When is Mutation Breeding Appropriate?

Mutation breeding is the strategy of choice when it is essential to retain the character- istics of an existing cultivar, but correct a specifi c defect in it. The high heterozygos- ity of most citrus cultivars makes it very dif- fi cult to recover progeny with very similar characters from hybridization-selection. The relatively small population sizes that most citrus breeders can produce and eval- uate further limit hybridization-selection in many crops. These conditions occur in all important fruit cultivar types except man- darins, so it would seem that mutation breeding should be widely used. This has not been the case because, for most traits, mutation breeding is also limited by the breeder’s ability to produce and evaluate very large populations. Mutation breeding procedures (other than genetic engineering) do not allow the breeder to target particular

genes or traits. Tissue exposed to mutagenic agents will have random mutations, most of which have no effect or undesirable effects. Therefore, it is necessary to screen large populations to have a reasonable chance of fi nding the desired mutation. Increasing the dosage of mutagen will increase the fre- quency of mutations, but also increases mortality. At high dosages, the desired mutation will often be accompanied by undesirable mutations which mean that the mutant line cannot be used directly as a cultivar. The one trait that clearly can be obtained relatively easily with mutation breeding is seedlessness, presumably because the mechanism does not require mutation of a single gene or genes. This is discussed in detail below. Alterations in other traits can also be obtained by muta- tion breeding, but success is less certain and the effort required is likely to be greater.

The place of mutation breeding in citrus cultivar development will depend greatly on the cost and acceptance of genetic engineering. Genetic engineering promises to achieve the same type of end result as mutation breeding: introduction of a specifi c trait into an existing cultivar. However, genetic engineering is more pre- cise because it depends not on a chance mutation, but on addition or change in a specifi c target gene. Induced mutations are not generally different from the natural mutations responsible for variation in

nature. Therefore, mutation breeding does not face the same level of controversy as the ‘unnatural’ changes induced by genetic engineering (Predieri, 2001). Furthermore, approval for release of cultivars produced by genetic engineering is presently very expensive in most countries, while culti- vars from mutation breeding are not subject to special regulation.

The stages of a mutation breeding proj- ect involve fi rst choosing a target tissue to expose to the mutagenic agent, choosing a mutagen, exposing the tissue, propagating plants from the exposed tissue, screening these plants for the desire mutation and then evaluating selections that carry the mutation, either directly or as parents. Each of these stages is discussed below.

Tissue choice

The breeder can initiate a mutation breed- ing project with one of several different types of tissues, each with certain advan- tages and disadvantages (Table 16.1). Choice of tissue may also depend on the target trait and type of mutagen to be used. The major advantage of treating hap- loid tissues, generally pollen or microspores, it that chimeras are not pro- duced. The treated pollen is then used to fertilize untreated egg cells, seeds are col- lected, and seedlings are grown and evalu- ated. The major disadvantages of this approach are, for fruit traits, the time to evaluation is long because the seedlings will be juvenile, and it may be diffi cult to distinguish mutants from recombinants in

the sexual progeny that result. Furthermore, this approach results in hybrids few of which will resemble the parent cultivar. Pollen irradiation has rarely been used in citrus.

Treatment of seeds is widely practised in seed-propagated crops, but is less common in citrus and other fruit trees. It is easy to treat large numbers of seeds, but identifi cation of mutants is easy only if most embryos are of nucellar origin. It is not clear whether the frequency of chimeras is similar after treatment of seeds and buds. In both cases, the target cells occur in meris- tems composed of many cells, and chimerism seems fairly likely, but there are few data on rates of chimerism on which to base a conclusion.

Treatment of tissue cultures with muta- gens has been used by several researchers. The fi rst step involves establishing a cul- ture capable of regeneration. This can be diffi cult in some cultivars and is at least tedious if it is desired to induce the muta- tion in many cultivars. The type of tissue culture influences the likelihood of chimerism and time to evaluation. If proto- plasts from an embryogenic callus system are used, then most embryos are thought to regenerate from single cells and therefore chimerism is unlikely, but the regenerated plants will be juvenile. Chimeras are also minimized if plants are regenerated from protoplasts produced before or after muta- gen treatment. If stem segments from mature tissue are cultured, then organized meristems are already present and chimeras are as likely as with seeds or budwood. An advantage of tissue culture systems is that

Table 16.1. Advantages and disadvantages of treating various tissue types with mutagens.

selection for certain traits can be applied to the culture, increasing the chances of recov- ery of a desired mutation.

