Nature Microbiology

Competition assays are a mainstay of modern microbiology, offering a simple and cost-effective means to quantify microbe-microbe interactions in vitro. Here we demonstrate a key weakness of this method that arises when competing microbes interact via toxins, such as those secreted via the type VI secretion system (T6SS). Timelapse microscopy reveals T6SS-armed Acinteobacter baylyi bacteria can maintain lethal T6SS activity against E. coli target cells, even under selective conditions intended to eliminate A. baylyi. Further, this residual killing creates a density- and T6SS-dependent bias in the apparent recovery of E. coli, leading to a misreporting of competition outcomes especially where target survival is low. We also show that incubating A. baylyi / E. coli co-cultures in liquid antibiotic prior to selective plating can substantially correct this bias. Our findings demonstrate the need for caution when using selective plating as part of T6SS competition assays, or assays involving other toxin-producing bacteria.

Conventional genomic risk classification of Listeria monocytogenes assigns clonal complexes to hypervirulent (CC1, CC2, CC4, CC6) or hypovirulent (CC9, CC121) categories based on population-level frequency ratios, leaving all remaining diversity in an undifferentiated "intermediate" category that carries no defined risk assessment. We analysed 436 genomes from confirmed invasive listeriosis across 19 countries using multi-dimensional genomic profiling of virulence and persistence determinants and demonstrate that this approach systematically misclassifies a major fraction of clinically relevant L. monocytogenes. Amphitrophic lineages -- carrying simultaneous genomic competence for clinical virulence (functional inlA, mean virulence score 52.7 +/- 6.6) and industrial persistence (SSI-1 in 94.1%, mean persistence score 66.8 +/- 11.6) -- constitute 31.0% of invasive disease, within 3.6 percentage points of the established hypervirulent category (34.6%). Of these 135 amphitrophic clinical isolates, 91.1% were classified as "intermediate" under conventional taxonomy. The five principal amphitrophic CCs (CC8, CC7, CC3, CC5, CC88) appear with indistinguishable dual-fitness genotypes in both clinical and food-chain datasets, establishing that the same organisms persist in processing facilities and cause invasive human disease. Decomposition of the species-level virulence-persistence trade-off (rho = -0.523) by trophic strategy reveals it to be a Simpsons paradox: no within-strategy correlation is significantly negative, and the only significant signal is a positive amphitrophic correlation (rho = +0.221, p = 0.010) indicating synergy rather than trade-off. Multi-dimensional profiling increases risk-stratified detection from 32.3% (conventional) to 65.6% of clinical isolates -- a 103% improvement. These findings demonstrate that clonal complex identity alone leaves one-third of clinically significant L. monocytogenes uncharacterised, and that effective One Health genomic surveillance requires simultaneous assessment of virulence and persistence at the isolate level.

The envelope stress response (ESR) of Gram-negative enteric bacteria senses fluctuations in nutrient availability and environmental changes to avert damage and promote survival. It has a protective role towards antimicrobials but direct interactions between ESR components and antibiotic resistance genes have not been demonstrated. Here we report specific interactions between the two-component conjugative pilus expression (Cpx)RA signal transduction system and the recently described mobile colistin resistance (MCR-1) protein. Purified MCR-1 is specifically cleaved by the serine endoprotease DegP within a structurally conserved periplasmic bridging domain. Cleavage-site mutations in MCR-1 render derivatives either protease-resistant or degradation-susceptible with widely differing consequences for colistin resistance. Transfer of the degradation-susceptible mutant to strains that lack either DegP or its regulator CpxRA restores expression and colistin resistance. MCR-1 production in Escherichia coli induces a Cpx-dependent ESR and imposes growth restriction in strains lacking either DegP or CpxRA, effects that are reversed by transactive expression of DegP. MCR-1 production impairs bacterial motility indicating dissipation of cytoplasmic transmembrane potential. Indeed, growth in media with low pH dramatically increases both MCR-1-dependent phosphoethanolamine (PEA) modification of lipid A as well as colistin-resistance activity. In vitro transferase- and lipid A reconstitution-assays demonstrate that MCR-1 is highly active at acidic pH. Acquiring MCR-1 also renders strains more resistant to antimicrobial peptides. Thus, a conserved motif within MCR-1 induces components of the ESR to confer resilience to stimuili commonly encountered in the environment such as to changes in pH and towards antimicrobial peptides. Excipient allosteric activation of the DegP protease specifically inhibits growth of isolates carrying mcr-1 plasmids indicating that a targeted strategy can lead to the elimination of transferable colistin resistance in Gram-negative bacteria.

