Uses and Opportunities for Ethyl Methanesulfonate Mutagenesis in Maize

MUTAGENESIS METHODS

Multiple approaches exist for making mutants in maize. The most widely used is mutagenesis by chemical treatments that alkylate DNA base residues. The addition of an alkyl group to a nucleoside, the base residue in DNA, interrupts canonical base-pairing and ultimately causes the misincorporation of an alternative nucleotide residue during semiconservative DNA replication (Meselson and Stahl 1958). Thus, these structural changes increase the mutation rate in a replication-dependent manner and typically result in heritable single-nucleotide substitutions. As such, chemical mutagenesis can produce a series of mutant alleles that vary in phenotypic severity. Multiple chemicals resulting in a variety of different alkylations with varied preferences for the four bases and positions on each base cause idiosyncratic patterns of mutation accumulation (Singer 1975; Horsfall et al. 1990; Shrivastav et al. 2010; Williams 2016). The most commonly used chemical mutagen in maize is EMS, an alkylating agent and potent mutagen that reacts specifically with the hydroxyl on the sixth position of a guanine residue to produce O-6-ethylguanine (Loveless 1958; Brookes and Lawley 1961). The adoption of EMS for maize mutagenesis owes much to the work of Dr. Gerry Neuffer and Dr. Ed Coe (Neuffer and Coe 1978), who generated improved protocols for EMS mutagenesis in maize and a large number of maize mutants via this technique. O-6-ethylguanine pairs readily with thymine during semiconservative DNA replication. In the subsequent replication event, the thymine residue will pair with adenine, and the granddaughter cell will receive a copy of the gene with an A:T pair, thereby resulting in a G-to-A transition. The mutation spectrum induced by EMS is overwhelmingly G-to-A, and the complementary C-to-T change, with >99% of mutations induced by EMS in plants exhibiting this change (Tsai et al. 2011; Henry et al. 2014; Addo-Quaye et al. 2017). However, insertion and deletion events can also result from EMS mutagenesis, and are more likely than point mutations to result in phenotypic impacts. When alleles of phenotypic impact are screened and molecularly characterized in model systems following EMS mutagenesis, a significant percentage (3%–7%) contain indels, presumably from failures in the repair of DNA breaks (Coulondre and Miller 1977; Schaaper and Dunn 1991).

Most other mutagenesis approaches use insertion or deletion as the primary mechanisms of genomic change. As a result, the spectrum of phenotypes returned is different, as insertion and deletion events are more likely to generate knockout alleles. Point mutations, which are generated by alkylating agents like EMS, result in both knockout alleles and hypomorphic alleles due to protein coding changes. The spectrum of phenotypes can reflect what is referred to as an allelic series, from mutations with weak and strong disruptions of gene function returned by chemical mutagenesis. In addition to weak alleles, changes in protein coding sequence can also generate new functions (neomorphic alleles) and dominant-negative alleles that are also not possible via gene disruption mutagenesis. Physical mutagenesis with ionizing radiation, which damages DNA including via double-stranded breaks, was the first mutation induction mechanism used by humans on maize (Stadler 1928; Sachs 2009). Other popular methods for generating gene disruptions in maize include insertion mutagenesis with transposons such as Robertson's Mutator (Robertson 1986) and the Activator/Dissociator DNA transposons (McClintock 1953, 1956). More recently, the advent of directed DNA cleavage by bacterial CRISPR-associated proteins (Cas) and a guide RNA designed to be complementary to a specific genetic target have revolutionized the generation of alleles in user-specified locations of the genome (Doudna and Charpentier 2014; Lorenzo et al. 2024). While the nicks in DNA produced by Cas proteins favor deletion variants, single-nucleotide changes are also possible, and an allelic series can be created.

