Brigham and Womens Hospital Genetics Division Boston, MA 02215
Recent progress in the development of resources for genomic analysis, highlighted by the release of both proprietary (Celera) and public domain assemblies of the mouse genome (http://www.ncbi.nlm.nih.gov/genome/guide/mouse), has been paralleled by a renewed enthusiasm for phenotype-driven strategies for gene discovery, such as mutagenesis using treatment with the chemical N-ethylnitrosourea (ENU). One remarkable aspect of this revival is the emergence of diverse strategies for the application of ENU mutagenesis, with respect to both the phenotypes being studied and the organization of research programs. This variety embraces both dominant and recessive screens, large and small programs, and phenotypic analyses ranging from examination for developmental defects to cell biological and biochemical assays to analysis of physiology and behavior. The recent Workshop on ENU Mutagenesis, held July 2528, 2002, and sponsored by the US Department of Energy, Celltech R and D, Inc., and the National Institutes of Health, provided an excellent opportunity to review progress in this field.
Successes
Treatment with ENU is a highly efficient means to generate mutations. With some caveats (see below), it can be reported that ENU mutagenesis is a highly productive method for generating interesting mutant mice. In fact, in many respects it has proven more effective than investigators anticipated, which has created a logistical challenge with respect to the need for simultaneous husbandry, characterization, distribution, and archiving of many lines of mice. However, this success has also emboldened investigators to target more finely cut phenotypes in their investigations by employing more specific assays. These include, for example, responses to specific pharmacological agents and analysis of subtle patterning differences using transgenic reporter genes. The near-universal success also confirms that the standard protocols for mutagenesis are robust. Most investigators employ a fractionated treatment regiment of three weekly intraperitoneal injections, although investigators at Harwell (http://www.mut.har.mrc.ac.uk/) continue to have success using a single treatment regimen.
ENU-induced mutations can be cloned.
What a difference a year makes. Investigators affiliated with programs large and small, academic and commercial, report success in identifying sequence changes in mutated genes. A number of factors have contributed to this. First and foremost has been progress in developing tools and resources for mouse genomic analysis (see, for example, the UCSC mouse genome browser at http://genome.ucsc.edu/cgi-bin/hgGateway?db=mm2). Note that while the assembly of a draft mouse genome sequence is undoubtedly the most useful of these, a number of investigators succeeded in identifying ENU-induced mutant loci prior to its completion. The decreasing cost and increasing capacity for DNA sequencing has facilitated a fairly straightforward approach in which cDNAs and/or genomic regions containing exons (including splice junctions) of candidate genes are sequenced even when a mutation has been resolved to only moderate resolution (for example, 15 cM). This approach will likely be complemented by updated techniques for heteroduplex detection (for example, using an instrument developed by SpectruMedix for capillary-based temperature gradient electrophoresis; see http://www.spectrumedix.com/). However, it is clear that genetic mapping, even to moderate resolution, presents a substantial logistical challenge, as discussed below.
ENU-induced mutations are biologically informative.
The primary premise behind a phenotype-driven approach such as ENU mutagenesis is the presumption that it allows an unbiased investigation of the genetic bases of a heritable trait. Even early results in the characterization of ENU-induced mutations demonstrate that this aim is amply fulfilled. These results include both the identification of the biological consequences of mutations in uncharacterized genes and the assignment of novel roles to previously studied genes. In many cases the connection between the cloned genes and the mutant phenotypes is unexpected, validating the application of an analytic strategy that is not biased by presumptions regarding gene function. Notable among these is research by Katherine Andersons laboratory characterizing genes that affect neural development (1, 3). Among the genes characterized in their analysis are several that might have been considered to have mundane "housekeeping" functions; yet have now been shown to have specific roles in the Sonic hedgehog signaling pathway.
In addition, the fact that ENU can induce both null and hypomorphic mutations has proven to be extremely productive in characterizing the biological role of even well-studied genes. For example, in our own research, we identified a line of mice that has an ENU-induced base change in a splice site of the FOG2 (Zfpm2) transcription cofactor. Although most of the homozygous mutant mice die in midgestation with severe heart defects similar to those found in the previously described knockouts of this gene, a subset survive to a late gestation but have severe diaphragm and lung defects (Ackerman KG, Herron BJ, Rao C, Huang H, Kochilas L, Epstein JA, Babiuk RP, Greer JJ, and Beier DR, unpublished observations). The role of FOG2 in the development of these organs was not previously recognized.
Challenges
ENU mutagenesis and complex phenotypes.
