Gut-innervating nociceptors regulate the intestinal microbiota to promote tissue protection.

Nociceptive pain is a hallmark of many chronic inflammatory conditions including inflammatory bowel diseases (IBDs); however, whether pain-sensing neurons influence intestinal inflammation remains poorly defined. Employing chemogenetic silencing, adenoviral-mediated colon-specific silencing, and pharmacological ablation of TRPV1+ nociceptors, we observed more severe inflammation and defective tissue-protective reparative processes in a murine model of intestinal damage and inflammation. Disrupted nociception led to significant alterations in the intestinal microbiota and a transmissible dysbiosis, while mono-colonization of germ-free mice with Gram+Clostridium spp. promoted intestinal tissue protection through a nociceptor-dependent pathway. Mechanistically, disruption of nociception resulted in decreased levels of substance P, and therapeutic delivery of substance P promoted tissue-protective effects exerted by TRPV1+ nociceptors in a microbiota-dependent manner. Finally, dysregulated nociceptor gene expression was observed in intestinal biopsies from IBD patients. Collectively, these findings indicate an evolutionarily conserved functional link between nociception, the intestinal microbiota, and the restoration of intestinal homeostasis.

The E3 ligases Itch and WWP2 cooperate to limit TH2 differentiation by enhancing signaling through the TCR.

The mechanisms by which the sensitivity of naive CD4+ T cells to stimulation by the cognate antigen via the T cell antigen receptor (TCR) determines their differentiation into distinct helper T cell subsets remain elusive. Here we demonstrate functional collaboration of the ubiquitin E3 ligases Itch and WWP2 in regulating the strength of the TCR signal. Mice lacking both Itch and WWP2 in T cells showed spontaneous autoimmunity and lung inflammation. CD4+ T cells deficient in Itch and WWP2 exhibited hypo-responsiveness to TCR stimulation and a bias toward differentiation into the TH2 subset of helper T cells. Itch and WWP2 formed a complex and cooperated to enhance TCR-proximal signaling by catalyzing the conjugation of atypical ubiquitin chains to the phosphatase SHP-1 and reducing the association of SHP-1 with the tyrosine kinase Lck. These findings indicate that targeted ubiquitination regulates the strength of the TCR signal and differentiation toward the TH2 lineage.

tRNA modification dynamics from individual organisms to metaepitranscriptomics of microbiomes.

tRNA is the most extensively modified RNA in cells. On average, a bacterial tRNA contains 8 modifications per molecule and a eukaryotic tRNA contains 13 modifications per molecule. Recent studies reveal that tRNA modifications are highly dynamic and respond extensively to environmental conditions. Functions of tRNA modification dynamics include enhanced, on-demand decoding of specific codons in response genes and regulation of tRNA fragment biogenesis. This review summarizes recent advances in the studies of tRNA modification dynamics in biological processes, tRNA modification erasers, and human-associated bacteria. Furthermore, we use the term "metaepitranscriptomics" to describe the potential and approach of tRNA modification studies in natural biological communities such as microbiomes. tRNA is highly modified in cells, and tRNA modifications respond extensively to environmental conditions to enhance translation of specific genes and produce tRNA fragments on demand. We review recent advances in tRNA sequencing methods, tRNA modification dynamics in biological processes, and tRNA modification studies in natural communities such as the microbiomes.

Interleukin-33 Induces the Enzyme Tryptophan Hydroxylase 1 to Promote Inflammatory Group 2 Innate Lymphoid Cell-Mediated Immunity.

Group 2 innate lymphoid cells (ILC2s) regulate immunity, inflammation, and tissue homeostasis. Two distinct subsets of ILC2s have been described: steady-state natural ILC2s and inflammatory ILC2s, which are elicited following helminth infection. However, how tissue-specific cues regulate these two subsets of ILC2s and their effector functions remains elusive. Here, we report that interleukin-33 (IL-33) promotes the generation of inflammatory ILC2s (ILC2INFLAM) via induction of the enzyme tryptophan hydroxylase 1 (Tph1). Tph1 expression was upregulated in ILC2s upon activation with IL-33 or following helminth infection in an IL-33-dependent manner. Conditional deletion of Tph1 in lymphocytes resulted in selective impairment of ILC2INFLAM responses and increased susceptibility to helminth infection. Further, RNA sequencing analysis revealed altered gene expression in Tph1 deficient ILC2s including inducible T cell co-stimulator (Icos). Collectively, these data reveal a previously unrecognized function for IL-33, Tph1, and ICOS in promoting inflammatory ILC2 responses and type 2 immunity at mucosal barriers.