The archaeal phylum Nanobdellota (formerly Nanoarchaeota) was previously represented by four complete genomes. We present 208 complete Nanobdellota genomes from Oxford Nanopore metagenomes of the Baltic Sea water column and Fennoscandian groundwater (69-201 m below sea level), rotated to the ORC1/Cdc6 replication origin -- a 52-fold expansion of complete-genome representation. Across the ar53 supermatrix and a Nanobdellota-tuned 71-marker supermatrix on 1,239 taxa, the named GTDB orders within Nanobdellota are recovered as monophyletic clades, including the three orders that dominate our environmental sampling: Woesearchaeales, Pacearchaeales, and the GTDB placeholder order SCGC-AAA011-G17. This is consistent with the existing GTDB R232 order-level circumscription. We retire the SCGC-AAA011-G17 placeholder name, replacing it with a complete-genome-anchored SeqCode nomenclatural chain (Maxwellarchaeales ord. nov., Maxwellarchaeaceae fam. nov., Maxwellarchaeum gen. nov., and Maxwellarchaeum balticum sp. nov.) without altering the order-level circumscription. Pacearchaeales and Maxwellarchaeales retain no central or energy metabolism beyond Form III RuBisCO, PEP synthase, and ferredoxin; Woesearchaeales retains partial glycolysis and a V/A-type ATPase. A 4,262-tip phylogeny of rbcL (the RuBisCO large-subunit gene) identifies nine candidate archaea-to-Patescibacteriota Form III RuBisCO transfer events -- including one to a Baltic Minisyncoccia -- versus two reciprocal candidates, consistent with archaea-to-CPR being the more frequently identified direction in our data. All 256 Nanobdellota genomes (208 complete + 48 high-quality non-circular), the ar71 marker set with its 1,239-taxon ML tree, 154 Nanobdellota-trained HMMs for KEGG-ortholog detection in DPANN proteomes (94 ROBUST), and the 4,262-tip rbcL reference tree are released as a community resource, alongside the full analysis archive -- alignments, intermediate trees, structural predictions, and per-step scripts -- at Zenodo (DOI 10.5281/zenodo.20174424; see Using the resource).

Pelagibacter, the largest genus within the SAR11 clade, is the most abundant bacterium in the ocean, yet the vast majority of its species-level diversity remains uncharacterized at the genomic level. Here we present 135 complete Pelagibacter genomes -- the largest such collection assembled to date -- comprising 75 from Oxford Nanopore metagenomes of the San Francisco Estuary (SFE), 31 from a deeply sequenced station within the same transect, and 29 from public databases. These genomes define 52 species at 95% ANI, of which 44 (85%) are taxonomically novel. An expanded phylogeny incorporating 89 additional high-quality NCBI genomes confirms that our collection captures the phylogenetic backbone of the genus, with genomes from Hawaii, Namibia, and the Sargasso Sea nesting within SFE clades. The pangenome is open (14,862 singletons, 62%), driven by two distinct mechanisms. First, a universal hypervariable region (HVR) at a conserved chromosomal position (7-15% from dnaA) is present in all 135 genomes, anchored by tRNA genes at both boundaries (Phe/His and Arg). The HVR carries genome-specific surface polysaccharide biosynthesis genes with a GC age gradient -- highest GC at the tRNA boundaries, lowest in the center -- consistent with a two-ended phage insertion model. Only this HVR is positionally conserved across the genus; the three other hypervariable regions previously described in a single reference genome are not. Second, scattered genomic islands throughout the chromosome contribute the remaining singleton content, including chimeric islands with genes from four bacterial phyla. Biosynthetic pathway reconstruction reveals auxotrophies that are phylogenetically structured, not uniform: biotin, reduced sulfur, and glycine are genus-wide dependencies, while isoleucine, pantothenate, histidine, and glyoxylate cycle capacity vary across lineages with significant phylogenetic clustering. Structural annotation with ESMFold and Foldseek resolved 3,125 hypothetical proteins; 1,222 remain uncharacterized by any method, including a 47-amino-acid protein conserved in two-thirds of all genomes within a fixed operonic context -- independently predicted by two gene callers yet matching nothing in any database. A controlled depth comparison at one station demonstrates that standard metagenome sequencing systematically underestimates Pelagibacter diversity, with three species recovered only at elevated depth and the species count at that station more than doubling (9 vs 4).

