3.1. Synchronous dewatered sludge disintegration and protein recovery
The protein concentrations in supernatant after alkaline-thermal (AT) pretreatment and enzyme lysis (EL) were 4171.25 ± 23.75 mg/L (148.71 mg/ g VSS) and 3400.71 ± 43.15 mg/L (121.23 mg/ g VSS), respectively (Fig. S2), This was significantly higher than the concentration of 108.24 ± 0.67 mg/L (3.86 mg/ g VSS) in untreated DS, indicating that these two methods fulfilled effectively synchronous sludge disintegration and protein extraction. AT pretreatment led to a higher release in the protein concentration compared with EL. When the pH adjusted to 3, the protein precipitated and the concentrations decreased to 1158.12 ± 16.87 mg/L (41.29 mg/ g VSS) in the AT-treated sludge, which was lower than that in the EL (2054.28 ± 78.3 mg/L). AT pretreatment presented a higher protein recovery (72.23%) compared to EL (39.59%), thus it was chosen to recover soluble proteins from the DS followed by electro-fermentation of residual component after protein recovery.
The concentrations of heavy metals in DS and recovered protein by AT pretreatment were further measured (Table S4). Through the protein recovery process, 59.77 − 63.77% of Al, Ca, and Pb, and 88.65 − 90.12% of Cu, Cr and Ni were removed. While Fe and Zn were not accumulated in recovered pellet. The concentrations of some toxic heavy metals (Pb, AS, Cd, Hg, and Se) were below the limit of the hygienic standard for feeds (GB 13078 − 2001) (Table S4). In addition, the recovered protein was proved to support Pseudomonas aeruginosa PAO1 growth as yeast extract and peptone substitute, with a comparable cell growth amount in RP-1000, RPY-3700, and RPP-1800 compared to those in LB medium (Figure S3).
3.2. EFS facilitates degradation of organic matter
For the TS treatment, EFS fed with TS (EFS-TS) achieved a 28.80% increase in the SCOD degradation efficiency during the first batch compared to the open-circuit EFS fed with TS (AD-TS) (51.44 ± 2.59%) (Fig. 2A). Interestingly, the SCOD degradation rate of EFS-TS in the first 7 days (246.04 mg COD/L-day) was 7.36-fold higher than that in AD-TS (one-way ANOVA, p < 0.05, n = 3), suggesting that EFS accelerated organics degradation. The EFS-TS obtained higher TCOD removal of 75.91 ± 1.15% compared with 71.35 ± 1.11% in the AD-TS (Fig. 2B). Polysaccharide and protein degradation were faster in the EFS-TS than in the AD-TS during the first 5 − 9 days (Fig. 2C and D). Polysaccharide removal efficiency accounted for 51.41 − 61.24% in the EFS-TS and AD-TS (Figure S5A), while protein concentration showed a slight increase and ended up no significantly difference in the first batch (one-way ANOVA, p < 0.05, n = 3). In the second batch, the tendency of COD removal was similar to that of the first batch (Figure S4A and B), while the protein and polysaccharide concentrations dropped by 48.15 ± 0.34% and 35.66 ± 3.44% in the EFS-TS, respectively, which was significantly enhanced compared with those in the AD-TS (39.9 ± 3.15% and 5.97 ± 2.73%) (one-way ANOVA, p < 0.05, n = 3) (Figure S5B).
For the RS treatment, the SCOD of EFS-RS and AD-RS fluctuated around about 800 mg/L to 1300 mg/L during the first batch and remained relatively constant, indicating that RS released more soluble organics. Furthermore, the TCOD removal of EFS-RS (21.33 ± 1.99%) was higher than that of the AD-RS. The EFS-RS showed 1.35 times higher polysaccharide degradation efficiency (31.51 ± 6.65%) than the AD-RS (Fig. 2C and D). Although the protein concentration increased, the increment of EFS-RS (74.14 ± 4.26%) was lower than that in AD-RS (91.29 ± 5.94%), suggesting more protein degradation in EFS-RS (Figure S5A). The tendency of organics removal in the second batch was similar to that of the first batch (Figure S4A and B). All EFSs exhibited significantly enhanced TCOD removal, faster protein and polysaccharide degradation compared to AD (one-way ANOVA, p < 0.05, n = 3). These results demonstrated that EFS accelerated disintegration and hydrolysis of WAS.