Electrochemical reactor dictates site selectivity in N-heteroarene carboxylations.

Pyridines and related N-heteroarenes are commonly found in pharmaceuticals, agrochemicals and other biologically active compounds1,2. Site-selective C-H functionalization would provide a direct way of making these medicinally active products3-5. For example, nicotinic acid derivatives could be made by C-H carboxylation, but this remains an elusive transformation6-8. Here we describe the development of an electrochemical strategy for the direct carboxylation of pyridines using CO2. The choice of the electrolysis setup gives rise to divergent site selectivity: a divided electrochemical cell leads to C5 carboxylation, whereas an undivided cell promotes C4 carboxylation. The undivided-cell reaction is proposed to operate through a paired-electrolysis mechanism9,10, in which both cathodic and anodic events play critical roles in altering the site selectivity. Specifically, anodically generated iodine preferentially reacts with a key radical anion intermediate in the C4-carboxylation pathway through hydrogen-atom transfer, thus diverting the reaction selectivity by means of the Curtin-Hammett principle11. The scope of the transformation was expanded to a wide range of N-heteroarenes, including bipyridines and terpyridines, pyrimidines, pyrazines and quinolines.

Electrochemically driven cross-electrophile coupling of alkyl halides.

Recent research in medicinal chemistry has suggested that there is a correlation between an increase in the fraction of sp3 carbons-those bonded to four other atoms-in drug candidates and their improved success rate in clinical trials1. As such, the development of robust and selective methods for the construction of carbon(sp3)-carbon(sp3) bonds remains a critical problem in modern organic chemistry2. Owing to the broad availability of alkyl halides, their direct cross-coupling-commonly known as cross-electrophile coupling-provides a promising route towards this objective3-5. Such transformations circumvent the preparation of carbon nucleophiles used in traditional cross-coupling reactions, as well as stability and functional-group-tolerance issues that are usually associated with these reagents. However, achieving high selectivity in carbon(sp3)-carbon(sp3) cross-electrophile coupling remains a largely unmet challenge. Here we use electrochemistry to achieve the differential activation of alkyl halides by exploiting their disparate electronic and steric properties. Specifically, the selective cathodic reduction of a more substituted alkyl halide gives rise to a carbanion, which undergoes preferential coupling with a less substituted alkyl halide via bimolecular nucleophilic substitution to forge a new carbon-carbon bond. This protocol enables efficient cross-electrophile coupling of a variety of functionalized and unactivated alkyl electrophiles in the absence of a transition metal catalyst, and shows improved chemoselectivity compared with existing methods.

ILC3s select microbiota-specific regulatory T cells to establish tolerance in the gut.

Microbial colonization of the mammalian intestine elicits inflammatory or tolerogenic T cell responses, but the mechanisms controlling these distinct outcomes remain poorly understood, and accumulating evidence indicates that aberrant immunity to intestinal microbiota is causally associated with infectious, inflammatory and malignant diseases1-8. Here we define a critical pathway controlling the fate of inflammatory versus tolerogenic T cells that respond to the microbiota and express the transcription factor RORγt. We profiled all RORγt+ immune cells at single-cell resolution from the intestine-draining lymph nodes of mice and reveal a dominant presence of T regulatory (Treg) cells and lymphoid tissue inducer-like group 3 innate lymphoid cells (ILC3s), which co-localize at interfollicular regions. These ILC3s are distinct from extrathymic AIRE-expressing cells, abundantly express major histocompatibility complex class II, and are necessary and sufficient to promote microbiota-specific RORγt+ Treg cells and prevent their expansion as inflammatory T helper 17 cells. This occurs through ILC3-mediated antigen presentation, αV integrin and competition for interleukin-2. Finally, single-cell analyses suggest that interactions between ILC3s and RORγt+ Treg cells are impaired in inflammatory bowel disease. Our results define a paradigm whereby ILC3s select for antigen-specific RORγt+ Treg cells, and against T helper 17 cells, to establish immune tolerance to the microbiota and intestinal health.

Inulin fibre promotes microbiota-derived bile acids and type 2 inflammation.

