In natural and engineered ecosystems, diverse species interact in complex ways to form highly efficient microecologies. One key orchestrator of these interactions is autoinducer-2 (AI-2), a signaling molecule that plays a crucial role in microbial community assembly, metabolic flux, and resilience to environmental disturbances. This review provides the systematic synthesis of AI-2’s dual structural dynamics (S-THMF-borate/R-THMF interconversion) and its context-dependent roles in mediating bacterial crosstalk. It also reveals the receptor diversity (such as LuxP and LsrB) of AI-2 in bacterial kingdom and the signal transduction mechanism. Systematically elaborated on AI-2’s regulation of cellular metabolic flux and its ability to autonomously exhibit a series of coordinated behaviors in response to environmental changes. The review explores the ramifications of AI-2 on bacterial community interactions in synthetic biology and natural ecosystems. The wide application of AI-2-mediated interspecific communication in various fields including host health, agriculture, industry and environmental ecology has also been widely discussed. Factors influencing AI-2 production are thoroughly examined, including internal factors such as strain specificity, cell density, growth form and the phenotypic heterogeneity. Additionally, external biological factors (such as nutritional status and environmental stress) and abiotic factors (aggregation, diffusion, and flow) are discussed in detail. By examining knowledge gaps in AI-2-mediated spatial heterogeneity and multi-QS system coordination, this work charts a roadmap for harnessing microbial communication in chemical engineering and environmental sustainability.

The escalating crisis of antimicrobial resistance poses a devastating and immediate threat to human life. Antimicrobial peptides (AMP) are a promising antibiotic substitute to combat antimicrobial resistance. Compared with the traditional wet-lab screening approaches, computational models have largely improved the efficiency of predicting antimicrobial peptide. However, most computational models overlook or underutilize the evolution and structural information of peptides, which is crucial for understanding the peptide functions. Here, we proposed a sophisticated deep learning model to predict AMPs, Antimicrobial Peptide Bilinear Attention Network (AMPBAN), which incorporates peptide evolution features from ESM3 protein language model, structure features from ESMFold predicted with equivariant graph neural network (EGNN), and the joint information from sequence and structure learned via Bilinear Attention Network. AMPBAN consistently demonstrated superior accuracy and generalization compared to nine state-of-the-art AMP prediction models across multiple independent benchmarks. Furthermore, an ablation study confirms that our multimodal fusion strategy significantly refines the integration of sequence and structural signals, yielding superior predictive balance over single-modality models. This framework provides a robust tool for the accelerated discovery of novel AMPs and the advancement of next-generation antimicrobial drug development. The datasets, source code and models are available at https://github.com/baiwenhuim/ampban.

The gut microbiome plays an active role in host health, producing gut metabolites that influence host digestive and immune function while also mediating microbial crosstalk. Dietary fiber is a major source of important fermentation byproducts that are generally implicated in gut community stability and host wellbeing, but dissecting microbe-specific contributions to polysaccharide metabolism in the context of a complex gut community is challenging with conventional model organisms. Using the American cockroach (Periplaneta americana) as a model omnivore, we use chemically-defined synthetic diets to identify how complex gut microbial communities respond to two of the most abundant plant polysaccharides, xylan and cellulose. To do so, we fed cockroaches synthetic diets containing one of these fibers or a mix of both in differing ratios. Through both 16S rRNA gene profiling and RNA-seq, we show that mixed fibers enrich for organisms characteristic of the source fibers as well as additional organisms only enriched in mixed-fiber diets. Through an organism-centric pangenome approach, we identify the impact of these fibers on gut microbiome activity. We found that gut communities responded strongly to xylan, with Bacteroidota belonging to Bacteroides, Dysgonomonas, and Parabacteroides producing xylan-active CAZymes at high levels. Multiple groups of Bacillota also responded strongly to a xylan diet, but appeared to act as cross-feeding secondary degraders, producing primarily xylosidases and transcripts associated with xylose utilization. In contrast, cellulose diets were associated with higher transcriptional activity among Fibrobacterota, which are typically a minor component of the cockroach gut microbiome but were the primary producers of CAZymes associated with cellulose and cellobiose degradation. These experiments provide new insight into gut microbial metabolism of these complex plant polysaccharides. Further, they highlight the utility of the cockroach model and synthetic diets to answer fundamental questions about gut microbial responses to different polysaccharides alone and in combination.