The VFAs concentration of TS increased after AT pretreatment, while acetate was the predominant SCFA detected in all reactors (Fig. 3A and B). During the EFS fed with TS, the concentration of acetate decreased sharply in the first 7 days and then gradually reduced to 109.60 ± 6.81 mg/L, while that of AD-TS accumulated in the first 11 days and slightly decreased to 973.60 ± 20.64 mg/L afterwards. The VFAs concentration of EFS/AD-RS was much lower than EFS/AD-TS, but the variation exhibited a similar trend in that the acetate degraded faster in the EFS compared with the AD (Fig. 3C and D).
3.3. EFS enhanced ammonium recovery
The hydrolysis and acidogenesis of protein-rich sludge released a large amount of ammonium into the supernatant (Fig. 4A). With the current flow in two-chamber of EFS, cations migrated from the anode chamber to the cathode chamber, resulting in ammonium enriching in the catholyte. The NH4+-N concentration in the anode chamber of EFS-TS decreased to 100.00 ± 5.56 mg-N/L after 17 days of operation in the first batch, while it increased to 248.80 ± 28.11 mg-N/L in the catholyte (Fig. 4A and B) achieving a 71.33% ammonium recovery from sludge digestate. By contrast, the NH4+-N concentration in sludge suspension in the AD-TS increased to 326.85 ± 6.25 mg-N/L after 17 days (Fig. 4A). Meanwhile, there was still 115.74 ± 8.64 mg-N/L of NH4+-N in the catholyte of AD-TS owing to the diffusion of concentration gradient. The second batch of operation showed similar results, with 299.07 ± 10.98 mg-N/L of NH4+-N migrating to the catholyte of EFS-TS, leading to 79.63% ammonium recovery efficiency (Figure S7).
Less NH4+-N migrated to the cathode chamber of the EFS-RS due to lower current and NH4+-N release (Figure S7D). NH4+-N of 44.97% in the EFS-RS was migrated to the catholyte during the first batch, higher than that in the AD-RS (26.89%) (Fig. 4). During the second batch, the NH4+-N concentrations of the EFS-RS and AD-RS increased to 260.65 ± 6.69 mg-N/L and 194.72 ± 19.78 mg-N/L in the cathode chamber, respectively. Hence, the EFS-RS achieved 49.88% ammonium recovery from the fermented sludge liquor (Figure S7). In addition, proton migration in the EFS led to a decrease of pH at the anode (Figure S8A), while proton consumption and hydroxyl ions (OH−) production at the cathode led to an increase in pH (Figure S8B).
3.4. EFS shaped microbiome of biofilm and suspension
The Pielou evenness, Shannon, and Simpson indices of the suspension samples were higher, indicating higher community diversity and more evenness distribution. The Chao1 estimator showed that the low abundance species in the suspension samples was higher than in the anode biofilms (Table S3). Principal coordinates analysis (PCoA) revealed a distinct difference in microbial community structure between the RS and TS, and electrode biofilm and suspension. However, the suspension of the EFS and AD had similar bacterial composition (Fig. 5Aand B). This suggests that EFS shaped microbial community structure.