Spontaneous mutants, historically called sports, contribute to the discovery of novel alleles. In maize, at the time of this writing (December 11, 2024), there are 325 maize mutants present at the Maize Genetics Cooperation Stock Center operated by the U.S. Department of Agriculture-Agricultural Research Service in Champaign, Illinois, that were the result of spontaneous events that occurred during the normal growing of maize (https://www.maizegdb.org) (Woodhouse et al. 2025, in this collection). Some of these mutants are among the classical mutants of maize, which provided foundational understanding of the genetics of maize growth and development. In addition to mentioning the occasional recovery of dramatic phenotypes due to natural mutations in maize, we would be remiss not to give evolution and natural variation their due here. The study of natural populations and maize plant breeding is the study of spontaneous mutations that happened in the past. This variation, sometimes aptly referred to as standing variation, has also informed our understanding of plant growth, development, evolution, and improvement. Genetic studies often consider natural variation and induced variation separately, but these mutagenesis approaches are not exclusive. The combination of induced variation with natural diversity has been widely exploited in maize to discover natural alleles that modify mutant phenotypes. This approach identifies modifier alleles in the standing variation of maize and thus describes new loci in pathways of known function.

USE OF EMS IN MAIZE

To create germinally transmitted alleles, mutations introduced into DNA in a cell must be carried by the germline to the next generation. In maize, transmission can be accomplished by mutagenizing either pollen or the shoot apical meristem (SAM) cells in seeds. In angiosperms, the entire aboveground body of the sporophyte is derived from a small number of meristematic cells called the SAM. In many flowering plants, the pollen and egg cells arise on the same flower and share a cell lineage in the SAM. As a result, the meristematic cells in the seed gives rise to both the pollen and egg cells. This shared lineage makes it possible to mutagenize seeds and recover homozygous recessive mutants in the first progeny of the mutagenized plants. In maize seeds, the cell lineages that give rise to the egg and pollen mother cells on the ear and tassel structures, respectively, are already specified and distinct at seed maturity. As a result, seed mutagenesis induces independent mutations in these two lineages. A seed-mutagenized maize plant (the M1 generation for first mutant generation) results in the recovery of heterozygous mutants from the ear and tassel lineages. Thus, progeny (the second generation or M2) are effectively F1 progeny with two independently mutagenized lineages. Selfing of an M2 plant (second generation)—meaning that one has to perform the mutagenesis, grow the progeny, and then screen for recessive mutants in the third generation (M3)—can recover homozygous recessive alleles. If these crosses are all done in the field, recovering recessive mutants is a 3 year process. The requirement of a second generation to recover recessive alleles, the poor fitness and performance of the M1 mutagenized plants, and the added complexity of seed handling have discouraged many researchers from doing seed mutagenesis in maize. Instead, mutagenesis of mature pollen, which results in every progeny carrying independent mutations in the heterozygous condition, has been preferred (see Khangura et al. 2025a, in this collection). Screening for dominant mutants in the M1 and M2 progeny of seed-mutagenized maize has identified many dominant mutations (Neuffer 1994). In our experience, the mutagenesis rates of seed treatment are much higher than in pollen mutagenesis, which may have contributed to the recovery of so many dominant and semidominant gain-of-function alleles in maize.

In the first documented use of EMS for maize pollen mutagenesis, several methods were attempted to deliver EMS into pollen (Neuffer and Ficsor 1963). Methods that failed to recover any mutants attempted to introduce EMS into the plant body, in the hope that it would translocate to the germline. These included immersing cut leaves from young plants in dilute EMS solutions, injecting EMS solution into young tassel shoots, and applying EMS solution to the developing tassel through a cotton wick to a drilled hole in the stem. The first successful method involved 6 h of soaking the tassel with aqueous EMS solution 3–5 days before pollination. This approach required 15–30 mL of EMS solution per plant and had a mutation rate of approximately one novel allele in 12,000 M1 plants. Given the low mutation rate observed in this study and the sheer exposure risk to researchers in the field or greenhouse, a better method of treating pollen was sought. Unlike earlier attempts to use an aqueous EMS solution, Dr. Neuffer and Dr. Ed Coe later promoted the use of EMS in paraffin or mineral oil for creating a maize pollen slurry to prolong pollen viability, allowing additional time for EMS mutagenesis to occur (Neuffer and Coe 1978). This method has been used for decades and was published as a brief procedure in The Maize Handbook (Neuffer 1994) and as a protocol in Mutants of Maize (Neuffer et al. 1997), and has been further updated in Methods in Molecular Biology: Plant Embryogenesis (Settles 2020). The use of paraffin oil for pollen incubation resulted in a higher mutagenesis rate and, more importantly, reduced the exposure of researchers to EMS. The history of experiments that standardized pollen paraffin oil suspension in maize has been detailed previously (Neuffer et al. 2009).