Dietary fibres can exert beneficial anti-inflammatory effects through microbially fermented short-chain fatty acid metabolites^1,2, although the immunoregulatory roles of most fibre diets and their microbiota-derived metabolites remain poorly defined. Here, using microbial sequencing and untargeted metabolomics, we show that a diet of inulin fibre alters the composition of the mouse microbiota and the levels of microbiota-derived metabolites, notably bile acids. This metabolomic shift is associated with type 2 inflammation in the intestine and lungs, characterized by IL-33 production, activation of group 2 innate lymphoid cells and eosinophilia. Delivery of cholic acid mimics inulin-induced type 2 inflammation, whereas deletion of the bile acid receptor farnesoid X receptor diminishes the effects of inulin. The effects of inulin are microbiota dependent and were reproduced in mice colonized with human-derived microbiota. Furthermore, genetic deletion of a bile-acid-metabolizing enzyme in one bacterial species abolishes the ability of inulin to trigger type 2 inflammation. Finally, we demonstrate that inulin enhances allergen- and helminth-induced type 2 inflammation. Taken together, these data reveal that dietary inulin fibre triggers microbiota-derived cholic acid and type 2 inflammation at barrier surfaces with implications for understanding the pathophysiology of allergic inflammation, tissue protection and host defence.

Neuropeptide regulation of non-redundant ILC2 responses at barrier surfaces.

Emerging studies indicate that cooperation between neurons and immune cells regulates antimicrobial immunity, inflammation and tissue homeostasis. For example, a neuronal rheostat provides excitatory or inhibitory signals that control the functions of tissue-resident group 2 innate lymphoid cells (ILC2s) at mucosal barrier surfaces1-4. ILC2s express NMUR1, a receptor for neuromedin U (NMU), which is a prominent cholinergic neuropeptide that promotes ILC2 responses5-7. However, many functions of ILC2s are shared with adaptive lymphocytes, including the production of type 2 cytokines8,9 and the release of tissue-protective amphiregulin (AREG)10-12. Consequently, there is controversy regarding whether innate lymphoid cells and adaptive lymphocytes perform redundant or non-redundant functions13-15. Here we generate a new genetic tool to target ILC2s for depletion or gene deletion in the presence of an intact adaptive immune system. Transgenic expression of iCre recombinase under the control of the mouse Nmur1 promoter enabled ILC2-specific deletion of AREG. This revealed that ILC2-derived AREG promotes non-redundant functions in the context of antiparasite immunity and tissue protection following intestinal damage and inflammation. Notably, NMU expression levels increased in inflamed intestinal tissues from both mice and humans, and NMU induced AREG production in mouse and human ILC2s. These results indicate that neuropeptide-mediated regulation of non-redundant functions of ILC2s is an evolutionarily conserved mechanism that integrates immunity and tissue protection.

BA.2.12.1, BA.4 and BA.5 escape antibodies elicited by Omicron infection.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron sublineages BA.2.12.1, BA.4 and BA.5 exhibit higher transmissibility than the BA.2 lineage1. The receptor binding and immune-evasion capability of these recently emerged variants require immediate investigation. Here, coupled with structural comparisons of the spike proteins, we show that BA.2.12.1, BA.4 and BA.5 (BA.4 and BA.5 are hereafter referred collectively to as BA.4/BA.5) exhibit similar binding affinities to BA.2 for the angiotensin-converting enzyme 2 (ACE2) receptor. Of note, BA.2.12.1 and BA.4/BA.5 display increased evasion of neutralizing antibodies compared with BA.2 against plasma from triple-vaccinated individuals or from individuals who developed a BA.1 infection after vaccination. To delineate the underlying antibody-evasion mechanism, we determined the escape mutation profiles2, epitope distribution3 and Omicron-neutralization efficiency of 1,640 neutralizing antibodies directed against the receptor-binding domain of the viral spike protein, including 614 antibodies isolated from people who had recovered from BA.1 infection. BA.1 infection after vaccination predominantly recalls humoral immune memory directed against ancestral (hereafter referred to as wild-type (WT)) SARS-CoV-2 spike protein. The resulting elicited antibodies could neutralize both WT SARS-CoV-2 and BA.1 and are enriched on epitopes on spike that do not bind ACE2. However, most of these cross-reactive neutralizing antibodies are evaded by spike mutants L452Q, L452R and F486V. BA.1 infection can also induce new clones of BA.1-specific antibodies that potently neutralize BA.1. Nevertheless, these neutralizing antibodies are largely evaded by BA.2 and BA.4/BA.5 owing to D405N and F486V mutations, and react weakly to pre-Omicron variants, exhibiting narrow neutralization breadths. The therapeutic neutralizing antibodies bebtelovimab4 and cilgavimab5 can effectively neutralize BA.2.12.1 and BA.4/BA.5, whereas the S371F, D405N and R408S mutations undermine most broadly sarbecovirus-neutralizing antibodies. Together, our results indicate that Omicron may evolve mutations to evade the humoral immunity elicited by BA.1 infection, suggesting that BA.1-derived vaccine boosters may not achieve broad-spectrum protection against new Omicron variants.

