Introduction

Currently, the total amount of municipal wastewater continues to increase each year with the intensification of human activities1. Amidst ongoing challenges posed by climate change2,3 and energy crises4,5, the conventional activated sludge (CAS) process, widely used for wastewater treatment, faces significant hurdles due to its high energy consumption and carbon emissions6. In contrast, the microalgal-bacterial granular sludge (MBGS) process presents a viable, energy-efficient, and environmentally sustainable alternative. This innovative method harnesses solar energy to produce oxygen7,8, aiding in the assimilation of organics and nutrients into biomass, thereby reducing carbon emissions9. Furthermore, the MBGS process can absorb industrial carbon dioxide (CO2)10 and methane (CH4)11, positioning it as an effective solution for achieving carbon neutrality in wastewater treatment12,13,14.

Understanding the removal pathways of nutrients and organics by MBGS is crucial, especially for complex organics, yet remains underexplored. Regarding nitrogen metabolism in non-aerated MBGS systems, most ammonia nitrogen (NH4+-N) is converted into glutamine or glutamate for microbial assimilation15. Both nitrate nitrogen (NO3--N) and nitrite nitrogen (NO2--N) are transformed to NH4+-N prior to metabolism16, while organic nitrogen, such as urea, is hydrolyzed to NH4+-N via intracellular urea amidolyase17. Phosphorus is removed through microbial assimilation and poly-phosphate accumulation9, with phosphate being transported into cells by oxidative phosphorylation and then removed by poly-phosphate kinase catalysis15.

However, significant knowledge gaps persist in our understanding of the metabolic processes for complex organics, particularly concerning the degradation mechanisms of substances like starch, glycerol and fatty acid found in municipal wastewater. Laboratory-scale studies often use simple carbon sources such as acetate and glucose15,18, whereas real municipal wastewater contains complex organics19, whose degradation mechanisms remain unclear. Considering the natural occurrence of algae-bacteria symbiosis in aquatic ecosystems20 and the influx of complex organics due to human activities21,22, further exploration of this symbiosis in organic degradation is essential. Since carbon metabolism is closely linked to greenhouse gas emissions23 and organic removal is often tied to nutrient removal24, understanding the mechanisms behind microalgal-bacterial symbiosis in the removal of complex organics is urgent.

Moreover, the specifics of cooperative interactions between algae and bacteria within the MBGS system remain poorly defined, which hinders the further application of this technology. It has been established that the critical aspect of algae-bacteria interactions involves the exchange of carbon dioxide (CO2) and oxygen (O2), essential for the efficient removal of organics by the MBGS process9. While previous studies have primarily focused on identifying microbial community components and exploring bacterial coherence, there has been insufficient attention to the dynamics of algae-bacteria interactions within this relationship25,26.

This study aims to uncover the removal pathways of complex organics and the dynamics of algae-bacteria interactions within the MBGS system. We have evaluated the performance of the MBGS process, investigated upregulated functional genes under diurnal fluctuations, tracked the evolution of the microbial community, and clarified the interactions between algae and bacteria within the system. Our research enhances the understanding of the mechanisms behind complex organics removal facilitated by algae-bacteria interactions in the MBGS process, contributing essential knowledge for its engineering applications. More importantly, this study advances our understanding of carbon cycling in natural aquatic ecosystems, mediated by microalgal-bacterial symbiosis.