Soils harbour immense biosynthetic gene cluster (BGC) diversity that can mediate microbial interactions, yet this potential is still mapped unevenly across the tree of life. Ktedonobacteria--a class of actinomycete-like bacteria within phylum Chloroflexota--are widespread in terrestrial environments and repeatedly dominate pioneer communities in extremely oligotrophic volcanic bare-ground soils; however, their secondary metabolism and genome architecture remain poorly characterised. Here, we integrate targeted cultivation using volcanic soils from Mount Zao with genome-resolved metagenomics and public genomes to analyse 183 ktedonobacterial genomes. Using antiSMASH and BiG-SLiCE, we identified 1,546 BGCs comprising 1,162 non-redundant gene-cluster families (GCFs). In our dataset, nearly one quarter of genomes encode [≥]10 distinct GCFs, and several family-level clades show mean GCF counts comparable to those in genus Streptomyces. Most ktedonobacterial BGCs are highly divergent from reference collections and exhibit unusually low intra-genomic redundancy, suggesting broad, underexplored chemotypes. Long-read assemblies from ten strains reveal recurrent 1.6-3.5 Mb chromid-like secondary replicons with chromosome-like composition but distinct maintenance signatures. These replicons are consistently enriched in BGCs and mobility-associated genes, with mobility loci concentrated near BGC boundaries. Collectively, our results expand the current knowledge of the phylogenetic landscape of soil biosynthetic diversity and highlight chromid-like secondary replicons as major genomic reservoirs for specialised metabolism in Ktedonobacteria.

Halogenated one-carbon (C1) compounds, such as dichloromethane (DCM), drive critical fluxes in global carbon and halogen cycles. While the genus Dehalobacter is canonically defined by obligate organohalide respiration, its physiological and ecological roles in anaerobic C1 metabolism have remained fundamentally ambiguous. Here, we document a paradigm-shifting metabolic capacity within a sediment-derived microbial consortium: the autonomous fermentation of DCM by a novel population, Candidatus Dehalobacter formatiformans strain J1. Over successive transfers, strain J1 outcompeted co-existing Dehalobacterium formicoaceticum to become the overwhelmingly dominant population (>80% relative abundance), converting DCM stoichiometrically to acetate and formate (4:1) without auxiliary substrates. Genome-resolved metagenomics revealed that strain J1 couples a distinct mec gene cassette--mediating methyl-transfer reactions during DCM activation--to a complete Wood-Ljungdahl pathway for efficient C1 assimilation. Crucially, strain J1 lacks the complete genetic repertoire for de novo cobamide biosynthesis. Physiological validation confirmed that this fermentative pathway is strictly dependent on exogenous cobamides, exposing a profound reliance on community cross-feeding. These findings reveal an unexpected acetogenic lifestyle within Dehalobacter, a lineage historically viewed as comprising obligate organohalide-respiring bacteria. More broadly, this work identifies cobamide-dependent methyl-transfer metabolism as an ecological control on anaerobic DCM fermentation and expands the known roles of Dehalobacter in carbon-halogen cycling in anoxic environments.