EMS mutagenesis of seeds can be successfully used in maize but requires a greater understanding of maize development. The maize seed contains multiple cells that independently contribute to the vegetative plant. As a result, seed mutagenesis has a much larger mutational target than the small number of cells that contribute to the pollen and egg germlines. Seed mutagenesis results in mutant sectors that occur in repeatable patterns on the plant body as the result of developmental processes and the size of the plant body already present in the seed. For example, the size of the tissue sector derived from each cell in a mature seed is different depending on the location of the cell in the embryo's apical meristem. Earlier-emerging leaves have larger cell populations already specified in the embryo, and as a result, mutagenesis produces smaller sectors in earlier-emerging leaves. For example, approximately the first five leaves have already separated from the meristem in the maize seed and are referred to as the plumule. Only mutagenized cells that result in phenotypes that are visible in very narrow sectors of the leaf, such as cell-autonomous changes in color, can be visually screened for new mutations in these already initiated leaves. Within the seed meristem, cell populations fated to become the next leaves have already been specified. The earlier-emerging leaves from this cohort have larger cell populations committed to them than the later leaves, again resulting in larger mutant sectors in the later leaves. There are many cells in the seed that will give rise to small sectors and a more limited number of cells that will give rise to sectors that contribute to half of a leaf or more. As a result, the population of independent mutagenized chromosomes assessed in a screen is much larger when the phenotype is readily visible in smaller leaf sections (e.g., cell-autonomous and easy to see). In our experience, phenotypes that are easy to score in 1/16th of the leaf can be seen in many leaves and therefore have a larger mutational target than phenotypes that are only clear when visible in half or more of a leaf. Leaves that have already been specified, either in the plumule or as committed primordia in the SAM, neither result in sectors that are contiguous between leaves nor share cells that contribute to either the ear or the tassel. Although it is highly variable across genetic backgrounds, we estimate that there are more than 50 cell lineages visible for cell-autonomous foliar phenotypes. Even if we carry out a weaker seed mutagenesis that is only as strong as we can achieve in pollen, we can recover independent mutations using 1/50th of the field footprint in a single generation. When cells in the meristem that are not yet specified are mutagenized, the iterative nature of plant development results in a mutant sector that extends over multiple successive leaves. In our experience with maize, we observe sectors present in multiple successive leaves above the ear, consistent with mutations that arise in the cells of the SAM in the seed. The trade-off inherent in this approach is that very few of these mutations occur in lineages that enter the tassel or ear and, as such, cannot give rise to heritable variation. In the classic genetics era, the lack of heritable variation limited the utility of the seed mutagenesis approach in maize to studying cell-autonomous phenotypes and the linkage of mitotic clades and development. With the low cost of sequencing, it is now possible to molecularly identify genes of phenotypic consequence via sequencing. To capture heritable mutations from seed treatments with EMS, most phenotypes visible in the leaf subtending the ear and sectors in the leaves subtending the tassel should be focused on, as these leaves share cell lineages with the respective germlines of these inflorescences.

With the proper experimental design and suitable phenotypes visible in the sporophyte generation, seed treatment can be used to recover mitotically heritable lineages with novel mutations resulting in visible phenotypes. The facility of recovering tissue from sectors on chimeric leaves and extracting and sequencing DNA from affected and unaffected neighbor tissue (called the somatic sector approach) can allow researchers to generate, phenotype, and molecularly describe multiple mutant alleles following seed mutagenesis of maize and without the requirement of transmitting the mutant allele to the next generation. This strategy is a classical approach pioneered in ionizing radiation and transposon mutagenesis (Stadler 1928; McClintock 1956; Robertson 1986; Sawers et al. 2006; Karre et al. 2021). The somatic sector approach works well when the phenotype is visible in the growing plant and the mutagenesis dose applied does not interfere with plant growth and prevent the accurate scoring of the phenotype. For instance, we used a lower dose of mutagen so as to not cause generalized sterility when seeking new alleles of a male sterile gene in sorghum (Xiao et al. 2025). One use for the sector approach is knocking out a dominant mutant to recover multiple loss-of-function alleles (Karre et al. 2021), which can aid in molecular identification of the causative locus. Another is to knock out the wild-type allele in a recessive mutant heterozygote to identify multiple loss-of-function alleles and molecularly identify a recessive allele. For the sector approach to work in either case, the expression of the mutant phenotype must be possible in the presence of neighboring wild-type tissue and the mutagenesis must affect a sector large enough to be visible to the researcher. This approach will work well for some phenotypes. For example, mutations disrupting the production of chlorophyll cause yellow–green leaves and this phenotype is easy to screen for. For others, such as a phenotype in the aboveground portion of the plant affected by root morphology (e.g., wilting), the recovery of phenotypically affected aboveground tissue (wilted leaves) would not lead to isolation of the causative polymorphism, which would presumably be found in the root tissue. In theory, one could dig up the roots of wilted plants and sequence candidate genes from DNA extracted from root tissue. However, an experimental design to engage in a targeted EMS mutagenesis using EMS treatment of pollen would result in the causative polymorphism being present throughout the plant. This experiment would therefore be easier to execute and likely more successful for some mutants (see Khangura et al. 2025a, in this collection).