Mapping genomic loci implicates genes and synaptic biology in schizophrenia.

Schizophrenia has a heritability of 60-80%1, much of which is attributable to common risk alleles. Here, in a two-stage genome-wide association study of up to 76,755 individuals with schizophrenia and 243,649 control individuals, we report common variant associations at 287 distinct genomic loci. Associations were concentrated in genes that are expressed in excitatory and inhibitory neurons of the central nervous system, but not in other tissues or cell types. Using fine-mapping and functional genomic data, we identify 120 genes (106 protein-coding) that are likely to underpin associations at some of these loci, including 16 genes with credible causal non-synonymous or untranslated region variation. We also implicate fundamental processes related to neuronal function, including synaptic organization, differentiation and transmission. Fine-mapped candidates were enriched for genes associated with rare disruptive coding variants in people with schizophrenia, including the glutamate receptor subunit GRIN2A and transcription factor SP4, and were also enriched for genes implicated by such variants in neurodevelopmental disorders. We identify biological processes relevant to schizophrenia pathophysiology; show convergence of common and rare variant associations in schizophrenia and neurodevelopmental disorders; and provide a resource of prioritized genes and variants to advance mechanistic studies.

Global histone H2B degradation regulates insulin/IGF signaling-mediated nutrient stress.

Eukaryotic organisms adapt to environmental fluctuations by altering their epigenomic landscapes and transcriptional programs. Nucleosomal histones carry vital epigenetic information and regulate gene expression, yet the mechanisms underlying chromatin-bound histone exchange remain elusive. Here, we found that histone H2Bs are globally degraded in Caenorhabditis elegans during starvation. Our genetic screens identified mutations in ubiquitin and ubiquitin-related enzymes that block H2B degradation in starved animals, identifying lysine 31 as the crucial residue for chromatin-bound H2B ubiquitination and elimination. Retention of aberrant nucleosomal H2B increased the association of the FOXO transcription factor DAF-16 with chromatin, generating an ectopic gene expression profile detrimental to animal viability when insulin/IGF signaling was reduced in well-fed animals. Furthermore, we show that the ubiquitin-proteasome system regulates chromosomal histone turnover in human cells. During larval development, C. elegans epidermal cells undergo H2B turnover after fusing with the epithelial syncytium. Thus, histone degradation may be a widespread mechanism governing dynamic changes of the epigenome.

E3 Ligase VHL Promotes Group 2 Innate Lymphoid Cell Maturation and Function via Glycolysis Inhibition and Induction of Interleukin-33 Receptor.

Group 2 innate lymphoid cells (ILC2s) are a specialized subset of lymphoid effector cells that are critically involved in allergic responses; however, the mechanisms of their regulation remain unclear. We report that conditional deletion of the E3 ubiquitin ligase VHL in innate lymphoid progenitors minimally affected early-stage bone marrow ILC2s but caused a selective and intrinsic decrease in mature ILC2 numbers in peripheral non-lymphoid tissues, resulting in reduced type 2 immune responses. VHL deficiency caused the accumulation of hypoxia-inducible factor 1α (HIF1α) and attenuated interleukin-33 (IL-33) receptor ST2 expression, which was rectified by HIF1α ablation or inhibition. HIF1α-driven expression of the glycolytic enzyme pyruvate kinase M2 downmodulated ST2 expression via epigenetic modification and inhibited IL-33-induced ILC2 development. Our study indicates that the VHL-HIF-glycolysis axis is essential for the late-stage maturation and function of ILC2s via targeting IL-33-ST2 pathway.

Glycine decarboxylase activity drives non-small cell lung cancer tumor-initiating cells and tumorigenesis.