Archaeal extrachromosomal elements (ECEs) are arguably the least well understood of all genetic elements, and few have > 200 kbp ("jumbo") genomes. Here, we report circular, jumbo ECEs with genomes of up to 535 kbp in length that associate with anaerobic methane-oxidizing Methanoperedens archaea. Notably, a 409-kbp genome related to jumbo ECEs is integrated into a subset of the ~4.2 Mbp Methanoperedens chromosomes at the tRNA-Asp genes. This represents the largest integrative element in Archaea and supports the jumbo ECE - host association. Multiple genome alignment and phylogenetic analyses suggest that the large sizes were developed by extensive DNA acquisition from Methanoperedens. The newly identified ECEs encode, and in some cases express, metabolic genes such as tetrahydromethanopterin S-methyltransferase exclusively involved in methane oxidation, and genes for nitrogen and sulfur compound transformations. Also encoded are defense systems, some of which are absent in hosts, such as hybrid Type I/Type III-A CRISPR-Cas systems. In contrast to viruses and plasmids, they have host-like replication machinery and occur at stable copy ratios of 1.44 ± 0.24 : 1 to the host. Overall, our results reveal a spectrum of jumbo ECEs of Methanoperedens, ranging from plasmid-like to minichromosome-like.

Virus-induced genome editing (VIGE) using compact RNA-guided endonucleases is a transformational new approach in plant biotechnology, enabling tissue-culture-independent and transgene-free genome editing. We recently established a transgene-free VIGE approach for heritable editing at single loci in Arabidopsis by delivering ISYmu1 TnpB (Ymu1) and its guide RNA (gRNA) via Tobacco Rattle Virus (TRV). Here, we greatly improved this approach by devising a multiple gRNA expression system and by utilizing an engineered high-activity Ymu1 variant (Ymu1-WFR) to develop an efficient multiplexed genome editing approach.

The emergence of antibiotic-resistant bacteria has rendered conventional antibiotic treatments ineffective, necessitating the development of antibacterial agents with unique mechanisms of action. To address this challenge, researchers have increasingly resorted to synthetic and bioengineered nano-materials to augment the antibacterial activity of nonantibiotic antibacterials (nonantibiotic antibacterial agents), including antimicrobial peptides (AMPs), metallic nanoparticles (MNPs), bacteriophages (phages), and phage derivatives such as endolysins, which are under extensive investigation. In this review, we discuss how modifications and syntheses of these agents, leveraging advancements in nanoscience and nanotechnology, have and can significantly enhance their antibacterial properties and overcome limitations such as cytotoxicity, instability, and poor bioavailability for in vivo or clinical use. Furthermore, we highlight supramolecular strategies for improved delivery, including phage-based, AMP-based, and endolysin-based systems and their demonstrated efficacy against persistent bacterial infections. Additionally, we highlight how the integration of artificial intelligence and machine learning ultimately promises to revolutionize the design, optimization, and clinical translation of these precision antimicrobials, paving the way for targeted and highly effective treatments.

