Investigating the Human Intestinal Virome

Hsieh, Shen-Yuan (2020) Investigating the Human Intestinal Virome. Doctoral thesis, University of East Anglia.

Preview

PDF Download (17MB) | Preview

Abstract

The human intestinal virobiota consists of highly complex and diverse viruses and virus-associated genes (termed the “virome”), dominated by bacteriophages. Numerically, viruses have been considered the most abundant and diverse biological entities on Earth, estimated to be approximately 1031 in number. In the human gastrointestinal tract (GIT), virus-to-microbe ratio (VMR) may be close to 1:1, while it may reach 20:1 at mucosal surfaces and within the mucus layer, in total numbering 1010-1015 virus-like particles (VLPs). Recent studies suggested that changes in the intestinal virome may lead to chronic GI-inflammation and intestinal microbial dysbiosis, thereby triggering diseases such as myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Thus, this thesis aimed to develop robust and reliable protocols for characterising the human faecal microbiome that can be applied to analysing the virome in patients with severe ME/CFS.

This thesis first aimed to develop a reliable and reproducible protocol for VLP isolation from human faeces and for VLP quantification using a digital image analysis (DIA)-based method. Protocols were then optimised for VLP DNA extraction to obtain DNA of sufficient quality and quantity for next generation sequencing (NGS). As part of these studies, I developed a bioinformatics pipeline for viral metagenomic analysis, and also, I determined the extent of PCR amplification bias in virome-enriched, uncultivated virus genomes (UViGs) by comparing the methods of linker amplified shotgun library (LASL) and non-amplified shotgun library (NASL) preparations. The optimised protocols and bioinformatics pipeline were then applied to the initial analysis of the faecal virome of severely affected ME/CFS patients and same household healthy control individuals (SHHC).

The optimised protocol is comprised of (1) homogenisation of faecal samples by vortexing without the use of bead-beating, followed by incubation on ice to facilitate the release of VLPs from solid materials; (2) partition of crude faecal matter, dietary debris and virions/VLPs by two-round of high-speed centrifugation; (3) sequential filtration using 0.8 μm and 0.45 μm filter; (4) PEG precipitation; (5) DNase and RNase treatment; (6) proteinase K digestion; (7) viral capsid lysis with SDS lysis buffer; (8) Phenol/Chloroform/Isoamyl alcohol extraction; (9) DNA purification using silica-based spin columns, and (10) DNA concentration using a vacuum-based condenser. Using three independent stool samples to evaluate reproducibility, VLP DNA yields were between 67.2 ng and 94.8 ng per gram of faeces. For VLP quantification, manual counting-based DIA method was more accurate and reliable than automated counting-based method.

Using an optimised bioinformatics pipeline to analyse UViGs from PCR and non-PCR virome-derived datasets, I found that misrepresentation of certain viruses may occur after amplification in their relative abundance. In alpha diversity, the UViGs from non-PCR datasets generally have higher richness and diversity than those from PCR datasets, suggesting that PCR is likely to lower viral richness and diversity. Moreover, the major differences in beta diversity were more likely to be driven by a high level of intestinal virome individuality between donors, while amplification bias may have a minor effect on the beta diversity of viruses in the PCR datasets. In addition, in an initial analysis of comparing UViG similarity networks, I found that there is no significant difference between both datasets but further investigation is required.

In comparing faecal samples from ME/CFS and SHHC, VLPs with nucleic acid-containing capsids in SHHC samples were higher than those of severe ME/CFS patients, although the variation and diversity of VLP were seen in both sets of faecal samples. Transmission electron microscopy (TEM) analysis identified Siphoviridae as the most prominent virus in both ME/CFS patients and SHHC VLP samples. Moreover, giant Siphoviruses were occasionally detected, suggesting potential novel strains are present in these samples. The biological meaning of these findings is not clear and requires further investigation. In ongoing work, the optimised protocol and bioinformatic pipeline is being applied to investigating the composition of the intestinal virome in severe ME/CFS patients and SHHC.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Medicine and Health Sciences > Norwich Medical School
Depositing User:	Chris White
Date Deposited:	18 Oct 2021 09:14
Last Modified:	31 Dec 2021 01:38
URI:	https://ueaeprints.uea.ac.uk/id/eprint/81750
DOI:

Downloads

Actions (login required)

View Item