Identification of the factors critical to the tumor-initiating cell (TIC) state may open new avenues in cancer therapy. Here we show that the metabolic enzyme glycine decarboxylase (GLDC) is critical for TICs in non-small cell lung cancer (NSCLC). TICs from primary NSCLC tumors express high levels of the oncogenic stem cell factor LIN28B and GLDC, which are both required for TIC growth and tumorigenesis. Overexpression of GLDC and other glycine/serine enzymes, but not catalytically inactive GLDC, promotes cellular transformation and tumorigenesis. We found that GLDC induces dramatic changes in glycolysis and glycine/serine metabolism, leading to changes in pyrimidine metabolism to regulate cancer cell proliferation. In the clinic, aberrant activation of GLDC correlates with poorer survival in lung cancer patients, and aberrant GLDC expression is observed in multiple cancer types. This link between glycine metabolism and tumorigenesis may provide novel targets for advancing anticancer therapy.

Folds from fold: Exploring topological isoforms of a single-domain protein.

Expanding the protein fold space beyond linear chains is of fundamental significance, yet remains largely unexplored. Herein, we report the creation of seven topological isoforms (i.e., linear, cyclic, knot, lasso, pseudorotaxane, and catenane) from a single protein fold precursor by rewiring the connectivity of secondary structure elements of the SpyTag-SpyCatcher complex and mutating the reactive residue on SpyTag to abolish the isopeptide bonding. These topological isoforms can be directly expressed in cells. Their topologies were confirmed by combined techniques of proteolytic digestion, fluorescence correlation spectroscopy (FCS), size-exclusion chromatography (SEC), and topological transformation. To study the effects of topology on their structures and properties, their biophysical properties were characterized by differential scanning calorimetry (DSC), heteronuclear single quantum coherence nuclear magnetic resonance spectroscopy (HSQC-NMR), and circular dichroism (CD) spectroscopy. Molecular dynamics (MD) simulations were further performed to reveal the atomic details of structural changes upon unfolding. Both experimental and simulation results suggest that they share a similar, well-folded hydrophobic core but exhibit distinct folding/unfolding dynamic behaviors. These results shed light onto the folding landscape of topological isoforms derived from the same protein fold. As a model system, this work improves our understanding of protein structure and dynamics beyond linear chains and suggests that protein folds are highly amenable to topological variation.

The impact of age and number of mutations on the size of clonal hematopoiesis.

Clonal hematopoiesis (CH) represents the clonal expansion of hematopoietic stem cells and their progeny driven by somatic mutations. Accurate risk assessment of CH is critical for disease prevention and clinical decision-making. The size of CH has been showed to associate with higher disease risk, yet, factors influencing the size of CH are unknown. In addition, the characteristics of CH in long-lived individuals are not well documented. Here, we report an in-depth analysis of CH in longevous (≥90 y old) and common (60~89 y old) elderly groups. Utilizing targeted deep sequencing, we found that the development of CH is closely related to age and the expression of aging biomarkers. The longevous elderly group exhibited a significantly higher incidence of CH and significantly higher frequency of TET2 and ASXL1 mutations, suggesting that certain CH could be beneficial to prolong life. Intriguingly, the size of CH neither correlates significantly to age, in the range of 60 to 110 y old, nor to the expression of aging biomarkers. Instead, we identified a strong correlation between large CH size and the number of mutations per individual. These findings provide a risk assessment biomarker for CH and also suggest that the evolution of the CH is influenced by factor(s) in addition to age.

KCTD10 regulates brain development by destabilizing brain disorder-associated protein KCTD13.

KCTD10 belongs to the KCTD (potassiumchannel tetramerization domain) family, many members of which are associated with neuropsychiatric disorders. However, the biological function underlying the association with brain disorders remains to be explored. Here, we reveal that Kctd10 is highly expressed in neuronal progenitors and layer V neurons throughout brain development. Kctd10 deficiency triggers abnormal proliferation and differentiation of neuronal progenitors, reduced deep-layer (especially layer V) neurons, increased upper-layer neurons, and lowered brain size. Mechanistically, we screened and identified a unique KCTD10-interacting protein, KCTD13, associated with neurodevelopmental disorders. KCTD10 mediated the ubiquitination-dependent degradation of KCTD13 and KCTD10 ablation resulted in a considerable increase of KCTD13 expression in the developing cortex. KCTD13 overexpression in neuronal progenitors led to reduced proliferation and abnormal cell distribution, mirroring KCTD10 deficiency. Notably, mice with brain-specific Kctd10 knockout exhibited obvious motor deficits. This study uncovers the physiological function of KCTD10 and provides unique insights into the pathogenesis of neurodevelopmental disorders.

High-throughput identification of regulatory elements and functional assays to uncover susceptibility genes for nasopharyngeal carcinoma.