Whole-genome sequencing is revolutionizing bacterial outbreak investigation but its application to the clinic remains limited. In 2020, prospective and retrospective surveillance detected a Pseudomonas aeruginosa outbreak with 253 isolates collected from 82 patients in 26 wards of a hospital. Its origin was dated to the late 90s, just after the facility opened, and patient-to-patient and environment-to-patient cases of transmission were inferred. Over time, two epidemic subclones evolved in separate hosts and hospital areas, including newly opened wards, and hospital-wide sampling confirmed reservoirs persisted in the plumbing. Pathoadaptive mutations in genes associated with virulence, cell wall biogenesis, and antibiotic resistance were identified. While the latter correlated with the acquisition of phenotypic resistances to 1st (cephalosporin), 2nd (carbapenems) and 3rd (colistin) lines of treatment, maximum parsimony suggested that a truncation in a lipopolysaccharide component coincided with the emergence of a subclone prevalent in long-term infections. Since initial identification, extensive infection control efforts guided by routine, near real-time surveillance have proved successful at slowing transmission.

In a single Oxford Nanopore long-read metagenome from a Fennoscandian deep-groundwater borehole (KR0015B, Aspo Hard Rock Laboratory, Sweden), 791 protein clusters span at least one chromosomal contig and at least one co-sampled circular mobile element -- the cross-replicon LGT-candidate cohort. The 199 participating chromosomes are dominated by three small-genome / symbiont-associated lineages -- Patescibacteriota, Omnitrophota, and Nanobdellota -- but the per-chromosome participation rate tells a different story: Omnitrophota chromosomes participate at an order-of-magnitude higher rate (mean 56 cross-replicon clusters per genome), while Patescibacteriota and Nanobdellota dominate by compositional abundance only. Two large divergent circular mobile elements (233-kbp u20424375 and 123-kbp u29249220) -- each lineage-restricted within a single Omnitrophota genus, with sparse cross-phylum reach (only 12 of their combined 289 cross-replicon clusters involve a non-Omnitrophota partner) -- together account for 37% of the cohort and lack canonical plasmid or phage signatures. The 233-kbp element carries a Mu-class DDE transposase, is found integrated in one host chromosome at 99.3% nucleotide identity over 87% of element length, and carries an essentially complete bacterial big-operon r-protein cluster (31 r-protein KOs) as cargo with no rRNA genes -- a cargo profile with no published precedent in the mobile-element literature. Seven cross-replicon clusters span both domains; per-cluster phylogenies confirm gene-tree topologies that violate the species-tree expectation in 6 of 10 callable smoking-gun trees. We release the cross-replicon cluster table, integrated mobile-partner classifications, and chromosome taxonomy as a community resource. A parallel cross-chromosome catalog without the mobile-partner requirement contains 957 clusters, 95% of which carry no co-sampled circular plasmid or virus partner -- a chromosome-only LGT footprint that bounds the MGE-coupled cohort and is consistent with vehicle-free / direct-contact transfer in lineages whose close-contact symbiotic biology is well-documented.

The evolution of obligate host-association of bacterial symbionts and pathogens remains poorly understood. The Rickettsiales represent an order of obligate alphaproteobacterial endosymbionts and parasites that infect a wide variety of eukaryotic hosts, including humans, livestock, insects and protists. Induced by their host-associated lifestyle, Rickettsiales genomes have undergone reductive evolution, leading to small, AT-rich genomes with limited metabolic capacities. We describe several genomes of deep-branching, environmental alphaproteobacteria that branch basal to previously sampled Rickettsiales, and whose genome content are reminiscent of free-living and biofilm-associated lifestyles. Ancestral genome content reconstruction across the Rickettsiales tree revealed that the free-living to host-association transition of this group occurred more recently than previously anticipated, and likely involved the repurposing of a type IV secretion system. One-Sentence SummaryDeep-branching Rickettsiales provide insights into the evolution of obligate host-associated lifestyle

Toxin-antitoxin (TA) systems are ubiquitous genetic elements in bacterial genomes, but their functions are controversial. Although they are frequently postulated to regulate cell growth following stress, few null phenotypes for TA systems have been reported. Here, we show that TA transcript levels can increase substantially in response to stress, but toxin is not liberated. We find that the growth of an Escherichia coli strain lacking 10 TA systems encoding endoribonuclease toxins is not affected following exposure to six stresses that each trigger TA transcription. Additionally, using RNA-sequencing, we find no evidence of mRNA cleavage following stress. Stress-induced transcription arises from antitoxin degradation and relief of transcriptional autoregulation. Importantly, although free antitoxin is readily degraded in vivo, antitoxin bound to toxin is protected from proteolysis, preventing release of active toxin. Thus, transcription is not a reliable marker of TA activity, and TA systems likely do not strongly promote survival following stress.