Results and discussion

Performance of complex organics removal by MBGS

As depicted in Supplementary Figs. 1–3, the MBGS effectively removed complex organics, achieving a chemical oxygen demand (COD) removal efficiency of 82.4% when treating real municipal wastewater with a flat-plate reactor during the daytime on day 12. The carbon-to-nitrogen (C/N) ratio of the influent significantly influenced the removal efficiencies of COD and total nitrogen (TN) (Fig. 1a), which increased as the C/N ratio rose from 1 to 5 and then stabilized at ratios above 5 (Fig. 1b). In fact, the empirical formula C100H158O38N17P suggests an optimal C/N ratio of approximately 59,27, which facilitated pollutant removal predominantly through microbial assimilation9,24. However, notable fluctuations in the C/N ratio, particularly reductions to around 1 during the experiment’s final 10 days (Supplementary Table 1), adversely affected MBGS performance. Moreover, the average dissolved oxygen (DO) concentrations during nighttime positively impacted COD removal (Fig. 1c, d). DO levels exceeded 21 mg/L at 20:00, as shown in Supplementary Table 2, facilitating organic removal during the night28, which is attributed to the utilization of O2 produced by microalgae during the day by bacteria at night24. Additionally, the optimal granule size for MBGS has been identified as approximately 1.2 to 1.3 mm at 200 µ mol·m−2·s−1 light intensity7, suggesting that further optimization of granule size could improve organics removal.

Fig. 1: Impact of carbon-to-nitrogen (C/N) ratio, carbon-to-phosphorus (C/P) ratio, and dissolved oxygen (DO) concentrations on pollutant removal efficiencies.
figure 1

Correlation heat maps of the correlation between removal efficiencies of COD, TN, and TP with C/N ratio, C/P ratio, and DO concentrations during daytime (a) and nighttime (c). b An exponential fitting plot of removal efficiencies of COD and TN versus C/N ratios during the day. (d) A linear fitting plot of removal efficiency of COD at night versus DO concentrations. Notes: Symbols indicate levels of statistical significance: * for p < 0.05, ** for p < 0.01, *** for p < 0.001.

Full size image

Furthermore, MBGS exhibited superior settling performance in flat-plate bioreactors, evidenced by a lower 5-min sludge volume index (SVI5) value and higher density compared to inclined-plate bioreactors, as shown in Supplementary Fig. 4. This improved performance is attributed to the presence of complex organics in real municipal wastewater, which supported granule growth and settling performance more effectively than synthetic wastewater28.

Organics removal pathways

Research29 identified the main organic components in real municipal wastewater as carbohydrate, fat and oil, consisting of glycerol and fatty acid30, with carbohydrate primarily in the form of starch19. Metabolic pathways shown in Fig. 2a illustrate how starch, glycerol, and fatty acid were metabolized into α-D-glucose-1P, glycerone-P, and acetyl CoA through starch and sucrose metabolism (ko00500), glycerolipid metabolism (ko00561), and fatty acid degradation (ko00071) pathways, respectively. These intermediates were then channeled into the tricarboxylic acid (TCA) cycle to release CO231, which was absorbed by microalgae (Cyanobacteria and Chlorophyta) though the Calvin cycle and converted into glyceraldehyde-3P, subsequently synthesized into biomass32.

Fig. 2: The organics metabolism pathways and functional genes of MBGS process.
figure 2

a The metabolic pathways of starch, glycerol, fatty acid, and photosynthesis. The abundance, expression (b), and expression ratios (c) of functional genes related to organics metabolism and photosynthesis. Notes: SW1: Synthetic wastewater in a flat-plate bioreactor, RW1: Real municipal wastewater in a flat-plate bioreactor, SW2: Synthetic wastewater in an inclined-plate bioreactor, RW2: Real municipal wastewater in an inclined-plate bioreactor.

Full size image

During the experiment, the relative abundance of organics metabolism pathways became increasingly prominent, with glycolysis process particularly active, accounting for the highest proportion (0.86%) in the real municipal wastewater of inclined-plate reactors by the end of experiment (Supplementary Fig. 5a). Further analysis by CAZy annotation revealed the abundance of glycosyl transferases, glycoside hydrolases and carbohydrate-binding modules involved in the carbohydrate degradation9, with glycoside hydrolases playing a crucial role in decomposing complex organics (Supplementary Fig. 5b).