Large-scale genetic association studies have identified multiple susceptibility loci for nasopharyngeal carcinoma (NPC), but the underlying biological mechanisms remain to be explored. To gain insights into the genetic etiology of NPC, we conducted a follow-up study encompassing 6,907 cases and 10,472 controls and identified two additional NPC susceptibility loci, 9q22.33 (rs1867277; OR = 0.74, 95% CI = 0.68-0.81, p = 3.08 × 10-11) and 17q12 (rs226241; OR = 1.42, 95% CI = 1.26-1.60, p = 1.62 × 10-8). The two additional loci, together with two previously reported genome-wide significant loci, 5p15.33 and 9p21.3, were investigated by high-throughput sequencing for chromatin accessibility, histone modification, and promoter capture Hi-C (PCHi-C) profiling. Using luciferase reporter assays and CRISPR interference (CRISPRi) to validate the functional profiling, we identified PHF2 at locus 9q22.33 as a susceptibility gene. PHF2 encodes a histone demethylase and acts as a tumor suppressor. The risk alleles of the functional SNPs reduced the expression of the target gene PHF2 by inhibiting the enhancer activity of its long-range (4.3 Mb) cis-regulatory element, which promoted proliferation of NPC cells. In addition, we identified CDKN2B-AS1 as a susceptibility gene at locus 9p21.3, and the NPC risk allele of the functional SNP rs2069418 promoted the expression of CDKN2B-AS1 by increasing its enhancer activity. The overexpression of CDKN2B-AS1 facilitated proliferation of NPC cells. In summary, we identified functional SNPs and NPC susceptibility genes, which provides additional explanations for the genetic association signals and helps to uncover the underlying genetic etiology of NPC development.

Quantification of tRNA m1A modification by templated-ligation qPCR.

N1-methyladenosine (m1A) is a widespread modification in all eukaryotic, many archaeal, and some bacterial tRNAs. m1A is generally located in the T loop of cytosolic tRNA and between the acceptor and D stems of mitochondrial tRNAs; it is involved in the tertiary interaction that stabilizes tRNA. Human tRNA m1A levels are dynamically regulated that fine-tune translation and can also serve as biomarkers for infectious disease. Although many methods have been used to measure m1A, a PCR method to assess m1A levels quantitatively in specific tRNAs has been lacking. Here we develop a templated-ligation followed by a qPCR method (TL-qPCR) that measures m1A levels in target tRNAs. Our method uses the SplintR ligase that efficiently ligates two tRNA complementary DNA oligonucleotides using tRNA as the template, followed by qPCR using the ligation product as the template. m1A interferes with the ligation in specific ways, allowing for the quantitative assessment of m1A levels using subnanogram amounts of total RNA. We identify the features of specificity and quantitation for m1A-modified model RNAs and apply these to total RNA samples from human cells. Our method enables easy access to study the dynamics and function of this pervasive tRNA modification.

DeepPBI-KG: a deep learning method for the prediction of phage-bacteria interactions based on key genes.

Phages, the natural predators of bacteria, were discovered more than 100 years ago. However, increasing antimicrobial resistance rates have revitalized phage research. Methods that are more time-consuming and efficient than wet-laboratory experiments are needed to help screen phages quickly for therapeutic use. Traditional computational methods usually ignore the fact that phage-bacteria interactions are achieved by key genes and proteins. Methods for intraspecific prediction are rare since almost all existing methods consider only interactions at the species and genus levels. Moreover, most strains in existing databases contain only partial genome information because whole-genome information for species is difficult to obtain. Here, we propose a new approach for interaction prediction by constructing new features from key genes and proteins via the application of K-means sampling to select high-quality negative samples for prediction. Finally, we develop DeepPBI-KG, a corresponding prediction tool based on feature selection and a deep neural network. The results show that the average area under the curve for prediction reached 0.93 for each strain, and the overall AUC and area under the precision-recall curve reached 0.89 and 0.92, respectively, on the independent test set; these values are greater than those of other existing prediction tools. The forward and reverse validation results indicate that key genes and key proteins regulate and influence the interaction, which supports the reliability of the model. In addition, intraspecific prediction experiments based on Klebsiella pneumoniae data demonstrate the potential applicability of DeepPBI-KG for intraspecific prediction. In summary, the feature engineering and interaction prediction approaches proposed in this study can effectively improve the robustness and stability of interaction prediction, can achieve high generalizability, and may provide new directions and insights for rapid phage screening for therapy.