This concerns the identification, using in vitro selection procedures, of desirable genotypes with respect to field performance traits such as growth, production of useful metabolites, and resistance to stresses caused by agents such as salt, heat, cold, drought, disease, insects, metals and herbicides. Selection in vitro can be at the level of cells (or protoplasts), microspores, buds, shoots, embryos, or whole plants. Selection at the cell level is attractive in terms of the large numbers that can be screened, but involves the particular problems of the generation of genetic variation, and sometimes the requirement to regenerate from the selected cells. Selection procedures reviewed here are considered independently of the source of variation, although a majority of reports concern putative genetic variation induced in vitro directly.
The various approaches to in vitro selection for disease resistance have been reviewed by van den Bulk (1991). Cultures can be exposed to a toxin, toxin analogues, filtrate or the pathogen itself. The use of purified toxins as selection agents in culture is potentially effective when symptoms of the disease are caused by a toxin produced by the pathogen, and where the toxin operates at the level of the explant cultured (e.g. at the cell level). Selection with crude filtrates has been used in many studies, in particular when a filtrate exhibits phytotoxic activity but no well characterized toxin is known. Selection by co-culture with the pathogen has been of limited success, in particular due to difficulties in growing or controlling the growth of the pathogen in culture.
Corn plants regenerated from callus lines selected for resistance to the purified toxin from Helminthosporium maydis displayed resistance to the pathogen (Widholm 1988). Resistant plants were also obtained using selection on the basis of the insecticide methomyl, apparently acting as a toxin analogue (Kuehnle 1990, Kuehnle & Earle 1992). Elm microcuttings displayed wilting in response to both filtrate and coculture with Ophiostoma ulmi (Dorion & Bigot 1987, Dorion et al. 1988). In another experiment, rate of fungal growth in coculture with callus from mature trees was correlated with known field resistance of the trees to Dutch elm disease (Dormir 1992). Resistance of cultured loblolly pine embryos to infection by Cronartium quercuum spores showed high family correlations with field resistance (Frampton et al. 1983). The ranking of poplar clones for field resistance to Septoria musiva was similar to that derived by inoculating cultured leaf disks with spores (Ostry et al. 1988). A high correlation between callus resistance to filtrate and resistance of regenerated plants to Fusarium oxysporum was reported for alfalfa (Arcioni et al. 1987, McCoy 1988). In vitro resistance of potato plantlets cocultured with Phytophthora infestans was well correlated with known field resistance of cultivars (Tegera & Meulemans 1985). Protoplasts of two grape varieties reputedly resistant to Eutypa lata showed good viability in the presence of the filtrate, while those of three reputedly sensitive varieties died rapidly (Mauro et al. 1986). Rankings determined with in vitro inoculation of leaf disks reflected well field resistance of poplar clones to rust (Singh & Heather 1982).
Tomato plants regenerated from callus tissues resistant to toxins secreted into media displayed resistance to Pseudomonas solanacearum (Toyoda et al. 1989). Resistance of callus tissues to filtrates from P. syringae has been used as an assay for bean cultivars resistant to halo blight (Hartmann et al. 1986). Insensitivity of protoplasts to methionine sulfoxymine, a toxin analogue, has been used to regenerate tobacco plants resistant to wildfire disease caused by P. syringae. Sensitivity of poplar microcuttings cocultured with Xanthomonas populi was well correlated with clonal susceptibility to bacterial canker in field tests (Janssen 1989). Toxin resistant cells selected in callus cultures of a peach clone highly susceptible to bacterial leaf spot disease yielded plants some of which displayed much higher field resistance (Hammerschlag 1986). Direct inoculation of pear plantlets with Erwinia amylovora provided a resistance assay well correlated with field resistance to fire blight (Viseur et al. 1987). Selection of cell lines resistant to culture filtrate yielded regenerated plants with increased resistance to Phoma lingam (Sacristan 1985).
As indicated, many of the above involve selection among populations of somaclonal variant cells. There are many reports also of cell lines selected by exposure to toxin yielding regenerants which do not display increased resistance to the disease. Although possible instability of the variants is a confounding factor in many of these studies, some undoubtedly result from the fact that some diseases do not act at the cell level.
In principle, this method is straightforward. Herbicide is simply included in the medium at an appropriate concentration. Selection operates on the physiological basis that many herbicides also directly kill cells and tissues. Some herbicides, however, act by interfering with the functions of organised tissues, e.g. photosynthesis, and may not affect the growth of heterotrophic cell cultures (Chaleff 1986, Widholm 1988). Similarly, anatomical features such as the cuticle may influence in vivo but not in vitro absorption of herbicide (Smeda & Weller 1991).