The biosynthetic locus encoding the exopolysaccharide poly-N-acetyl-glucosamine (PNAG) is widely conserved across bacteria, including the WHO critical-priority pathogen Klebsiella pneumoniae (Kp). In Kp, PNAG synthesis is mediated by the pgaABCD operon, yet its lineage-specific regulation remains incompletely defined. Using a comparative genomics approach to interrogate the pgaABCD locus across the high-risk clonal Kp complex 258 (CC258) lineage, we identified a previously uncharacterised positive transcriptional regulator located immediately upstream of pgaA, which we designate pgaR. Phylogenetic analysis revealed recurrent evolutionary events affecting this regulatory region, including repeated deletion or truncation of pgaR and a G>A substitution upstream of the pgaR start codon. Functional characterisation demonstrated that loss of pgaR abolishes pgaABCD expression and PNAG production, whereas the upstream G>A substitution drives PNAG hyper-production. In vitro, Kp produce extensive extracellular PNAG networks under static growth conditions, consistent with a role in biofilm architecture. Despite this, PNAG expression was dispensable in murine pneumonia and peritonitis models, while PNAG hyper-production significantly attenuated virulence and disease severity, indicating a fitness cost associated with sustained overexpression. Collectively, we discovered PgaR as a novel gene regulator of the pgaABCD operon. We show a previously unrecognised lineage-specific layer of PNAG regulation in Kp and demonstrate that opposing PNAG phenotypes: loss and hyper-production, have independently and repeatedly emerged among clinical CC258 isolates, highlighting dynamic selection acting on biofilm-associated traits in this high-risk pathogen. ImportanceThe exopolysaccharide poly-N-acetyl-glucosamine (PNAG) is widely conserved in bacteria, including the WHO critical-priority pathogen Klebsiella pneumoniae. However, how PNAG production is regulated in high-risk lineages has remained unclear. Here, we identify PgaR as a previously unrecognised positive regulator of the pgaABCD operon in clonal complex 258, a globally disseminated and drug-resistant lineage. We show that natural genetic variation within this regulatory region leads to strikingly different PNAG phenotypes: complete loss of production or hyper-production. While PNAG contributes to extracellular matrix formation in vitro, it is dispensable for virulence in murine infection models, and sustained overproduction imposes a fitness cost. The repeated and independent emergence of both loss- and gain-of-function variants among clinical isolates reveals dynamic evolutionary pressures acting on biofilm-associated traits. These findings uncover a lineage-specific layer of PNAG regulation and highlight how modulation of surface polysaccharide expression shapes pathogen fitness and adaptation.

Lyme disease is the most common vector-borne disease in North America and Europe. The clinical manifestations of Lyme disease vary based on the genospecies of the infecting Borrelia burgdorferi spirochete, but the microbial genetic elements underlying these associations are not known. Here, we report the whole genome sequence (WGS) and analysis of 299 patient-derived B. burgdorferi sensu stricto (Bbss) isolates from patients in the Eastern and Midwestern US and Central Europe. We develop a WGS-based classification of Bbss isolates, confirm and extend the findings of previous single- and multi-locus typing systems, define the plasmid profiles of human-infectious Bbss isolates, annotate the core and strain-variable surface lipoproteome, and identify loci associated with disseminated infection. A core genome consisting of [~]800 open reading frames and a core set of plasmids consisting of lp17, lp25, lp36, lp28-3, lp28-4, lp54, and cp26 are found in nearly all isolates. Strain-variable (accessory) plasmids and genes correlate strongly with phylogeny. Using genetic association study methods, we identify an accessory genome signature associated with dissemination and define the individual plasmids and genes that make up this signature. Strains within the RST1/WGS A subgroup, particularly a subset marked by the OspC type A genotype, are associated with increased rates of dissemination. OspC type A strains possess a unique constellation of strongly linked genetic changes including the presence of lp56 and lp28-1 plasmids and a cluster of genes that may contribute to their enhanced virulence compared to other genotypes. The patterns of OspC type A strains typify a broader paradigm across Bbss isolates, in which genetic structure is defined by correlated groups of strain-variable genes located predominantly on plasmids, particularly for expression of surface-exposed lipoproteins. These clusters of genes are inherited in blocks through strain-specific patterns of plasmid occupancy and are associated with the probability of invasive infection.