EXPERIMENTAL DESIGN MATTERS

The selection of the material for mutagenesis and the experimental design used will determine what kinds of variants will be exposed. Depending on the goal of the researcher, different experimental designs must be used. The designs of multiple mutagenesis experiments in maize are well described in Candela and Hake (2008), including pollen mutagenesis for carrying out mutant screens, population generation for reverse genetics, and suppressor mutagenesis for knocking out dominant mutations. These investigators also discuss the utility of carrying out mutagenesis to generate chimeric plants that exhibit mutant sectors for developmental and genetic analysis of gene function and identities. The advent of cheaper massively parallel sequencing methodology has led a number of researchers to integrate these approaches into mutagenesis designs. Updates to the mapping process integrating EMS treatment and shotgun sequencing into maize pedigrees are also available (Liu et al. 2012; Klein et al. 2018; Best and McSteen 2022; Lorenzo et al. 2024).

The maize community is currently underutilizing the ability of EMS, and other chemical mutagens, to return multiple alleles at a single locus. Much of the failure that we have witnessed in efforts to harness mutagenesis, and the frustration and fear that limit the use of EMS for maize discovery research, are owed to the necessity of setting up the experiment to achieve the goal, and failures to do so. Appropriate planning often begins in years prior to the experiment, with the production of sufficient seed and pedigreed material for mutagenesis.

Before the advent of gene editing, one approach, called “targeted mutagenesis,” used the experimental crossing design to return additional mutant alleles at a target locus at high frequencies. In this arrangement it is the genetic design, rather than directed mutagenesis, that creates the target. These experiments operate differently depending on the inheritance of the trait and the tissue target of the mutagen. For example, when the mutation is recessive and pollen is mutagenized, a tester plant with a recessive mutation is crossed by wild-type pollen treated with EMS. The progenies are then screened for the appearance of the recessive phenotype. All mutant progeny in the M1 result from pollen contamination from the ear parent or neighboring plant carrying the same mutant allele itself, from gynogenetic haploidy, or (the desired outcome) from the induction of a new allele in wild-type pollen that fails to complement the mutation in the tester plant. In this way, an M1 screen is achieved in which the recessive phenotype of the mutant of interest is “targeted” by the design of the cross. This approach is extremely effective at recovering additional independent mutant alleles to molecularly identify the causative locus of a mutant phenotype and for generating allelic series to explore allele interactions and protein function. Although easiest when the recessive mutant is fertile as a homozygote, we have used this approach to recover novel alleles in mutants by crossing heterozygotes with EMS-treated pollen (Sauer et al. 2023). Using a heterozygote as the tester doubles the size of the population needed to recover new alleles in the phenotypic screen, as half the progeny are derived from a wild-type egg. A similar screen is possible for recessive mutants following treatment of seeds with EMS, but the material treated needs to be heterozygous for the recessive phenotype. If the phenotype is cell-autonomous and in an accessible part of the plant, examination of M1 mutagenized individuals for somatic sectors displaying the mutant phenotype on an otherwise wild-type plant indicates a chimera generated by mutagenesis in a limited cell population. The somatic sector results in the expression of the phenotype when a mutation of the wild-type allele fails to complement the “targeted” mutant and is exposed in situ. Similar experiments can be used to knock out a dominant mutant allele that has a phenotype that can be visibly observed in the sporophyte. Heterozygous seed can be mutagenized, and the resulting plants can be directly observed for loss of the phenotype in a sector due to disruption of the dominant allele, leaving a single copy of the wild-type allele. Similarly, pollen mutagenesis to knock out a dominant mutant using pollen from a heterozygous source is well described by Candela and Hake (2008), and we provide an updated protocol for EMS mutagenesis of pollen that can be used for this purpose (Khangura et al. 2025a, in this collection). Finally, when dominant mutant plants are fertile as homozygotes, knockout alleles can be generated by pollen mutagenesis using a simpler design. In such cases, pollen from dominant mutant homozygotes can be treated with EMS and placed on wild-type ears. All wild-type plants recovered are either pollen contaminants, haploids, or knockouts of the dominant allele. A simple PCR test or low-throughput sequencing can be used to demonstrate the presence of the original allele and determine the molecular identity of the induced intragenic suppressor. Our associated protocols (Khangura et al. 2025a,b), and protocols by previous investigators (Neuffer and Coe 1978; Neuffer 1994; Neuffer et al. 2009; Settles 2020), include the steps necessary to generate novel mutations causing previously unknown phenotypes, the generation of new phenotypically expressing alleles at known genes, and the induction of derivative alleles to help molecularly identify the basis of both recessive and dominant genetic variants.