Bacteroidota, Desulfobacterota, Firmicutes, and Proteobacteria were the foremost predominant phyla in all samples (Fig. 5C). The relative abundance of Bacteroidota decreased from 35.75 − 54.83% in suspension samples (AD-RS, AD-TS, EFS-RS-ss, and EFS-TS-ss) to 24.09 − 26.25% in the anode biofilm of EFS-TS (EFS-TS-bio) and EFS-RS (EFS-RS-bio). Conversely, the poised anode potential promoted the enrichment of Desulfobacterota which relative abundance accounted for 54.86% in the EFS-RS-bio and 44.20% in the EFS-TS-bio. In the EFS-TS-bio, Geobacter (38.80%), Hydrogenophaga (3.62%), and Petrimonas (9.32%) were enriched, while the EFS-TS-ss and AD-TS shared the same predominant genera (Rikenella, Lentimicrobium, Petrimonas and Azonexus). The relative abundance of Lentimicrobium, Petrimonas, and Azonexus increased by 31.70%, 43.55%, and 673.89%, respectively, in the EFS-TS-ss compared to AD-TS (Fig. 5D). In the EFS-RS-bio, Geobacter (51.27%), Lentimicrobium (12.00%), and Leptolinea (1.54%) were enriched compared with AD-RS, which had Clostridium (11.76%), Smithella (8.81%), Rikenella (5.18%), and Petrimonas (4.95%) as its predominant genera. The relative abundance of Lentimicrobium (20.96%), Rikenella (11.82%), and Leptolinea (2.55%) increased in the EFS-RS-ss compared with AD-RS.
The clustering analysis heatmap showed a clear distinction in archaeal communities between biofilm and suspension samples (Fig. 5E, Figure S9). The predominant phyla in all reactors belonged to Euryarchaeota, Candidatus Bathyarchaeota, and Candidatus Lokiarchaeota. Methanosarcina and Methanothrix, the acetoclastic and methylotrophic methanogens, respectively, dominated all communities. While the relative abundance of most methanogens reduced in the EFS-TS-bio and EFS-RS-bio compared to that in the AD-TS/RS, which is consistent with therir lower methane production (Figure S6). For example, the relative abundance of Methanosarcina decreased by 71.23% in the EFS-TS-bio compared with the AD-TS, while that of Methanothrix decreased by 62.02% in the EFS-RS-bio compared with the AD-RS.
3.5. Difference in functional gene enrichment between EFS and AD
The metagenomic analysis showed that Geobacter (46.94 − 48.51%) was the most predominant genera in the biofilm of the EFS (Figure S10). COG categories illustrated that the genes involved in signal transduction mechanisms (T) had 1.28 and 1.07 times higher relative abundances in the EFS-TS-bio and EFS-RS-bio, respectively, than those of AD (Figure S11). Clustering heatmap of KEGG also showed the shift of metabolic pathways between the EFS and AD reactors (Figure S12A), such as the increased abundances of fatty acids biosynthesis, TCA cycle, protein export, and two-component system encoding genes in the EFS biofilm. Furthermore, the abundance of putative transposase encoding gene related to replication and repair increased 5.62-fold in the EFS-TS-bio than that in the AD-TS (Figure S12B). Besides, genes encoding methyl-accepting chemotaxis protein, putative ABC transport system permease protein, and branched-chain amino acid transport system substrate-binding protein were identified to be increased in the EFS-TS-bio and EFS-RS-bio compared to the AD.
3.6. EFS enriched genes involved in carbon and nitrogen transformation
The critical pathways for sludge degradation include polysaccharides degradation, monosaccharide metabolism, and protein hydrolysis. Compared with the AD, the abundant genes involved in starch and sucrose metabolism pathways enriched in the EFS-TS and the EFS-RS, including alpha-amylase (treS), glucoamylase (gla1), and 4-alpha-glucanotransferase (malQ), especially for treS which abundance increased by 1.31 times in the EFS-TS-bio vs the AD-TS (Fig. 6A and B). Additionally, gene encoding fructose-bisphosphate aldolase (fbaA) exhibited 47.46% higher abundance in the EFS-TS-bio than the AD-TS. The genes involved in protein hydrolysis enzymes, including serine endopeptidases (gluP and degQ) and cysteine peptidases (pfpI) showed 1.67 − 5.17-fold higher abundance in the biofilms of the EFS-TS/RS vs the AD (Fig. 6C and D). The results that EFS-TS and EFS-RS increased the abundance of the genes involved in polysaccharides and protein metabolism were consistent with their higher polysaccharides and protein degradation efficiency, which provided a reasonable explanation for the enhancement of COD removal (Fig. 2 and S4).
