Characterising prophage diversity and function in the preterm infant pathogen, Staphylococcus epidermidis

Abstract

The human gut is a complex and dynamic ecosystem, shaped by a multitude of factors including diet, genetics, environment, and early microbial exposure. Initial microbial colonisation, thought to be at birth, is influenced significantly by environmental exposure and feeding practices. In preterm infants, this process is often disrupted, leading to colonisation by potentially pathogenic bacteria from sources such as the Neonatal Intensive Care Unit (NICU), increasing susceptibility to diseases such as necrotising enterocolitis (NEC) and late-onset sepsis (LOS). Staphylococcus epidermidis, a common skin commensal and frequent gut coloniser in preterm infants, is a leading cause of LOS due to its biofilm-forming ability and antibiotic resistance. Bacteriophages (phages), particularly temperate phages, further modulate gut microbial dynamics, and their interactions with bacterial hosts in the immature gut and can influence disease outcomes. Understanding these host– microbe–phage interactions is essential for improving neonatal gut health and preventing disease in vulnerable preterm populations. This thesis tested the hypothesis that prophage carriage and inducibility differ between S. epidermidis lineages and that antibiotic exposure in the NICU environment promotes prophage induction and potential dissemination of adaptive genes. To address this, the work aimed to evaluate and compare computational tools for prophage prediction in S. epidermidis, characterise prophage diversity and functional potential in both clinical and preterm infant isolates, and investigate the effects of antibiotic exposure on prophage induction and gene transfer potential.
To investigate bioinformatic tools for temperate phage (prophage) identification in sequenced S. epidermidis genomes, three prophage prediction tools were evaluated, which employ different computational approaches. The comparison used 89 humanassociated S. epidermidis genomes retrieved from the National Center for Biotechnology Information (NCBI). The investigation found that PHASTER overpredicted prophage regions in multiple instances by including bacterial host contamination in its predictions. Additionally, PHASTER predicted a substantial number of prophages that were not validated as phage by CheckV, and in some cases were identified as plasmids. ProphET yielded the fewest predictions and failed to identify two prophages, Staphylococcus phage SPBeta-like and PhiSepi-HH3, that were found in abundance by the other tools. VIBRANT had the most predictions with > 50% completeness as assessed by CheckV, and although it failed to predict some prophages that the others identified, it showed the most consistency. Thus, VIBRANT was chosen to be used in the subsequent studies. The 89 NCBI isolates were investigated further by characterising the S. epidermidis genomes and their predicted prophages. This analysis identified that the most abundant sequence types (STs) were ST2 and ST5, two of the most prevalent STs isolated from nosocomial infections which are commonly multi-drug resistant. VIBRANT predicted 249 prophage regions, an average of 2.8 per genome. ST2 had the greatest average prophage regions per genome. Annotation of the prophage regions revealed auxiliary metabolic genes (AMGs) and antibiotic resistance (AMR) genes in some prophage genomes, supporting the hypothesis that prophages contribute to host pathogenicity.
Building upon the understanding of prophage diversity in the NCBI’s collection of S. epidermidis genomes, attention was then directed to characterising prophages in S. epidermidis within the specific context of the preterm infant gut. To explore prophages infecting S. epidermidis in the preterm infant gut, 81 stool samples collected from very preterm infants (<32 weeks’ gestation) from the NICU in the Royal Victoria Infirmary, Newcastle upon Tyne, were sequenced using Illumina HiSeq X10 2x150bp chemistry. The data were assembled using a curated metagenomic assembly pipeline. Multilocus sequence typing (MLST) analysis identified 16 different STs, the most abundant being ST2, ST59 and ST439. 126 prophage regions were predicted and analysed for AMR, virulence determinants and auxiliary genes. ST2, a ST associated with nosocomial infections had the highest average prophages per genome. Genes that could aid the host’s fitness in disease were identified, and although their presence did not infer disease, they could still contribute to S. epidermidis’s pathogenicity. These findings indicate lineage-dependent differences in prophage content that may influence microbial adaptation and community structure within the preterm gut.
Given the widespread use of antibiotics in the NICU, and the hypothesis that such treatment may induce prophage activation, the effects of clinically relevant antibiotics on coagulase negative Staphylococcus (CoNS) strains isolated from preterm infants were examined. The samples were exposed to the minimum inhibitory concentration of clinically relevant antibiotics, and the resulting lysate was screened against multiple CoNS hosts to determine infection. The resulting data showed that various antibiotics were able to induce different prophages. Although, they seem strain-specific in their induction profiles. Inducible prophages also had acquired AMGs, showing that the induction of CoNS prophages within the neonatal gut could lead to the dissemination of these genes, and contribute to disease. These results support the hypothesis that antibiotic exposure can promote prophage-mediated gene transfer, potentially reshaping the resistome and functional capacity of the preterm infant gut microbiome.
Together, these findings enhance our understanding of prophage diversity in S. epidermidis and underscore the importance of integrating phage analysis into studies of the preterm infant gut microbiome. This research can enhance our understanding of prophage diversity and function within the preterm gut, which opens avenues for exploring targeted interventions to modulate the microbiome and improve infant health outcomes.

Date of Award	29 Oct 2025
Original language	English
Awarding Institution	Northumbria University
Supervisor	Andrew Nelson (Supervisor) & Darren Smith (Supervisor)