Functional gene analysis indicated that genes encoding 15 enzymes involved in organics metabolism and 5 enzymes related to photosynthesis were upregulated in treating real municipal wastewater compared to synthetic wastewater (Supplementary Table 3). Genes such as gapA (K00134), pdhD (K00382), pps (K01007), petC (K02636), petH (K02641), psbA (K02703), and psbD (K02706) were abundant (Fig. 2b), while gldA (K00005), gap2 (K00150), amyA (K01176), gpmA (K01834), petH, psbA, and psbD exhibited robust expression activity (Fig. 2c). Notably, gapA, pdhD, pps, gap2, and gpmA were involved in glycolysis process, whereas petH was linked to photosynthetic electron transport, enhancing CO2 fixation by boosting NADPH synthesis33. These suggest that MBGS could decompose organics to CO2 via aerobic heterotrophic bacteria and fix CO2 into microalgae biomass by upregulating functional genes involved in glycolysis and NADPH synthesis, respectively. Additionally, the upregulation of amyA in starch metabolism and gldA in glycerolipid metabolism enhanced the decomposition of starch and glycerol, respectively. Simultaneously, upregulated psbA and psbD genes of photosynthesis facilitated the production of electrons from H2O, which are used to reduce NADP+ to NADPH, thereby producing O234. The integration of these metabolic pathways is essential for the efficient conversion of complex organics into valuable biomass, highlighting the potential of MBGS in sustainable wastewater treatment.

Microbial community evolution

Supplementary Fig. 6a illustrates that real municipal wastewater significantly promoted the growth of Cyanobacteria, which, along with Desulfobacterota I, formed symbiotic relationships within the outer layers of MBGS (Supplementary Fig. 6b). Figure 3a indicates that the dominant microbial phyla within the MBGS were Proteobacteria, Bacteroidota, and Cyanobacteria, which constituted 51.3%, 10.6%, and 6.2% of the microbial community, respectively. The primary microbial classes included Alphaproteobacteria, Gammaproteobacteria, Bacteroidia, and Cyanobacteriia, with proportions of 28.3%, 23.0%, 10.1% and 6.3%, respectively (Fig. 3b).

Fig. 3: Community structure succession of MBGS with different wastewater and reactor.
figure 3

The circos maps of microbial community structure of MBGS at phylum (a) and class (b) level based on the metagenomic sequencing (top 20). The clustering heatmaps of microbial community structure of MBGS at phylum (c) and class (d) level based on the metatranscriptomics sequencing (top 20). Notes: SW1: Synthetic wastewater in a flat-plate bioreactor, RW1: Real municipal wastewater in a flat-plate bioreactor, SW2: Synthetic wastewater in an inclined-plate bioreactor, RW2: Real municipal wastewater in an inclined-plate bioreactor.

Full size image

The influence of real municipal wastewater was particularly notable in enhancing the growth of Chloroflexota (Anaerolineae), Acidobacteriota (Acidobacteriae), Verrucomimicrobiota (Verrucomicrobiae), and Sumerlaeota (Sumerlaeia), as demonstrated through cluster analysis (Fig. 3c, d). Proteobacteria are shown to play a crucial role in organics and nutrient removal by facilitating the decomposition of carbohydrates and synthesis of enzymes15. Cyanobacteria contribute to the absorption of nitrogen and phosphorus from wastewater35. Other microbial groups such as Bacteroidota36, Chloroflexota37, Verrucomicrobiota38, Actinobacteriota39, Acidobacteriota40, Desulfobacterota I41, Armatimonadota42, and Sumerlaeota43 are also involved in complex organics degradation. The elevated relative abundance of above bacteria in real municipal wastewater compared to synthetic wastewater (Supplementary Fig. 7) suggests their effective adaptation to complex organic substrates and highlights their potential roles in degradation processes. Moreover, the Alpha diversity index indicates that the MBGS in real municipal wastewater has higher species richness and diversity compared to synthetic wastewater (Supplementary Table 4). Additionally, real municipal wastewater was found to upregulate gene expression in Cyanobacteria (Cyanobacteriia) and Chlorophyta (Chlorophyceae) (Fig. 4), which is beneficial for providing the oxygen necessary for bacterial degradation of complex organics.