Treatment of buds has been widely used in citrus (e.g. Froneman et al., 1996). A major advantage is that trees propagated from such buds will not display the delayed flowering and thorninesss typical of seedlings. However, the probability of chimerism is high relative to treatment of single-cell systems. Lapins (1983) discusses the organization of primary and axillary buds in detail. A dormant leaf bud contains the primary leaf bud and several axillary bud primordia. These buds are typically induced to grow into shoots that may con- tain mutant cells. Axillary buds that are larger and more fully differentiated at the time of mutagen treatment are more likely to produce shoots, but these are more likely to be chimeric. The smaller axillary buds have fewer cells and therefore less proba- bility of giving rise to a chimera.

Mutagenic agents

The breeder can choose from a wide range of mutagenic agents. Each agent typically induces mutations by a particular mecha- nism and therefore results in a specifi c type or types of DNA alteration. Specifi c agents may be appropriate for a particular tissue or target trait. Mutagens can be broadly classi- fi ed as radiation, chemical, transposable element and pathogen.

Radiation is the most widely used type of mutagen for citrus. In part, this is because there has been considerable fund- ing for mutation breeding from national and international atomic energy agencies anx- ious to characterize the effects of radiation and fi nd useful applications. Although UV irradiation can be used, it generally pene- trates tissues less deeply than ionizing radi- ation. Ionizing radiation includes X-rays and g-rays, and has been the most widely used and effective type of mutagen for citrus breeding. It can induce a wide range of mutation types including chromosome breaks and rearrangements, and point

mutations. It is thought that seedless forms of citrus can be obtained at relatively high frequencies by irradiation because the resulting chromosome breaks frequently lead to rearrangements (inversions and translocations) that cause sterility. This supposition has not been verifi ed experi- mentally, but many researchers have reported a high frequency (up to 10%) of seedless selections from irradiation (Hearn, 1984, 1986; Vardi et al., 1996)

When using radiation, as well as other mutagens, the breeder must determine the appropriate dose and exposure to apply. The two most commonly used radiation types are X-rays and g-rays (generally from a 60Co source). Dose is the amount of muta- gen applied per unit time, while exposure is the total absorbed radiation. The current unit of absorbed radiation energy is the Gy (Gray), while most of the older literature reports exposure in rads or Roentgens (R, which measures exposure rather than dose). One Gy equals 100 rad. This unit is useful in calculating how long and at what distance to expose tissue to the radiation source. However, it is diffi cult to predict the effect of a given dose and exposure on a particular tissue. Relatively little is known about the effects of different dose rates for citrus tissues because most breeders have access to a single radiation source which emits at a constant rate. Breeders typically use the LD50 as a measure of the biological

response of tissue to a particular radiation

dose. While LD stands for lethal dose, and therefore the term implies survival, it is also possible to apply it to vegetative growth, i.e. a 50% reduction in vegetative growth. Sensitivity to ionizing radiation typically varies with the physiological con- dition of the tissue, particularly water con- tent, and the cultivar, so breeders generally identify an appropriate dose empirically. In the breeding programme at the University of California, Riverside, we use a dose rate of 4–8 Gy/min (50 Gy/min would be ideal if available) and exposure of 30–50 Gy with budwood to induce mutations to seedless- ness. Hearn (1986) also reported good results from irradiating grapefruit budwood

with 30–50 Gy from a 137Cs source. Vardi et al. (1996) found about 10–13% of low- seeded mutants in two mandarin cultivars after exposures of 27.5–65 Gy, but another cultivar had only about 8.5% after expo- sures above 40 Gy.

Chemical mutagens have been used less commonly in citrus than radiation. Different chemical mutagens induce differ- ent spectra of mutations. A very wide range of chemical mutagens are now available, but few of these have been evaluated in citrus. If vegetative tissues such as buds or meristems are to be treated, it is generally considered better to dissect away the outer layer of leaves that cover the bud or meris- tem to improve penetration of the chemical. Tissue cultures can be treated with muta- gens rather conveniently, as summarized by Predieri (2001), but their use for citrus mutation breeding has not been reported.

