Search for: All records

Abstract BackgroundThe annotation of protein sequences in public databases has long posed a challenge in molecular biology. This issue is particularly acute for viral proteins, which demonstrate limited homology to known proteins when using alignment, k-mer, or profile-based homology search approaches. A novel methodology employing Large Language Models (LLMs) addresses this methodological challenge by annotating protein sequences based on embeddings. ResultsCentral to our contribution is the soft alignment algorithm, drawing from traditional protein alignment but leveraging embedding similarity at the amino acid level to bypass the need for conventional scoring matrices. This method not only surpasses pooled embedding-based models in efficiency but also in interpretability, enabling users to easily trace homologous amino acids and delve deeper into the alignments. Far from being a black box, our approach provides transparent, BLAST-like alignment visualizations, combining traditional biological research with AI advancements to elevate protein annotation through embedding-based analysis while ensuring interpretability. Tests using the Virus Orthologous Groups and ViralZone protein databases indicated that the novel soft alignment approach recognized and annotated sequences that both blastp and pooling-based methods, which are commonly used for sequence annotation, failed to detect. ConclusionThe embeddings approach shows the great potential of LLMs for enhancing protein sequence annotation, especially in viral genomics. These findings present a promising avenue for more efficient and accurate protein function inference in molecular biology. 

Abstract Marine microbes are important in biogeochemical cycling, but the nature and magnitude of their contributions are influenced by their associated viruses. In the presence of a lytic virus, cells that have evolved resistance to infection have an obvious fitness advantage over relatives that remain susceptible. However, susceptible cells remain extant in the wild, implying that the evolution of a fitness advantage in one dimension (virus resistance) must be accompanied by a fitness cost in another dimension. Identifying costs of resistance is challenging because fitness is context‐dependent. We examined the context dependence of fitness costs in isolates of the picophytoplankton genusMicromonasand their co‐occurring dsDNA viruses using experimental evolution. After generating 88 resistant lineages from two ancestralMicromonasstrains, each challenged with one of four distinct viral strains, we found resistance led to a 46% decrease in mean growth rate under high irradiance and a 19% decrease under low. After a year in culture, the experimentally selected lines remained resistant, but fitness costs had attenuated. Our results suggest that the cost of resistance inMicromonasis dependent on environmental conditions and the duration of population adaptation, illustrating the dynamic nature of fitness costs of viral resistance among marine protists. 

Phages play a role in shaping ecosystems by controlling host abundance via cell lysis, driving host evolution via horizontal gene transfer, and promoting nutrient cycling. The genus Bradyrhizobium includes bacteria able to symbiotically nodulate the roots of soybean (Glycine max), providing the plant with a direct source of biologically fixed nitrogen. Optimizing this symbiosis can minimize the use of nitrogen fertilizers and make soybean production more sustainable. Phages targeting Bradyrhizobium may modify their hosts’ genotype, alter phenotypic traits such as symbiotic effectiveness, and mediate competition among strains for nodulation sites. Sixteen phages were isolated against B. diazoefficiens strain USDA110 and B. elkanii strains USDA94 and USDA31. Comparative analyses revealed host species-dependent diversity in morphology, host range, and genome composition, leading to the identification of three previously undescribed phage species. Remarkably, all B. elkanii phages shared a siphophage morphology and formed a single species with >97% nucleotide identity, even when isolated from farms separated by up to ~70 km, suggesting genomic stability across geographic scales. In contrast, phages isolated against B. diazoefficiens had a podophage-like morphology, exhibited greater genetic diversity, and divided into two distinct species. Although no phages were recovered against the B. japonicum strains or native Delaware Bradyrhizobium isolates tested, some Delaware Bradyrhizobium isolates showed susceptibility in a host range assay. The phage genomes demonstrated features predicting phenotypes. The phage terminase genes predicted headful packaging which promotes generalized transduction. The B. elkanii phages all carried tmRNA genes capable of rescuing stalled ribosomes, and all but one of the phages isolated against the two host species carried DNA polymerase A indicating greater phage control of genome replication. State-of-the-art structural annotation of a hypothetical gene shared by the B. diazoefficiens phages, having a mean amino acid identity of ~25% and similarity of ~35%, predicted a putative tail fiber function. Together this work expands the limited knowledge available on soybean Bradyrhizobium phage ecology and genomics. 

Viruses infecting aquatic microbes vary immensely in size, but the ecological consequences of virus size are poorly understood. Here we used a unique suite of diverse phytoplankton strains and their viruses, all isolated from waters around Hawai'i, to assess whether virus size affects the suppression of host populations. We found that small viruses of diverse genome type (3–24 kb genome size, 23–70 nm capsid diameter) have very similar effects on host populations, suppressing hosts less strongly and for a shorter period of time compared to large double‐stranded DNA viruses (214–1380 kb, 112–386 nm). Suppressive effects of larger viruses were more heterogeneous, but most isolates reduced host populations by many orders of magnitude, without recovery over the ~ 25‐d experiments. Our results suggest that disparate lineages of viruses may have ecological consequences that are predictable in part based on size, and that ecosystem impacts of viral infection may vary with the size structure of the viral community. 

Protein language models (pLMs) have revolutionized computational biology by generating rich protein vector representations, or embeddings—enabling major advancements inde novoprotein design, structure prediction, variant effect analysis, and evolutionary studies. Despite these breakthroughs, current pLMs often exhibit biases against proteins from underrepresented species, with viral proteins being particularly affected, frequently referred to as the “dark matter” of the biological world due to their vast diversity and ubiquity, yet sparse representation in training datasets. Here, we show that fine-tuning pre-trained pLMs on viral protein sequences, using diverse learning frameworks and parameter-efficient strategies, significantly enhances representation quality and improves performance on downstream tasks. To support further research, we provide source code for fine-tuning pLMs and benchmarking embedding quality. By enabling more accurate modeling of viral proteins, our approach advances tools for understanding viral biology, combating emerging infectious diseases, and driving biotechnological innovation. 

