Proteome-wide Investigation Of The Relationship Between Domain Features And Expression Characters Of Proteins

Domains represent the basic structural, functional and evolutionary units of proteins. Genome-wide examination of domain distribution characteristics provides a key to understanding the mechanisms and principles of genome evolution and the increasing of organism's complexity. Genome sequencing of human and other model organisms have opened the flood-gate of the analysis of protein domain character via comparative genomic sequence analyses. The common and species-specific structural and functional features have been discovered in such studies. More importantly, some regular patterns of protein evolution related to domain duplication, accretion, and combination are also gotten to known. However, Since gene expression is regulated space- and stage-specifically, genome-level analyses can not represent the characters in certain tissues/organs, and can not reveal the relationship between domain character and the quantitative feature of proteins, which are essential for deeply understanding tissue-specific structures/functions and the rules of the composition and functions of a bio-entity. In this study, we focus on the proteome-wide relationship between domain features and expression characters of proteins.Second only to the brain in complexity, the liver is the home of many mysteries of life, including metabolism, detoxication, immunity, etc. The global cognition of the structural characters of human liver proteome, however, is deficient currently. For the first time, we systematically explore the domain distribution characters in a predicted Human Liver Proteome (pHLP) and focus on the structural causes of biological complexity of human liver. Global domain distribution analyses indicate that domains with high promiscuity or high connectivity and proteins with multiple domains are significantly enriched in pHLP compared to the predicted Human Genome Encoding Proteome (pHGEP). However, relatively new-age domains are significantly depleted in pHLP, and vice versa. These findings reveal that the biological complexity of human liver is more remarkably resulted from the diverse domain organization instead of new domain types. Further analyses at individual domain level disclose a list of specific domains (or domain combinations) significantly enriched or depleted in pHLP representing liver-characteristically physiological functions, and the character analyses of these domains confirm the causes of biological complexity mentioned above.The analyses above reveal the differential domain distribution pattern between pHGEP and pHLP. It is valuable to explore if the differential distribution pattern exist widely in other tissues/organs. To this end, the published gene atlas of 73 human tissues/organs was used to further analyses. The results reveal that the domain type distribution character widely exists in almost every tissue/organ, which indicates a basic rule for the gene expression of genome. Besides examining the correlation of domain characters and protein expression in the quality view, we also explored the relationship between domain character and protein abundance. Three simple but important parameters including domain number (DN), domain coverage (DC) and protein-protein interacting domain coverage (PPI_DC) were used to this analysis. We investigated how these characters relate to protein abundance using the quantitative proteome datasets of 6 representative species (H. sapiens, M. musculus, D. melanogaster, C. elegans, S. cerevisiae, E. coli). Interestingly, we found that DN negatively correlates with protein abundance especially in higher eukaryote, while DC and PPI_DC positively correlate with protein abundance. Previous studies have indicated that DN is the simplest parameter to represent the complexity of protein structure and function, and DC is a metric to quantify the functional compactness of protein domain architecture. Moreover, PPI_DC represents the complexity of interaction network the protein involved in. Thus, our results enrich the knowledge of the relationship between protein complexity/ compactness and its abundance.Systematic analyses of domain character can not only disclose important rules but also provide valuable information about some important individual research targets. In the previous studies, we found that KRAB domain containing proteins (KRAB-ZFPs) are significantly enriched in human fetal liver of 22 week-age but depleted in human adult liver. In addition, 57.6% KRAB-ZFPs express at the haematopoietic stage of mouse fetal liver. Considering erythroid differentiation is the main process in the haematopoiesis at fetal liver, these data imply that some of KRAB-ZFPs may take part in the regulation of erythroid differentiation.To test this hypothesis and to uncover the role of KRAB-ZFPs in erythroid differentiation, murine erythroleukemia (MEL) cells were used to our study. KAP-1, the universal co-repressor of KRAB-ZFPs, was knocked-down and the effect on adult erythroid differentiation was examined. We found RNAi-mediated knockdown of KAP-1 in MEL cells increase proportion of erythroid differentiation, especially for gene expression of embryonic globin during the HMBA -induced differentiation process. This indicates at least part of KAP-1 associated complexes negatively regulate embryonic globin gene during adult erythroid differentiation, and there may be some novel regulators, eg KRAB-ZFPs, interacting with KAP-1 to repress embryonic globin gene. To further discover these novel regulators, proteins interacting with KAP-1 was isolated by endogenous immunoprecipitation and identified by LC-ESI-MS/MS. In the list of identified proteins, we selected Zfp445 for further investigation and found Zfp445 can negatively regulate embryonic globin gene expression. But the mechanism of this negative regulation need further study.