Fig. 4: Gene expression of dominant microorganisms in different wastewater and reactor.
figure 4

The gene expression ratios (cDNA/DNA ratios) of dominant microorganisms at the phylum (a) and class level (b). Notes: SW1: Synthetic wastewater in a flat-plate bioreactor, RW1: Real municipal wastewater in a flat-plate bioreactor, SW2: Synthetic wastewater in an inclined-plate bioreactor, RW2: Real municipal wastewater in an inclined-plate bioreactor.

Full size image

Algae-bacteria interactions of MBGS

The contributions of microorganisms to wastewater treatment significantly depend on their metabolic activity rather than their relative abundance44. Utilizing metatranscriptomics sequencing, co-occurrence networks of the top 20 dominant microorganisms in MBGS at both the phylum and class levels were analyzed, revealing intricate interactions that influence the system’s performance. These networks, depicted in Fig. 5, consist of 20 nodes forming three modules (Modules 1, 2, and 3) at the phylum level and two modules (Modules 1 and 2) at the class level, organized based on positive correlations between the microorganisms

Fig. 5: Interactions of algae and bacteria in MBGS.
figure 5

The co-occurrence networks between dominant microorganisms (top 20) of MBGS at phylum (a) and class (b) level. Notes: Nodes represented each dominant phylum or class, and the connections between nodes indicated a correlation between the two microorganisms. The red line indicated a positive correlation, while the blue line indicated a negative correlation. The size of the node was directly proportional to the relative abundance of the microorganisms in the entire sample. The edges thickness was proportional to the relevance degree.

Full size image

Notably, Cyanobacteria displayed a significant positive correlation with Desulfobacterota I45 (relevance degree of 0.66), indicating a synergistic interaction crucial for the function of MBGS. Similarly, Chlorophyceae showed a strong positive correlation with Verrucomicrobiae46 (relevance degree of 0.86). The symbiotic relationships between Chlorophyceae with Sumerlaeia, Chloroflexia47, Acidobacteriae48, and Fimbriimonadia were particularly strengthened in real municipal wastewater, as evidenced by enhanced gene expression at the end of the experiment (Fig. 4b). Additionally, competitive relationships existed, such as between Bacteroidota and Sumerlaeota (relevance degree of −0.94), due to overlapping metabolic functions in complex organics decomposition49.

Overall, the symbiotic relationships, particularly between Chlorophyceae and other groups such as Verrucomicrobiae, Sumerlaeia, Chloroflexia, Acidobacteriae and Fimbriimonadia, along with the interaction between Cyanobacteria and Desulfobacterota I, are provital for the effective decomposition of complex organic matter. The effective exchange of O2 and CO2 between algae and bacteria within these interactions is vital for enhancing organics removal efficiency in the MBGS process9.

Relationships between microbial community and functional genes

Figure 6 highlights the relationships between functional genes and microbial communities in MBGS. The genes of gap2 and gpmA, as the main functional genes with high expression activity in the glycolysis process50, the expression of the former was notably correlated with Acidobacteriae (relevance degree of 0.71), while the expression of the latter was strongly associated with Sumerlaeia and Fimbriimonadia (relevance degree of 0.86 and 0.79, respectively). Additionally, the amyA gene, active in starch metabolism and crucial for labile-C degradation51, correlated exclusively with Actinobacteriota52 (relevance degree of 0.71). The gldA gene, a key gene for glycerolipid metabolism, is carried by Bacteroidota53, and its expression was closely linked to Chloroflexia and Verrucomicrobiae (relevance degree of 0.86 and 0.83, respectively). Furthermore, the expression of the fadE gene, essential in fatty acid degradation54, was closely correlated with Desulfobacterota I (relevance degree of 0.71). Overall, Acidobacteriae, Sumerlaeia, and Fimbriimonadia played significant roles in enhancing the glycolysis process. Actinobacteriota was pivotal in starch decomposition55, while Chloroflexia and Verrucomicrobiae were instrumental in glycerol decomposition. Desulfobacterota I emerged as a key bacterium for fatty acid degradation.