Many chloroviruses replicate in Chlorella variabilis algal strains that are ex-endosymbionts isolated from the protozoan Paramecium bursaria, including the NC64A and Syngen 2-3 strains. We noticed that indigenous water samples produced a higher number of plaque-forming viruses on C. variabilis Syngen 2-3 lawns than on C. variabilis NC64A lawns. These observed differences led to the discovery of viruses that replicate exclusively in Syngen 2-3 cells, named Only Syngen (OSy) viruses. Here, we demonstrate that OSy viruses initiate infection in the restricted host NC64A by synthesizing some early virus gene products and that approximately 20% of the cells produce a small number of empty virus capsids. However, the infected cells did not produce infectious viruses because the cells were unable to replicate the viral genome. This is interesting because all previous attempts to isolate host cells resistant to chlorovirus infection were due to changes in the host receptor for the virus. 

ABSTRACT The ability of Bradyrhizobium spp. to nodulate and fix atmospheric nitrogen in soybean root nodules is critical to meeting humanity’s nutritional needs. The intricacies of soybean bradyrhizobia-plant interactions have been studied extensively; however, bradyrhizobial ecology as influenced by phages has received somewhat less attention, even though these interactions may significantly impact soybean yield. In batch culture, four soybean bradyrhizobia strains, Bradyrhizobium japonicum S06B (S06B-Bj), B. japonicum S10J (S10J-Bj), Bradyrhizobium diazoefficiens USDA 122 (USDA 122-Bd), and Bradyrhizobium elkanii USDA 76 T (USDA 76-Be), spontaneously (without apparent exogenous chemical or physical induction) produced tailed phages throughout the growth cycle; for three strains, phage concentrations exceeded cell numbers by ~3-fold after 48 h of incubation. Phage terminase large-subunit protein phylogeny revealed possible differences in phage packaging and replication mechanisms. Bioinformatic analyses predicted multiple prophage regions within each soybean bradyrhizobia genome, preventing accurate identification of spontaneously produced prophage (SPP) genomes. A DNA sequencing and mapping approach accurately delineated the boundaries of four SPP genomes within three of the soybean bradyrhizobia chromosomes and suggested that the SPPs were capable of transduction. In addition to the phages, S06B-Bj and USDA 76-Be contained three to four times more insertion sequences (IS) and large, conjugable, broad host range plasmids, both of which are known drivers of horizontal gene transfer (HGT) in soybean bradyrhizobia. These factors indicate that SPP along with IS and plasmids participate in HGT, drive bradyrhizobia evolution, and play an outsized role in bradyrhizobia ecology. IMPORTANCE Previous studies have shown that IS and plasmids mediate HGT of symbiotic nodulation ( nod ) genes in soybean bradyrhizobia; however, these events require close cell-to-cell contact, which could be limited in soil environments. Bacteriophage-assisted gene transduction through spontaneously produced prophages provides a stable means of HGT not limited by the constraints of proximal cell-to-cell contact. These phage-mediated HGT events may shape soybean bradyrhizobia population ecology, with concomitant impacts on soybean agriculture. 

Viruses face many challenges on their road to successful replication, and they meet those challenges by reprogramming the intracellular environment. Two major issues challenging Paramecium bursaria chlorella virus 1 (PBCV-1, genus Chlorovirus, family Phycodnaviridae) at the level of DNA replication are (i) the host cell has a DNA G+C content of 66%, while the virus is 40%; and (ii) the initial quantity of DNA in the haploid host cell is approximately 50 fg, yet the virus will make approximately 350 fg of DNA within hours of infection to produce approximately 1000 virions per cell. Thus, the quality and quantity of DNA (and RNA) would seem to restrict replication efficiency, with the looming problem of viral DNA synthesis beginning in only 60–90 min. Our analysis includes (i) genomics and functional annotation to determine gene augmentation and complementation of the nucleotide biosynthesis pathway by the virus, (ii) transcriptional profiling of these genes, and (iii) metabolomics of nucleotide intermediates. The studies indicate that PBCV-1 reprograms the pyrimidine biosynthesis pathway to rebalance the intracellular nucleotide pools both qualitatively and quantitatively, prior to viral DNA amplification, and reflects the genomes of the progeny virus, providing a successful road to virus infection. 

A large fraction of marine primary production is performed by diverse small protists, and many of these phytoplankton are phagotrophic mixotrophs that vary widely in their capacity to consume bacterial prey. Prior analyses suggest that mixotrophic protists as a group vary in importance across ocean environments, but the mechanisms leading to broad functional diversity among mixotrophs, and the biogeochemical consequences of this, are less clear. Here we use isolates from seven major taxa to demonstrate a tradeoff between phototrophic performance (growth in the absence of prey) and phagotrophic performance (clearance rate when consuming Prochlorococcus ). We then show that trophic strategy along the autotrophy-mixotrophy spectrum correlates strongly with global niche differences, across depths and across gradients of stratification and chlorophyll a . A model of competition shows that community shifts can be explained by greater fitness of faster-grazing mixotrophs when nutrients are scarce and light is plentiful. Our results illustrate how basic physiological constraints and principles of resource competition can organize complexity in the surface ocean ecosystem.