The cell wall is a polymeric exoskeleton that defines the size and shape of bacteria; it is composed of peptidoglycan and, in Gram-positive bacteria, wall teichoic acid polymers. Two systems synthesize peptidoglycan in rod-shaped bacteria, which are found pervasively across bacterial clades: the multi-protein Rod complexes synthesize anisotropic peptidoglycan, which is required for rod shape because it reinforces the cell wall along its circumference1-3, whereas the non-essential enzyme PBP1 synthesizes isotropic peptidoglycan1,4 (Fig. 1A). In Gram-positive bacteria, rod shape also requires wall teichoic acids5 for unknown reasons. Here, we show that wall teichoic acids promote rod shape by preventing the formation of nanoscopic pores in the Bacillus subtilis cell wall, which lead to amorphous growth by activating PBP1 and inhibiting Rod complexes. Depleting wall teichoic acids resulted in pores within minutes, coinciding with a rapid increase in PBP1-mediated synthesis. PBP1s ability to sustain teichoic acid-less growth depended on its intrinsically disordered domain. In contrast to previous steady-state measurements6,7, we found that wall teichoic acid depletion caused the transient arrest of Rod complexes prior to the onset of amorphous growth. Finally, one of the two synthetically lethal cell wall hydrolases in B. subtilis8, LytE, became essential during wall teichoic acid depletion, meaning that PBP1 and LytE execute a novel, amorphous mode of growth. Collectively, our results identify the cell wall, via its molecular-scale structure, as a non-canonical auto-regulator of its own synthesis. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=109 SRC="FIGDIR/small/610702v3_fig1.gif" ALT="Figure 1"> View larger version (41K): org.highwire.dtl.DTLVardef@154996forg.highwire.dtl.DTLVardef@12561c3org.highwire.dtl.DTLVardef@1356cacorg.highwire.dtl.DTLVardef@719d22_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOFig. 1:C_FLOATNO Inhibiting wall teichoic acid synthesis decreases Rod complex activity prior to cell shape loss.(A) Schematic of Gram-positive cell wall synthesis. (B) Cell length growth rate [Figure 1] for wild-type cells during 0.5 g/mL tunicamycin treatment. Blue line shows smoothed population median, error bars show standard deviation. Data shown for 2,048 discrete cell tracks from 3 biological replicates. Inset: micrographs taken (i) before and (ii) after 45 min tunicamycin treatment. Scale bars 5 m. (C) Time course for the percentage of processive Mbl filaments during 0.5 g/mL tunicamycin treatment. Error bars show 95% confidence intervals across biological replicates (bootstrap analysis). Analysis performed on 62,901 discrete filament tracks from 4 biological replicates. Inset shows representative fluorescent kymographs of Mbl motion across the cell waist in (i) LB and (ii) after 30 min tunicamycin treatment. Red dotted line follows a processive Mbl filament. Scale bars 1 m. (D) Mean fluorescent puncta per unit cell length for wild-type, {Delta}ponA and ponATP-cells during tunicamycin treatment, labeled with fluorescent D-amino acids. Error bars show 95% confidence intervals. Analysis performed over 2,017 segmentations from 4 biological replicates (wild-type), 857 segmentations from 3 biological replicates ({Delta}ponA) and 257 cells from 2 biological replicates (ponATP-). All timepoints show statistical significance for difference in mean between cell types (Students T-test, P<0.01). Dots show individual replicates. Measurements at 50 min tunicamycin treatment used a lower exposure time so are plotted separately. Inset shows micrographs of (i,iii,v) wild-type and (ii,iv,vi) {Delta}ponA cells at (i,ii) 0 min, (iii,iv) 30 min, and (v,vi) 50 min of treatment. Micrographs are identically saturated across cell types. Scale bars 5 m. (E) Percentage of peptidoglycan subunits in monomers, measured by HPLC (Methods). Two biological replicates per strain. Mean values calculated across biological replicates. Error bars show 95% confidence intervals. Individual datapoints shown in red. Statistical significance calculated using Students T-test, P<0.05. C_FIG