Fig. 6: Relationship between microbial communities and functional genes in MBGS.
figure 6

The co-occurrence networks between microbial communities and functional genes of MBGS at phylum (a) and class (b) level. Notes: Nodes represented each phylum, class or genes, and the connections between nodes indicated a correlation between microorganism and gene. The size of the node was directly proportional to the relative abundance of microorganism or function gene. The line indicated a positive correlation and the edges thickness was proportional to the relevance degree.

Full size image

The above interactions suggest that the removal mechanisms of complex organics in MBGS involved several interconnected pathways, as illustrated in Fig. 7. The upregulation of the gap2 and gpmA genes in the glycolysis facilitated the conversion of complex organics into CO2, mediated by Chlorophyceae with Acidobacteriae, Sumerlaeia, and Fimbriimonadia. The upregulated petH gene, involved in the NADPH synthesis by Cyanobacteria, played a crucial role in enhancing the fixation of CO2 into biomass. Concurrently, the amyA gene related to starch metabolism was upregulated in Actinobacteriota, improving starch decomposition. Similarly, the upregulated gldA gene in glycerolipid metabolism, influenced by the interactions between Chlorophyceae and Chloroflexia/Verrucomicrobiae, enhanced glycerol decomposition. Fatty acid was effectively decomposed by Cyanobacteria and Desulfobacterota I. Lastly, the upregulated expression of psbA and psbD genes carried by Chlorophyceae56, provided ample O2 to support the metabolic activities of the associated aerobic heterotrophic bacteria, thus facilitating the overall organics degradation.

Fig. 7: Organics removal mechanisms by MBGS process.
figure 7

The genetically regulated breakdown of starch, glycerol, and fatty acid by algae and bacteria into CO2, which then enters the Calvin cycle for biomass synthesis. Notes: Blue dotted box indicates the upregulated process, the blue ellipse indicates the upregulated functional gene, and red upward arrow indicates microorganism with up-regulated expression. Green box indicates bacteria that have a cooperative relationship with Chlorophyceae, while blue box indicates bacteria that have a cooperative relationship with Cyanobacteria.

Full size image

Implication

As a green process, understanding the pathways for complex organics removal in the MBGS process is crucial. To date, studies have predominantly focused on simpler carbon sources such as acetate and glucose15,18, leaving the mechanisms for complex organics less understood. This gap hinders the engineering applications of the MBGS process. Furthermore, while microalgal-bacterial symbioses ubiquitously exist in natural water that often contain complex organics from human activities57, the degradation mechanisms of these substances have not yet attracted attention, leading to an incomplete understanding of the biogeochemical carbon cycle.

In this study, the MBGS process has been shown to effectively convert complex organics into biomass during day-night cycles, facilitated by algae-bacteria interactions, thereby enriching organics from wastewater for resource recovery. However, achieving stable performance has been challenging due to suboptimal granule size and an imbalanced C/N ratio in the wastewater. Future studies will aim to optimize reaction conditions, such as adjusting stirring to regulate granule size, given its positive impact on photosynthetic efficiency and granule characteristics58, and adding inorganic carbon to enhance the C/N ratio of the influent, as MBGS can utilize additional CO2 to improve pollutant removal performance10. In addition, forthcoming research will delve deeper into the collaborative mechanisms within microalgal-bacterial symbioses, including the analysis of metabolites such as signaling molecules.

Materials and methods

Raw and synthetic wastewater

Raw wastewater was collected from the sewer manholes at a university’s canteen and dormitory in Wuhan, and from a wastewater treatment plant in Hangzhou. For comparative analysis, synthetic wastewater compositions are described in Supplementary Table 5. Variations in pollutant concentrations and pH levels of influent are detailed in Supplementary Table 6. The average water temperature during the experiment was maintained at approximately 24.5 °C.