Additionally, key genes involved in glycolysis pathway were identified to be more abundant in the EFS than AD, especially in the EFS-TS-bio, including hk, tal-pgi, pfkABC, fbaAB, tpiA, gapAN, gapor, pgk, apgM, eno, korAB, and porABDG, which catalyzed the hydrolysis of D-glucose to α-D-glucose-6P, followed by β-D-fructcose-6P oxidation to glyceraldehyde-3P and further to phosphoenol-pyruvate, and pyruvate conversion to acetyl-CoA (Fig. 7A). Among them, the gapor exhibited the highest increased abundance of 4.98-fold in the EFS-TS-bio vs the AD-TS. The relative abundance of gene encoded acetate kinase (ackA) increased to over 0.04% in the biofilm of EFS-TS/RS, whereas its abundance in the AD accounted for about 0.02%. Additionally, the abundance of gene encoded acetyl phosphotransferase (pta) accounted to 0.02% in the EFS-TS-bio and EFS-RS-bio while decreased to 0.01% in the AD (Fig. 7B). Furthermore, the genes involved in acetate oxidation had higher abundance in the EFS than those in the AD, suggesting that the capacity for acetate degradation was enhanced in the EFS.
The electrode biofilm and suspension hosted various pathways for nitrogen transformation (Fig. 7C). The abundance of genes encoding glutamate dehydrogenase (gdhA), hydroxylamine reductase (hcp), nitrogenase delta subunit (anfG), nitrogenase molybdenum-iron protein (nifDHK), and nitrite reductase (nrfAH), which catalyzed ammonium formation from L-glutamate, hydroxylamine, nitrogen, and nitrite, were higher in the EFS-TS-bio and EFS-RS-bio than in the AD, consistent with high ammonium recovery in EFS. The nifDHK and nrfAH genes showed 2.46 − 3.90-fold higher abundances in the anode biofilms than those of AD. On the other side, the abundance of genes encoding nitrite reductase (nirABD), which involved in the generation of ammonium from nitrite, was higher in the suspension vs the anode biofilm. In addition, the microorganism in suspension samples possessed more genes encoding ammonia monooxygenase (amoABC), hydroxylamine dehydrogenase (hao), periplasmic dissimilatory (napA), assimilatory nitrate reductase (nasA), membrane-bound nitrate reductase (narGH), nitrite reductase (nirKS), and nitrate/nitrite transporter (nrtP).
3.7. Abundant genes associated electron transfer in EFS
Genes related to electron transfer were abundantly enriched in electrode biofilm and suspension of EFS, but rarely detected in AD samples (Fig. 7B). Genes encoded cytochrome c peroxidase (macA) and cytochrome c-type biogenesis protein (ccdA) more enriched in electrode biofilm than in suspensions, which functioned in EET of Geobacter. The relative abundance of macA increased to 0.21 − 0.29%, while the abundance of ccdA increased to 0.25 − 0.28% in the electrode biofilm of EFS. The fixABCX associated electron transfer flavoprotein (ETF): quinone oxidoreductase complexes, pyrDII encoded dihydroorotate dehydrogenase electron transfer subunit, along with nuoA-N, rnfA-EG, and hdrA-E associated ETF-independent redox complexes were more abundant in the EFS-TS-bio and EFS-RS-bio than those in suspensions. In addition, the electron transport complex protein associated genes (rnfA-EG) also showed higher abundance in the suspentions of EFS compared with the AD. These results confirmed the promotion of EET pathways in EFS.