EMS mutagenesis in maize has been successfully adopted by a few laboratories. We believe that one of the limitations to its broader use by the maize community is due to fear of the harmful impact of EMS on researchers and a lack of experience and established protocols in the laboratory. This hesitation is particularly the case for the pollen mutagenesis methods, which are carried out in the field at anthesis (the time of pollen shed). For all EMS experiments, appropriate field design and preparations are necessary for safety. For pollen mutagenesis experiments, appropriate field layout at planting can minimize worker exposure to EMS during mutagenesis. By planting the field so that the tester plants are in a block or on an edge, researchers will not be tempted to walk through an area where EMS treatments were carried out. This layout improves both worker safety and confidence during EMS mutagenesis experiments. Pollen is treated in EMS as a slurry. Efficient movement of workers applying that slurry throughout the nursery can minimize the time that potential worker exposures can take place and critically reduce the time spent in personal protective equipment (PPE), which is hot and uncomfortable. Additional layers of PPE, recommended for worker safety from EMS, can increase the heat experienced by workers outdoors in summer nurseries. More time spent at high temperatures can pose its own risks directly, or indirectly through mistakes made in handling. By comparison, seed mutagenesis protocols can be carried out in the fume hood of a standard laboratory, and once the seeds are planted, the exposure to researchers is sequestered.

To improve the adoption and success of chemical mutagenesis experiments in maize, we provide two associated protocols that lay out, step-by-step, how to carry out EMS mutagenesis of maize by treating pollen or seed (Khangura et al. 2025a,b, respectively). Maize genetic backgrounds differ in their sensitivity to EMS by enough that a laboratory that is working with a specific line should run a pilot test for EMS sensitivity before engaging in a larger experiment. We describe doing so in the accompanying protocols. We also provide some guidelines about plant numbers, EMS doses, mutagenesis rates, and overall design considerations to help maximize success and encourage the use of these techniques in a broader set of circumstances. In addition, for geneticists interested in molecular identification of genes underlying various mutant phenotypes, certain experimental designs covered in these protocols can help achieve specific research objectives, and we make reference to these where possible.

CONCLUDING REMARKS

The use of chemical mutagens to generate genetic diversity in maize has unlocked the exploration of gene function, plant development, and biochemistry. Improved protocols and the capacity to directly sequence mutations in tissue samples collected from mutant plants make this technology particularly relevant for basic science discovery. The availability of pangenomic data (Hufford et al. 2021), including the variants present in multiple genetic backgrounds, makes questions in any maize variety conceivable. The opportunities now made possible by improved EMS protocols should usher in a new era of discovery in maize research.

COMPETING INTEREST STATEMENT

B.P.D. is an inventor on a patent (US-10595479-B2) for the recurrent mutagenesis of maize. R.S.K. and N.B.B. declare no competing interests.

AUTHOR CONTRIBUTIONS

Conceptualization: R.S.K., N.B.B., and B.P.D. Writing—original draft: R.S.K., N.B.B., and B.P.D. Writing—review and editing: R.S.K., N.B.B., and B.P.D.