Metasecretome phage display : a new approach for mining surface and secreted proteins from microbial communities : a thesis presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biochemistry at Massey University, Palmerston North, New Zealand

Abstract

The microbial community residing in the reticulo-rumen degrades plant material to provide energy for its ruminant host. The key enzymes and proteins for plant fibre degradation are secreted from the microbial cells, and form part of the ‘metasecretome’ - the collection of cell-surface and secreted proteins that mediate important interactions between the microbiota and their rumen environment. Consequently, the metasecretome represents a valuable repository of bioactivities with potential applications in enhancing fibre digestibility and feed efficiency in ruminant animals, and in improving the depolymerisation of lignocellulosic feedstocks for biofuel production. A new metasecretome phage display approach was developed in this thesis, with the aim to focus sequencing efforts on the metasecretome-encoding component of complex microbial community genomes (metagenomes). This was achieved by combining secretome-selective phage display at a metagenomic scale with next-generation sequence analysis. The ability of this approach to focus metagenome mining onto sequences encoding surface and secreted proteins from the highly fibrolytic rumen plant-adherent microbiota of a dairy cow has been assessed. The metasecretome selection protocol efficiently enriched for a broad spectrum of metasecretome protein coding sequences, both in terms of the taxonomic and functional diversity, and the membrane-targeting signals present. This allowed in silico identification of functionally diverse surface and secreted carbohydrate-active enzymes (CAZymes). In particular, the metasecretome dataset was enriched for sequences encoding putative components characteristic of cellulosomes, the cell-surface multi-protein structures specialised for the degradation of plant fibre. Over one-sixth of the putative CAZymes identified in the metasecretome dataset shared a low sequence similarity with putative CAZymes identified through previous genomic and metagenomic studies; hence this work has identified proteins that potentially have novel carbohydrate-active functions. Affinity screening of the metagenomic phage display library on amorphous cellulose and arabinoxylan significantly enriched for a putative serine/threonine protein kinase. In silico analyses have not associated this protein with recognised carbohydrate binding functions, thus the observed binding may have not been carbohydrate specific. Overall, the methodology developed in this thesis is applicable for the high-throughput metasecretome exploration and is complementary to existing strategies used for mining surface and secreted proteins of complex microbial communities.