The function of a gene is actually the function of its product(s), and can be conveniently classified at three levels:
• The biochemical level. For example, a gene product may be described as a kinase or a calcium-binding protein. This reveals little about its wider role in the organism;
• The cellular level. This builds in information about intracellular localization and biological pathways. For example, it may be possible to establish that a protein is located in the nucleus and is required for DNA repair, even if its precise biochemical function is unknown;
• The organismal level. This may include information about where and when a gene is expressed in different tissues, and its role in the processes of development and/ or physiology.
To obtain a complete picture of gene function, information is required at all three levels. Defined vocabularies have been developed to describe gene function across all genomes. The most commonly used information resources here are the Gene Ontology (GO) system (http://www.geneontology.org/) and the Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/).
Useful information about a gene’s function can be obtained by using specific labeled probes or antibodies to track expression at the RNA and/or protein levels in cells and tissues, and to identify interactions with other proteins and nucleic acid sequences. However, the most comprehensive insights into gene function come from different types of genetic screens. In principle, genetic screens can be classified into two major headings:
• Forward genetic screen—the starting point is an abnormal phenotype (often induced by exposure to a mutagen), and the task is to identify the locus of the genetic change that has produced the phenotype, and so connect genotype with phenotype;
• Reverse genetic screen—the starting point is a defined gene or genes, and the task is to induce genetic changes in a defined gene(s) to produce an abnormal phenotype that provides information on how the gene(s) normally function.
In the post-genome era, large gene collections are available, and because the functions of many of them are not known or poorly understood, reverse genetic screens are now commonly deployed. They can be carried out in cultured cells, as described below. Inevitably, assaying gene function in cultured cells has its limitations. In particular, no detailed information can be obtained about how a gene functions in regulating processes that involve an interaction between different types of cell in the body, whether it be in the context of physiology (such as in the nervous system) or embryonic development. Defining gene function in this wider context requires the genetic manipulation of model organisms.
Understanding gene function in model organisms
The most powerful way to understand gene function is to manipulate genes in the germ line, and then analyze the phenotype of the whole organism so that information can be obtained at each of the three levels listed above. That is not an option in humans (although we may gain valuable information from studying people with genetic mutations). Instead, we rely heavily on extrapolation from genetic manipulation of various model organisms to infer the functions of human genes. That may be done in different ways: by introducing precise changes at the DNA level to alter a gene; by blocking gene expression at the RNA or protein level; or by introducing and expressing transgenes so that the gene product is either produced in abnormally large quantities (overexpression) or produced in cells or tissues where it is not normally expressed—see Figure 1 for an overview.
Fig1. Principal ways of studying gene function in vivo by genetically modifying model eukaryotes. (A) Most methods seek to inactivate a gene to produce a mutant phenotype that can be correlated with a specific gene (peach boxes). Usually this is done at the level of DNA, where large-scale random mutagenesis screens have been performed in many models, including both invertebrates (such as D. melanogaster) and vertebrates (notably zebrafish and mice). Specific knockouts of pre-determined genes have also been conducted, notably by using homologous recombination in mice. The expression of a specific gene can also be selectively inhibited by targeting its transcripts—large-scale RNA interference-based genetic screens have been conducted in C. elegans and D. melanogaster, and gene knockdown with antisense morpholino oligonucleotides is commonly performed in zebrafish embryos. Blocking proteins with dominant negative mutant proteins, aptamers, or intrabodies also has potential roles in treating disease (by specifically down-regulating harmful genes) and is discussed in more detail in Chapter 22. (B) An alternative approach is to use a transgene to overexpress a specific gene or to express it in tissues or developmental stages in which it is not normally expressed (ectopic expression) in an attempt to produce a phenotype that will provide clues to the function of the gene.
Germ-line gene inactivation is the most common approach to defining the function of a pre-determined gene in model organisms. Genetically amenable invertebrate models such as Drosophila melanogaster and Caenorhabditis elegans have been invaluable for understanding gene function, and even unicellular yeasts have provided valuable information that has illuminated how some well-conserved genes work. Inevitably, however, because these models are evolutionarily distantly related to us, they cannot provide information on many functionally important human genes with vertebrate specific functions.
Of the vertebrate model organisms, two models stand out because of their amenability to genetic manipulation. The zebrafish is advantaged by a short generation time, production of a large number of eggs at mating, external fertilization (so that all aspects of development are readily accessible), and a transparent embryo (facilitating identification of developmental mutants). However, the most useful model organism to infer human gene function has been the mouse: being a mammal, its physiology is very simi lar to ours, it has a short generation time and litter sizes are large, and it has long been amenable to genetic experimentation (mutants have been produced and studied over decades). In mice, inactivation of a pre-determined gene has typically been done by mutating the gene in embryonic stem cells (ESCs) to produce a null allele, a gene knock out, after which the mutant gene is introduced into the germ line by transferring the mutant ESCs into the inner cell mass of a blastocyst that is then implanted in the uterus of a foster mother.
Many research labs have produced mutant mice by knocking out a pre-determined gene of interest. Recently, however, the field has been transformed by a systematic and comprehensive approach toward understanding gene function that seeks to make the mouse the first mammal for which there will be a truly comprehensive functional catalog. Two key developments toward achieving this aim are listed below.
• The International Knockout Mouse Consortium (IKMC). Various USA and European research centers are collaborating to make a public resource of mouse embryonic stem cells that collectively contains a null mutation for every gene in the mouse genome, and the targeting strategies allow for producing both standard null mutations and conditional gene knockouts, where a null mutation can be activated later in development once the embryo has progressed through critical stages of early embryogenesis. For a general background to the IKMC, see http://www.mousephenotype.org/about-ikmc, and for the targeting strategies used to make the gene knockouts, see http://www.mousephenotype.org/about-ikmc/ targeting-strategies.
• The International Mouse Phenotyping Consortium (IMPC; at http://www. mousephenotype.org). The IMPC is generating a knockout mouse strain for every protein-coding gene by using the ES cell resource generated by the IKMC. Systematic broad-based phenotyping is performed by each IMPC center using standardized IMPReSS (International Mouse Phenotyping Resource of Standardised Screens) procedures (http://www.mousephenotype.org/impress).
Genetic screens in cultured cells
Although germ-line manipulation is the gold standard in functional screens, useful information can also be obtained by carrying out global screens of genetic function in cultured human and mammalian cells, and a variety of different reverse genetic screens can be carried out (for a background, see the review by Tochitani & Hayashizaki [2007], PMID 17308666, under Further Reading).
Until quite recently, the most commonly used genetic screens in human and mammalian cells depended on gene silencing using RNA interference to suppress gene expression. Sometimes that involves making siRNA libraries: pairs of complementary siRNAs are synthesized for each gene in the genome and deposited as individual pairs in wells of huge microtiter plate arrays. Alternatively, one makes libraries of genes encoding short hairpin RNA; the idea is to mimic the way in which miRNA is naturally made (it forms a short transcript that spontaneously forms a hairpin RNA that then undergoes cleavage).
Because RNAi-based silencing does not completely disrupt gene expression, the modern preference now is to use an alternative: genome-wide gene knockout screens using genome editing with the CRISPR-Cas9 system. As an example, see Shalem et al, (2014) (PMID 24336571).
