In the wake of the rapid advancement of industrialization and urbanization, environmental pollution problems have become increasingly severe. Specifically, organic pollutants and heavy metal contamination pose a huge threat to ecosystems and human health 1, 2, 3. Microbial remediation technology, which leverages the metabolic capabilities of microorganisms to degrade or transform pollutants, is recognized as a green pollution control technology due to its high efficiency, low cost, and environmental friendliness 4, 5, 6••. However, traditional remediation methods relying on single microbial strains encounter limitations in terms of efficiency and adaptability in complex environments. Moreover, bioremediation based on single strains is ill-suited to handle the mounting challenge of combined pollution (i.e. the co-existence of different pollutants), which hinders practical application of microbial remediation.

Artificially designed communities (DesComs) can remarkably boost degradation efficiency and environmental adaptability through metabolic interactions among strains 7, 8, 9, 10. Two types of DesComs are commonly used in bioremediation: engineered communities (EngComs) and synthetic communities (SynComs) 11, 12. EngComs are constructed using the ‘top-down’ approach. Specifically, carefully selected environmental variables (e.g. enrichment, artificial selection, or directed evolution) are used to drive existing communities (such as natural microbial communities) through ecological selection to carry out biodegradation processes. In contrast, SynComs are built using the ‘bottom-up’ approach, where isolated or engineered strains are assembled into communities to achieve a biodegradation function. In the context of bioremediation, pollutant-degrading strains are often sourced externally rather than being native to the contaminated soils. Therefore, a synergistic approach that combines both top-down and bottom-up strategies has been proposed for constructing SynComs based on the engineering of natural microbial communities 8, 13••. For example, the process begins by introducing pollutant-degrading strains into the contaminated soils to form EngComs. Subsequently, key keystone strains are identified and selected from these EngComs. Finally, SynComs are assembled by combining the pollutant — degrading strains with the carefully chosen keystone strains [13].

The metabolic diversity inherent in DesComs enables them to meet the needs of simultaneous remediation of multiple pollutants. Notably, the composition and structure of a microbial community often determine its metabolic activity [14], and the microbial communities in the natural environment often lack the optimal structure for efficient remediation. By meticulously designing and optimizing the structures and functions of SynComs, the pollutant degradation rate by SynComs can be significantly improved. Therefore, the design and construction of SynComs are essential prerequisites for achieving high efficiency of microbial remediation [15]. Consequently, SynComs and their design principles have become a research hotspot in bioremediation. In this perspective, we discuss the latest advances in understanding the mechanisms by which SynComs enhance biodegradation, as well as the prevalently employed strategies for SynCom design and construction. Moreover, we specifically review the application of metabolic models in SynCom design.

Two types of polluted ecosystems, laboratory-scale and field-like conditions, are commonly used to test the pollutant-mitigation ability of microbes. A bacterial strain or microbial consortium can efficiently degrade pollutants in the lab but may prove ineffective in real-world scenarios. Therefore, bioremediation involves both microbial degradation and deploying these biological catalysts in complex environmental settings, often requiring technologies for their introduction, distribution, and performance monitoring. Here, biodegradation refers to mitigation of contaminants under laboratory conditions, whereas bioremediation involves such mitigation in field-like conditions (Table S1).