Streptococcus pneumoniae is a leading cause of childhood meningitis, sepsis and pneumonia despite widespread implementation of pneumococcal conjugate vaccines (PCVs). Serotype 2, once a major invasive serotype that nearly disappeared in the mid-20th century, is not included in current vaccine formulations. Recent reports from multiple countries suggest potential re-emergence of serotype 2. Here, we present 30 years of hospital-based surveillance from Bangladesh (1993-2022), where serotype 2 accounted for 7.8% of invasive pneumococcal disease cases. Infections occurred predominantly in very young infants (median age, 3 months) and were largely associated with meningitis (91.3%), with nearly 90% of isolates recovered from cerebrospinal fluid. Comparative analysis of otitis media and nasopharyngeal carriage isolates demonstrated high invasive propensity relative to other serotypes. Whole genome sequencing of 170 serotype 2 isolates from 21 countries revealed that all modern isolates belong to the globally disseminated lineage GPSC96, which is distinct from the prototypical laboratory strain D39 (GPSC622). Phylodynamic reconstruction dated the emergence of GPSC96 to the late 19th century, with continued global circulation and largely preserved antibiotic susceptibility. These findings highlight serotype 2 as a potential invasive pneumococcal threat in countries such as Bangladesh and supports consideration of its inclusion in the next-generation conjugate vaccines.

Plant roots release a wide array of metabolites into the rhizosphere, shaping microbial communities and their functions. While metagenomics has expanded our understanding of these communities, little is known about the physiology of their members in host environments. Transcriptome analysis via RNA sequencing is a common approach to learning more, but its use has been challenging because plant RNA masks bacterial transcripts. To overcome this, we developed randomly-barcoded promoter-library insertion sequencing (RB-PI-seq) and combined it with chassis-independent recombinase assisted genome engineering (CRAGE), using Pseudomonas simiae WCS417 as a model rhizobacterium. This method enables targeted amplification of barcoded transcripts, bypassing plant RNA interference and allowing measurement of thousands of promoter activities during root colonization. Our analysis revealed time-resolved dynamics of promoter activities, highlighting early transcriptional reprogramming as a key determinant of successful colonization. Additionally, we discovered that transcriptional activation of xanthine dehydrogenase and a lysozyme inhibitor are crucial for evading plant immune system defenses. This framework is scalable to other bacterial species and provides new opportunities for understanding rhizobacterial gene regulation in native environments.

Microbiome sequencing measures relative rather than absolute abundances, providing no direct information about total microbial load. Normalization methods attempt to compensate, but rely on strong, often untestable assumptions that can bias inference. Experimental measurements of load (e.g., qPCR, flow cytometry) offer a solution, but remain costly and uncommon. A recent high-profile study proposed that machine learning could bypass this limitation by predicting microbial load from sequencing data alone. To evaluate this claim, we assembled mutt, the largest public database of paired sequencing and load measurements, spanning 35 studies and over 15,000 samples. Using mutt, we show that published machine learning models fail to generalize: on average they perform worse than a naive baseline that always predicted the training set mean. These failures stem from covariate shift-limited shared taxa between studies, differences in community composition, and differences in preprocessing pipelines-that silently derail model inputs. In contrast, Bayesian partially identified models do not attempt to impute microbial load, but instead propagate scale uncertainty through downstream analyses. Across 30 benchmark datasets, Bayesian partially identified models consistently outperformed normalization and machine learning approaches, providing a principled and reproducible foundation for microbiome inference.