Metagenome-assembled genomes (MAGs) are routinely recovered from metagenomic studies, yet the population genetic information embedded within these datasets remains largely underutilized. Analyzing within-species genetic variation can reveal adaptive evolution, selection pressures, and ecological dynamics that are hidden when MAGs are treated as homogeneous entities. Existing tools address individual analysis steps in isolation, requiring manual integration and creating barriers for researchers without extensive bioinformatics expertise. Here we present PopMAG, a Nextflow pipeline and interactive Shiny application that automates population genetics analysis of MAGs. PopMAG integrates quality control, community profiling, competitive read mapping, functional annotation, and microdiversity estimation into a single reproducible workflow. The pipeline calculates key population genetics metrics including nucleotide diversity (π), pN/pS ratios, fixation index (FST), Levins index and SNVs counts with results consolidated into an interactive visualization platform for metadata-driven exploration. We demonstrate PopMAGs utility through analysis of longitudinal cystic fibrosis lung metagenomes, where we detect signatures of antibiotic-driven selection in Pseudomonas aeruginosa efflux pump genes coinciding with treatment intervention. Availability and implementation: PopMAG and corresponding documentation are publicly available at https://github.com/daasabogalro/PopMAG

A methodology is presented to integrate DNA methyltransferase genes of interest into the single copy, IncPbeta; R Factor conjugal plasmid, R702. This is to enable DNA transfer, or improved frequency thereof, by mirroring the recipient native DNA methylation signature prior to transfer of plasmid DNA by conjugation, evading native restriction modification systems in the organism of study, without target strain modification. As proof of concept, the bacteriophage derived methyltransferase phi3TI, known to facilitate DNA transfer into the industrially important model solventogenic organism, Clostridium acetobutylicum ATCC 824, was inserted into the single copy R702 to create a methylating strain. Plasmids extracted from this strain were transferred by electroporation at a comparable frequency to the established method using the multi-copy accessory plasmids pAN1 or pAN2, harbouring phi3TI. The methodology was further exemplified and employed to enable high frequency of DNA transfer into the clinically relevant Clostridioides difficile ribotype 027 outbreak strain, R20291. The modification (hsdM) and specificity (hsdS) components of the native Type I restriction modification system of R20291 was inserted into R702 to create a functional methylating conjugal donor Escherichia coli strain. This approach of coupling methylation and the native transfer functions of R702 creates a system which can easily be mobilised into a genotypically appropriate E. coli strain; creating an in-vivo methylating donor strain which may be utilized to protect and transfer DNA to the organism of study by conjugation.

Autotransporters are important gram-negative bacterial cell-surface proteins and are virulence factors enabling surface attachment and adhesion to other bacteria. These proteins are composed of a signal peptide, a β-barrel that serves as an anchor in the outer membrane, and an extracellular passenger domain responsible for adhesion. Autotransporters rely on BamA for their insertion in the outer membrane (OM), but specific helper proteins, such as TpgA and SadB have been described to promote the surface exposure of trimeric autotransporters (TAAs), a specialized subclass of autotransporters forming homotrimer adhesins. To identify domains or proteins that could help TAAs secretion, we analyzed a recent dataset of all trimeric autotransporters found across the bacterial tree of life. While we did not find additional potential helper proteins for OM translocation, we found that the extended signal peptide (ESPR), sometimes found in TAAs, is associated with longer adhesins both for TAAs and type Va autotransporter adhesins. ESPRs are found in all bacteria but Fusobacteriia and Alphaproteobacteria. We also identified in Burkholderia, Veillonellales and Pasteurellales a DUF2827 domain proteins as potential glycosyltransferases constantly associated with TAAs. Finally, we describe the existence of extra periplasmic domains in some TAAs, featuring either a coiled-coil domain or a peptidoglycan-binding domain. Our research show that there is a strong phylogenetic separation between Terrabacteria, almost invariably displaying additional periplasmic domains, and Gracilicutes (represented by Proteobacteria) where they are largely absent. This suggests that the presence these domains might be correlated with specific Terrabacteria OM features. Using the diderm Firmicute Veillonella parvula as a model, we demonstrate that the absence of periplasmic domains in TAAs leads to a significant protein degradation, yet they are not essential for adhesin trimerization or secretion. Additionaly, we show that the SLH domains of V. parvula TAAs excludes them from the septum during division, but that this exclusion is not crucial for adhesin function or stability in the tested conditions. Altogether, these results illuminate the genetic flexibility and modularity of autotransporters, enhancing our understanding of this important class of adhesins in diderm bacteria.