Treatment with ENU has proven to be a potent method for generating mutations that affect organogenesis, and success has been reported in the analysis of specific systems including the neural tube, skin, eye, ear, and skeleton. In addition, mutant lines with heritable defects in immune function, cell biological phenotypes, and serum biochemistry have been identified, and in several cases, the mutant gene has been identified. However, progress in applying ENU mutagenesis to more complex physiological, neurological, and behavioral phenotypes has been slower. Given the ambitious nature of these programs, this is not entirely unsurprising, and a numbers of factors have contributed. One is the substantial logistical problem of scaling up phenotype assays that are time- and technology-intensive to be compatible with a high-throughput protocol. Another factor is the task of differentiating a true mutant from a normal outlier when using a quantitative or subjective assay. As confidence in the potency of mutagenesis technology increases, this can be addressed by raising the threshold of what is considered as variant, which will increase the yield of true mutant lines. Results presented from the Northwestern University Center for Functional Genomics clearly showed that a "hands-free" phenotyping program (based on video recording of mouse activity) could identify mice with unambiguously variant behavior which has proved heritable (http://genome.northwestern.edu/).
Utilization of mutant lines.
One expectation regarding ENU mutagenesis programs was that the mutant lines obtained would be considered a valuable resource by the larger community. While numerous productive collaborations have been forged, it is also clear that these novel mutant lines have not been as widely sought as anticipated. Several factors have contributed, including lack of awareness (despite multiple publications and informative websites), generic difficulties in mouse shipment and import, and reluctance by investigators to undertake the mapping and positional cloning tasks required for mutation characterization. Lack of map information for a mutation represents a significant disincentive, since there is little enthusiasm for characterizing a potential remutation of a known gene. Several approaches to mapping ENU-induced mutations are being employed, including high-throughput genome-wide mapping protocols (likely to become SNP-based as mouse genome sequence characterization continues), haplotype-based approaches requiring small numbers of affected animals (5), and region-specific protocols using deletions or, more successfully, inversions as balancer chromosomes (Ref. 6; http://www.mouse-genome.bcm.tmc.edu/ENU/MutagenesisProjDes.asp). However, the logistics and costs for mapping crosses do not scale well and potentially impose a significant burden on the centers generating large numbers of mutants.
Relative utility of ENU mutagenesis.
Given the rapid development and characterization of the mouse genome sequence, coupled with potent new technologies for generating targeted mutations (2) or sequence-tagged gene-trap reagents (4), it is necessary to examine the utility of committing large amounts of resources to generating mutations that require a substantial cloning effort for their characterization. This is particularly the case as investigators (and funding agencies) consider increasingly feasible high-throughput strategies to generate mutations of every known or presumptive gene. These are increasingly attractive to the extent they are embryonic stem (ES)-cell based, as this can simplify distribution issues (although they require that investigators have ready access to cost-effective facilities for blastocyst injection, which is not universally the case). However, this expedience must be balanced against the utility of obtaining unbiased insight into the genetic basis of a trait. As previously noted, even early efforts have yielded unexpected insights and have identified mutations in novel genes whose function were otherwise unknown. The importance of this cannot be overstated, as this approach is virtually guaranteed to identify genes that have fundamental biological roles and, furthermore, has the potential for identifying novel pathways for therapeutic intervention in human disease.
Moving Forward
Results from screens employing developmental analyses and cell biological assays demonstrate that ENU mutagenesis protocols are robust. This will prove fruitful not only in the near-term, but will continue to be productive as investigators ask more focused questions using specific assays. These results also demonstrate how crucial it is to have assays that distinguish unambiguously normal variants from induced mutations and are amenable to high throughput. Although ENU mutagenesis is highly efficient, it is still necessary to look at hundreds to thousands of normal mice (for any individual phenotype) before identifying a true mutant line. The constraints on phenotyping are a particular challenge for complex physiological and behavioral phenotypes. For these investigations, it may require a better marriage between investigators with experience in these studies, to devise appropriate assays, and geneticists, to ensure that support is present for mapping and cloning of identified genes. (In many cases this has been achieved, and it is simply necessary to allow sufficient time for these programs to become fully active.) The European Community has proposed to approach this with an ambitious program called Eumorphia (http://www.eumorphia.org/), in which they plan a distributed effort to identify strategies for phenotype definition and assay. The US funding agencies and their advisors will have to determine how their resources will be split between analogous centralized programs vs. more specific investigator-driven research. This review becomes more complicated as it is necessary to consider both the relative utility of the rapidly evolving tools for genome manipulation and the inherent logistical complexities of larger mouse-intensive efforts. Although our present dilemma is the need to find the right balance between these phenomenally potent technologies, the positive aspect of this challenge will be the remarkable opportunities that exist for the study of mammalian biology.
REFERENCES