Experimental procedure

MBGS was sourced from prior research16, characterized by a mean size of 4.98 mm, density of 1.04 g/cm3, and an SVI5 of 40.20 mL/g. The experimental setup included two identical continuous-flow flat-plate bioreactors with vertical baffles and two identical inclined-plate bioreactors at a 45° angle (Supplementary Table 7). These bioreactors were operated under LED lights (MBTL-T8-18, Hangzhou Mobate Biotechnology Co., Ltd., China) at a light intensity of 180 µmol·m²·s¹ with a 12-hour light/dark cycle (Supplementary Fig. 8). The initial volatile suspended solids (VSS) concentrations of MBGS in the continuous-flowing bioreactors were approximately 2.6 g/L, with a hydraulic retention time (HRT) maintained at 12 h. Water samples were collected daily at 8:00 and 20:00, then filtered through 0.45-µm filters for subsequent analysis. Water temperature and pH values of effluent throughout the experiment are presented in Supplementary Fig. 9.

Analytical methods

Measurements for COD, TN, NH4+-N, NO3--N, NO2--N, TP, SVI5, and VSS followed the standard methods59. Granule size and density of MBGS were determined according to established protocols11,60. Granule morphology and microbial compositions were observed using a fluorescence microscope (Rx50, Ningbo Shunyu Instrument Co., Ltd, China) and a scanning electron microscope (Gemini 500, Zeiss, Germany). The instruments used for measuring light intensity, DO concentration, water temperature, and pH value were consistent with those described in a previous study28. Statistical analyses and data fitting methods were as previously described10.

Metagenomic and metatranscriptomic sequencing analysis

Total DNA was extracted using the Mag-Bind Soil DNA Kit (M5636-02, Omega Bio-tek, USA), with concentrations assessed by the Quantifluor-ST fluorometer by measuring absorbance at 260 nm and 280 nm (E6090, Promega, USA) and the Quant-IT PicoGreen dsDNA detection kit (P7589, Invitrogen, USA). DNA quality was then evaluated through 1% agarose gel electrophoresis. Total RNA was extracted using the RNA PowerSoil® Total RNA Isolation Kit (12866-25, MoBio, USA). The extracted RNA was assessed for integrity on a 1.5% agarose gel. RNA concentrations were quantified using a UV spectrophotometer (NC-2000, Thermo Scientific, USA), with a required minimum concentration of 50 ng/μl, and RNA integrity number (RIN) of ≥5.5, as measured by a biological analyzer (2100 Bioanalyzer, Agilent, USA).

Sequencing analysis of MBGS was performed using the Illumina NovaSeq sequencing platform. Total metagenomic DNA and cDNA synthesized from mRNA via metatranscriptomics were fragmented using the Whole Genome Shotgun (WGS) strategy. Fragment libraries of appropriate lengths were constructed and sequenced from both ends (paired-end, PE). Species annotation, sequence splicing, gene prediction and functional annotation of metagenomic and metatranscriptomic sequencing were detailed as described61,62. Analyses of pollutant metabolism pathways and the relative abundance of CAZy classes in carbon metabolism were conducted using the KEGG and CAZy databases, respectively.

Co-occurrence networks analysis

Based on the metaranscriptomics sequencing data from MBGS in four bioreactors, Sparcc rank correlation coefficients between microorganisms, as well as between microbial communities and functional genes, were calculated using the software Mothur63. Co-occurrence networks were then constructed for correlations where r > 0.6 and p < 0.05. The co-occurrence networks, derived from metatranscriptomic sequencing, utilized the concept of degree to describe the interactions between algae and bacteria, as well as between microbial communities and functional genes64,65. The co-occurrence networks were subsequently imported into Gephi 0.9.2 (http://gephi.github.io/) for visual representation.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.