Predicting microbiome function remains challenging as microbial interactions scale from pairwise encounters to emergent community properties. This is particularly true of disease protective microbial consortia, where pathogen invasion has typically been studied either in terms of single biocontrol agents or in terms of microbiome diversity at the full community level, but rarely in between. Focusing on a 16-member synthetic tomato phyllosphere bacterial community, we combined reciprocal spent-media growth assays of over 600 pairwise and community-level combinations with comparative genomics to dissect the ecological and metabolic drivers of community interactions. Across the interaction network, negative interactions dominated, with community-derived spent media consistently exerting stronger inhibitory effects on bacterial growth across the community than any single-species filtrate. While two isolates (Exiguobacterium sibiricum and Bacillus thuringiensis) exhibited strong inhibitory effects in monoculture assays, community spent media analyses revealed that no single strain was responsible for the pathogen-suppressive phenotype observed in community, indicating that protection against Pseudomonas syringae is an emergent property of the particular community composition. Furthermore, using correlations and cross-validated multivariate models, inhibition strengths were poorly predicted by either genomic annotations or phenotypic strategies. Instead, community context strongly constrained environmental modification and buffered strain-specific effects observed in isolation. Together, these results demonstrate that microbial community function cannot easily be inferred from pairwise interactions or individual strain properties alone, and that both direct and indirect interactions shape phyllosphere community structure and function, with emergent properties such as pathogen suppression arising from collective properties rather than the presence/absence or dominance of individual keystone taxa.

Degradation of complex carbohydrates in the gut is a key trait of Bacteroides species. Some glycans are metabolised selfishly releasing few or no oligosaccharide breakdown products from complex polysaccharides, whereas others release oligosaccharides and cross feed other microbes. The outer cell wall of many fungi commonly found in the gut consists of highly -mannosylated proteins which have been shown to be metabolised in a selfish manner by Bacteroides thetaiotaomicron. We show that the species Bacteroides salyersiae releases branched manno-oligosaccharides during growth on mannan and that these act as a nutrient source for Bacteroides spp. that are unable to degrade polymeric mannan. Molecular characterisation of the locus responsible for mannan degradation reveals that it contains multiple glycoside hydrolases and glycan binding proteins targeted to the Type 9 Secretion System, a Bacteroidetes specific secretion system that allows the secretion of large folded proteins across the outer membrane. More commonly found in oral and environmental Bacteroidetes, here the T9SS enables B. salyersiae to locate large, multimodular enzymes and glycan binding proteins outside the cell to target a complex, branched polysaccharide. This points to a previously unknown role of the T9SS in glycan metabolism in gut Bacteroides.

Habitat transitions are central to microbial ecology and evolution and have been extensively studied across vastly different environments, such as between saline and non-saline environments. However, microbial habitat transitions along other large-scale environmental gradients remain poorly studied. This is particularly true for transitions involving the cryosphere, despite building evidence suggesting the Cryogenian as important for evolutionary radiation. Here, we investigated ecosystem transitions and related genomic adaptations of the cosmopolitan cryospheric Polaromonas bacterium. We constructed a pangenome from 282 high-quality genomes, sourced from glaciers, glacier-fed streams, lakes, wetlands, groundwater, rivers, and soils. Phylogenetic reconciliation revealed that the ancestral Polaromonas genome radiated from glacier ecosystems into various downstream environments through multiple independent transitions. These transitions were marked by extensive horizontal gene transfer and gene loss, with mobile genetic elements, such as plasmids and prophages playing key roles in genomic diversification. Predicted ancestral genomes encoded versatile metabolic and stress-response capacities, supporting adaptation to fluctuating and extreme conditions in the various cryospheric habitats. Compared to the ancestral Polaromonas genome, distinct genomic signatures were associated with specific habitats: glacier-fed stream lineages possess expanded stress tolerance repertoires, glacier lineages gained chemolithotrophic and anaerobic pathways, lake and wetland genomes acquired phototrophic functions, and soil lineages expanded substrate transport and stress tolerance. Together, our findings highlight the role of genomic plasticity in the ecological success of Polaromonas, and also underscore the cryosphere as a potential evolutionary cradle from which lineages dispersed and adapted to downstream aquatic and terrestrial environments.