances of other OTUs were <3% (Figure 2A; Supplementary Table S2).

Furthermore, an accurate estimation of Bacteria and Archaea cell densities was performed by CARD-FISH (Figure 3). Bacterial cell abundances accounted for 1.2E + 07 ± 1.6E + 06 cells g^-1 dry sediment and increased up to 3.3E + 07 ± 9E + 06 and 3E + 08 ± 1.6E + 07 cell numbers g^-1 dry sediment, respectively, before and after the lactate addition (Figure 3). Archaea also increased during the treatments from 4.3E + 06 ± 1.7E + 06 cells g^-1 dry sediment detected in the original marine sediment to 3.44E + 07 ± 7.4E + 06 and 1.11E + 08 ± 2.7E + 07 cells g^-1 dry sediment quantified in the sediment before and after lactate addition, respectively (Figure 3).

Chloroflexi were quantified by CARD-FISH in the original marine sediment (3.6E + 06 cells g^-1 dry sediment) and increased up to 2.0E + 07 cells g^-1 dry sediment after lactate addition. Among Chloroflexi, D. mccartyi cells increased overtime representing 25% and 64% of total Chloroflexi in the marine sediment and at the end of the treatment, respectively (Figure 4A).

Moreover, the occurrence of D. mccartyi 16S rRNA and reductive dehalogenase genes (pceA, tceA, vcrA, bvcA, pcbA1, pcbA4, and pcbA5) was ascertained and quantified by qPCR (Figures 4B,C). A total of 8.33E + 06 D. mccartyi 16S rRNA gene copies g^-1 dry sediment were found in the original marine sediment. The anaerobic incubation and the addition of lactate enhanced the total D. mccartyi 16S rRNA genes, whose abundances accounted for 1.77E + 07 and 3.11E + 07 gene copies g^-1 dry sediment, respectively (Figure 4B). D. mccartyi strains carrying pceA, tceA, and vcrA genes were found in the original marine sediment at low abundances ranging between 1E + 02 – 5E + 04 gene copies g^-1 dry sediment and did not drastically increase during the treatments (Figure 4B). Diversely, D. mccartyi strains carrying pcbA5, pcbA4, pcbA1 genes were more abundant. Indeed, 7E + 05 and 5.4E + 05 gene copies g^-1 dry of pcbA5 and pcbA4 genes were detected in the original marine sediment followed by pcbA1 gene (2.5E + 05 gene copies g^-1 dry sediment). Despite D. mccartyi strains carrying PCB-dechlorinase genes increased during the treatments (Figure 4B), they represented only a small fraction of total D. mccartyi detected by qPCR. Indeed, in the original marine sediment and after lactate addition, D. mccartyi strains carrying not identified reductive dehalogenase genes accounted for 80–90% of D. mccartyi 16S rRNA gene copies. Diversely, at the end of the anaerobic incubation, where most of the PCB RD was observed, D. mccartyi strains carrying pcbA5 were mainly found and represented 60% of D. mccartyi 16S rRNA gene copies (Figure 4C).

Further insight into the microbial community composition was performed by preparing a 16S rRNA gene clone library with DNA extracted from the marine sediment and at the end of the treatments. The PCR primers used in the clone library allowed the harvest of 16S rDNA from bacterial members in the community. The clone inserts were sequenced and a total of 77 partial or nearly complete 16S rRNA gene sequences were obtained from DNA extracted from the original marine sediment (Supplementary Table S3) and after lactate addition (Supplementary Table S4). Almost all clones obtained from the original marine sediment fell into previously described phyla of the bacterial domain, with the majority being members of the γ-Proteobacteria (27%) and ε-proteobacteria (27%) phyla. Other clones were in the δ-Proteobacteria (11.5%), α-Proteobacteria (11.5%), Planctomycetes (7.7%), Verrucomicrobia, Acidobacteria, Cyanobacteria, and candidate division OP8 (~4% each) (Supplementary Table S3). Moreover, the phylogenetic analysis of the 16S rRNA gene sequences obtained from the original marine sediment showed that only ≈35% of the clones were closely related to known bacteria with similarities ranging between 97 and 99% (Figure 5). They were mainly affiliated to Sulfurovum aggregans (ε-proteobacteria), Alcanivorax venustensis (γ-proteobacteria), Brevundimonas spp. Marine-1 (α-proteobacteria). Remarkably, most of the clones showed low similarity (≤92%) to known microorganisms and they were phylogenetically located within δ-proteobacteria, Candidate division OP8, Planctomycetes, Verrucomicrobia. Several sequences were also found phylogenetically related to Chromatiales members (γ-proteobacteria) and in particular located within the Thioalkalispiraceae, Ectothiorhodospiraceae, and Chromatiaceae families. Additionally, the sequence KU302736 was phylogenetically located within the Chloroplast lineage and the most close sequences belonged to the Picea glauca (white spruce), the latter known to promote microbial biodegradation of PCBs via the release of phytochemicals upon fine root death (Slater et al., 2011).

After the addition of lactate, 16S rRNA gene sequences obtained from the contaminated sediment were mostly related to ε-proteobacteria (33.3%) and γ-proteobacteria (21%). Other clones were in the Cyanobacteria (12.5%), Firmicutes (8%) and to a lesser extent in the Deinococcus–Thermus, Gemmatimonadetes, Planctomycetes, δ-proteobacteria, Bacteroidetes, Firmicutes, TA06, and Rs-D42 (~ 4% each) (Supplementary Table S4). Phylogenetic analysis conducted on the 16S rRNA gene sequences collected after lactate addition showed that ≈ 76 % of the clones showed similarity ≥97% to known microorganisms (Figure 6). Most of these sequences were closely related to Sulfurovum aggregans (ε-proteobacteria), followed by several sequences phylogenetically related to Stenotrophomonas rhizophila (γ-proteobacteria), Rhizobium (α-proteobacteria), Desulfosarcina (δ-proteobacteria), Deinococcus geotermalis (Deinococcus-Thermus) and Aeribacillus pallidus (Firmicutes). Nevertheless, several sequences showed low similarity to known microorganisms and were phylogenetically related to uncultured Pseudomonas, Caenimonas members of β-proteobacteria, Bacteroidetes members close to Reichenbachiella lineage, uncultured Gemmatimonadetes, Planctomycetes and TA06. Sequences phylogenetically related to Chloroplast were also retrieved (Figure 6).