Most available published reports concern the selection of somaclonal variant cell lines showing resistance to herbicide. An exception is that of Yenne et al. (1987), where responses of existing commercial cultivars were compared. In another interesting exception, rape microspores were subjected to mutagenic agents, plated, and then early stage embryos subjected to range of herbicides. Survivors completed the regeneration and colchicine doubling phase and plants were regenerated for testing (Beversdorf & Kott 1987)
Resistance to most major herbicide classes has been selected for in vitro, and at least partially resistant plants regenerated in many cases. Atrazine-resistant plants were regenerated from cells selected in green cultures of Nicotiana plumbaginifolia (Cseplo et al. 1985). In another study, cotyledonary node plus epicotyl explants of soybean were cultured on atrazine-containing medium. Some explants yielded organogenic shoots from which plants were regenerated. Some of these plants displayed resistance to atrazine (Wrather & Freytag 1991). Calli of Nicotiana debneyi displaying very high levels of tolerance to amitrole were selected by stepwise exposure to increasing concentrations. Regenerated plants displayed tolerance as calli (Swartzberg et al. 1985). Amitrole tolerance at the whole plant level was reported in tobacco plants regenerated from cell lines selected in vitro (Chaleff 1986). A similar result was achieved with maize (Anderson et al. 1987). Plants regenerated from resistant cell lines of tobacco displayed tolerance to the sulfonylurea herbicides chlorsulfuron and sulfometuron methyl (Chaleff 1986, 1986b). Plants regenerated from hybrid poplar leaf explants subjected to selection on media containing sulfometuron methyl were tolerant of herbicide levels lethal to control plants (Michler & Haissig 1988, Michler 1988). Regeneration of shoots from poplar leaf explants exposed to glyphosate gave rise to glyphosate tolerant plants (Michler & Haissig 1988). In pea, in vitro sensitivity of some commercial cultivars showed some correlation with field sensitivity to glyphosate (Yenne et al. 1987). Tobacco plants tolerant to picloram were regenerated from resistant cell lines (Chaleff 1986). For tomato, some tolerance to paraquat was recorded in plants regenerated from cells selected on media containing this herbicide (Chaleff 1986).
Instability of the tolerance is a feature of several of these reports, and there are several examples of poor correlations. In birdsfoot trefoil (Lotus), for example, plants regenerated from cells surviving in chlorsulfur on media actually were more sensitive than the controls (MacLean & Grant 1987). There are many examples also of resistant cell lines whose capacity to give rise to resistant plants has yet to be tested.
This is also a simple procedure — plants are regenerated from explants displaying tolerance to salt added to tissue culture media. Salt resistance in plants includes both avoidance and tolerance mechanisms. Resistance in Atriplex for example depends on the anatomical and physiological integrity of the whole plant and not on cellular properties. Nevertheless, cellular tolerance mechanisms exist which are more amenable to in vitro selection techniques. Elevated levels of the amino acid proline are believed to protect plant tissues against stress by acting as N-storage compound, osmosolute and hydrophobic protectant for enzymes and cellular structures, and proline effected salt tolerance has been reported in several crops (Jain et al. 1991).
Cell lines tolerant of elevated levels of salt in the medium have been selected in many studies. In some of these studies, plants regenerated from the cell lines have also displayed increased tolerance in greenhouse and field trials, e.g.: alfalfa (Winicov 1990, 1991), Coleus blumei (Ibrahim et al. 1992), Brassica juncea (Jain et al. 1991) and Citrus sinensis (Spiegel-Roy & Thorpe 1986). Plants regenerated did not display such tolerance in tobacco (Watad et al. 1991). All of the these reports concern variation induced in culture. Stability of tolerance through subsequent sexual generations has been evident in some studies (Ibrahim et al. 1992, Jain et al. 1991), but not in others (Lucas et al. 1989).
Once again, this is simple in principle - explants are cultured on media containing high levels of metal salts. In practice, however, complex interactions among various nutrients, and pH effects, must be taken into account. As for salt tolerance, plants can resist high levels of metals such as aluminium and cadmium by various avoidance (e.g. decreased uptake) or tolerance (e.g. altered enzyme structures, precipitation in cytoplasm) mechanisms (Tal 1983). These operate at different levels, and therefore good correlations between in vitro and whole plant responses are not necessarily expected.
Cell lines resistant to elevated concentrations of metal ions have been selected in many studies, including those involving aluminium, cadmium, mercury, zinc, lead, and manganese. Increased tolerance to aluminium in subsequently regenerated plants was demonstrated for Nicotiana plumbaginifolia (Meredith et al. 1988) and rice (Van Sint Jan & Bouharmont 1992). Most studies involve selection among variants induced by the culture process. Meredith et al. (1988), however, cited some studies where responses of callus cultured from cultivars known to vary in sensitivity to metals were examined. Differences among Agrostis genotypes in resistance to zinc were maintained in callus cultures (Wu & Antonovics 1978), while cultivar differences in sensitivity to zinc (Christianson 1979), and manganese (Petolino and Collins 1985), were not correlated with callus responses for bean and tobacco respectively.
Cultures are exposed to high temperatures and survivors selected. Tolerance is presumably based on reduced incidence of or resistance to the effects of protein denaturation, reported to be the primary cause of heat injury (Quamme & Stushnoff 1983).