Transposable elements (TEs) are natu- ral genetic elements with the ability to move from one genome location to another, either with or without duplication of the element. A variety of different classes have been identifi ed (Feschotte et al., 2002). TEs can be used to induce mutations, a strategy known as transposon tagging (Maes et al., 1999) because a gene mutated by a transpo- son insertion can be cloned using the trans- posable element as a ‘tag sequence’. No TEs in citrus have been shown to be active, but there is indirect evidence of such activity since some clementine cultivars derived by mutation have differences in DNA markers detected using primers from retrotranspo- son elements (Breto et al., 2001). TEs can be transferred from one plant species to another, but do not always work in unre- lated species. If a TE inserts into a gene or regulatory region, it can induce a mutation that alters expression of that gene. At pres- ent, use of TEs as an experimental mutagen- esis system for citrus has not been reported.

Targeted mutation, in which mutations are induced in a specifi c gene or DNA sequence, is not yet effi cient in plants. If this technique were successfully devel- oped, it would certainly reinvigorate muta- tion breeding research because it would

open up an alternative path to apply genomics tools to citrus, one which is likely to be less controversial than transformation.

Propagation of mutagen-treated materials

When the plant material has been treated with the mutagen, the next step is to propa- gate trees from the potentially mutant cells. The precise method depends on the tissue that was treated, since this determines appropriate propagation mechanisms and infl uences the probability of chimeras. If trees or buds are treated with radiation, then DuPlooy et al. (1993) suggest that the mv1 generation should be budded on to rootstocks and the mv2 generation buds from these trees should be budded on to a new set of rootstock and the resulting trees planted in the fi eld for evaluation. This is expected to reduce the frequency of chimeras, but detailed results were not reported. In the UCR programme, to induce mutations to low seed content, we irradiate budwood, propagate trees from this and evaluate these mv1 trees. About 50% of irradiated buds result in trees that are planted in the fi eld for evaluation. Most buds that grow are the primary bud, but sec- ondary buds also push in some cases. Of these, perhaps 1–5% are low seeded, the percentage varying greatly with cultivar. We have rarely observed chimeras in this material in that all fruit from a tree is as seedy as that of the original parent, or all fruit have greatly reduced seed number. However, this does not test for periclinal chimeras. Selections are then re-propagated using standard budding methods to pro- duce about 60–100 trees for trials in several locations. It is expected that chimeras may be revealed at this stage. A rapid method to test for chimerism would be a valuable addition.

Screening

Screening refers to initial evaluation of the mutant population. The method chosen

will depend on the target trait. If trees are not expected to be juvenile then they can be screened fairly quickly and a high planting density can be used with little diffi culty. If trees result from mutagen treatment of seed, pollen or tissue cultures so that juvenility is expected, then they are unlikely to fruit until 5–7 years old. Training trees to a long central leader (Vardi and Spiegel-Roy, 1988) and certain cultural practices can speed fruiting, but may require more labour than is available to some programmes. Trees expressing juvenility must either be planted at wider spacing or pruned more severely to maintain plantings. In any case, when the trait is expressed in a reasonable percentage of the population, the breeder can begin screening. Each tree and sector of the tree should be evaluated for the trait, and for other visible mutations. If sectoral chimeras occur, these may be detectable as branches with different leaf colour, intern- ode distance or other visible differences that correlate with fruit character effects. Evaluation of fruit for 2–3 years is advisable to reduce the number of selections that do not consistently express the mutation. An example of a mutation breeding programme is shown in Fig. 16.2.

Evaluation

Field evaluation of selections from muta- tion breeding is similar to that for other cul- tivars, but a few specifi c considerations are relevant. Cultivars from mutation breeding are less likely to be genetically stable and to be hybrids, and therefore particular atten- tion should be paid to evaluating unifor- mity within and between trees. Trees should be subject to normal cultural prac- tices that might infl uence stability. Pruning in particular may lead to layer substitutions which cause reversion of chimeras, so branches that regrow after pruning should be carefully studied. On the other hand, if the cultivar treated with mutagen has already been characterized for performance attributes, it should be possible to evaluate the selections for somewhat fewer years

than would be required for a hybrid cultivar that may have many novel attributes and/or defects.