Plants regenerated from cotton cell cultures exposed to regular high temperature treatments themselves yielded callus with an elevated tolerance to high temperatures (Trolinder & Shang 1991), although whole plant responses were apparently not examined in this study.
For this purpose, explants are exposed to low temperatures and survivors selected. Methods include those involving exposure to the frost temperatures expected, through to immersion in liquid nitrogen. Plants are unable to avoid the low temperatures, and only tolerance mechanisms enable the plant to survive adverse effects. Resistance operates mainly at the cellular level, e.g. accumulation of antifreeze substances, dehydration to minimize freezable water and increased capacity for supercooling, and good correlations between cellular and whole plant responses are therefore expected (Tal 1983). The involvement of proline has been reported (Teulieres et al. 1989, Tantau & Dorffling 1991). Chilling tolerance is apparently an additive, multigenic character, and different mechanisms control resistance at different developmental stages (Paull et al. 1979).
For eucalypt clones representing a range of taxa, viability of protoplasts after cooling to -10°C was well correlated with field resistance determined by assessing sprout regrowth after frost injury (Teulieres et al. 1989). Nowak et al. (1992) concluded that testing the ability of cultured whole plantlets to withstand freezing could be used to rank alfalfa germplasm for cold tolerance. Some callus lines surviving immersion in liquid nitrogen gave rise to plants significantly more tolerant of low temperatures than unselected controls (Kendall 1991).
Methods employed have generally involved the attempted selection of cell lines showing tolerance to osmotic agents such as polyethylene glycol. Drought resistance in plants can operate through avoidance and tolerance mechanisms. Avoidance involves the maintenance of high internal water potential in the presence of external water stress, while drought tolerators are able to endure dehydration of the protoplasm (Quamme & Stushnoff 1983).
Cell lines resistant to stress caused by polyethylene glycol have been identified for Douglas fir (Leustek & Kirby 1990) and some crop species (Tal 1983). Stability of tolerance to water stress induced by mannitol after regeneration of plants and subsequent culture of callus explants has been demonstrated for a Prunus hybrid (Ochatt & Power 1988), but correlations with whole plant responses to drought stress are not available.
Several reports exist of plants or cell lines selected for tolerance to a particular stress displaying elevated tolerance also to other stresses. Plants and cell lines selected for resistance to one herbicide are often cross-resistant to other, sometimes unrelated, herbicides (Hughes 1983). Cell lines selected for high temperature tolerance were tolerant also of increased water stress in cotton (Trolinder & Shang 1991), and high salt levels in tobacco (Harrington 1989). Tobacco cell cultures selected for cadmium tolerance displayed tolerance to heat shock and exposure to cold (Huang & Goldsbrough 1988). Selection for salt tolerance in callus of sorghum yielded some cell lines resistant to insects (Isenhour 1988). Although an underlying genetic basis has yet to be demonstrated in these particular studies, such examples of cross-tolerance suggest the existence of common factors in stress tolerance. A protective effect of proline against a range of stressful conditions has been demonstrated (Tantau & Dorffling 1991).
Resistance to amino acid analogues has been used to identify genotypes producing increased levels of certain amino acids in which plants are often deficient, and stability in regenerated plants and their progeny has been demonstrated in some cases (Widholm 1988). Tobacco cell lines synthesizing high levels of nicotine have been selected on the basis of resistance to nicotinic acid (Robins et al. 1987). Variation in the production of the antihypertensic alkaloids serpentine and ajmalicine by both plants and callus cultures was examined among 20 genotypes of Catharanthus roseus. Poor correlations suggest that accumulation in callus cultures will not be a useful method for selecting plants producing increased levels (Roller 1978).
The gene determining parthenocarpy in tomatoes is recessive and difficult to detect under field conditions, but can be distinguished readily on the basis of ovary size when cultured on media with gibberellic acid (Young 1990). Higher heritability under in vitro rather than field conditions is also the basis for in vitro screening for the presence of bracts and pink discolourations (both undesirable traits) in cauliflower (Kalia 1986). Ethylene-resistant cultivars of Begonia display greater “keepability” under house conditions than ethylene sensitive cultivars, and an in vitro selection system has been developed involving exposure of plantlets to ethylene and retention of those retaining green foliage (Hvoslef-Eide 1991). Although cell lines vary for growth rate in culture, this parameter is poorly correlated with plant growth in the field (Widholm 1988), preventing selection for yield on this basis.
In summary, many recent publications have reported useful correlations between in vitro responses and the expression of desirable field traits, most commonly for disease resistance although some positive results are available also for tolerance to herbicides, metals, salt and low temperatures. These reports concern selection among variants (at both the sporophyte and gametophyte level), originating from sexual recombination, and also those purportedly induced in culture. Least ambiguous are reports of assays involving genetic variation generated prior to culture, and these are most numerous for disease resistance. Successful in vitro assay procedures have been reported for many plants and diseases (both fungal and bacterial), and the potential clearly exists for developing protocols for others. The majority of reports of in vitro selection concern selection among populations of somaclonal variants, and these should be treated with a little more caution. There is little evidence to suggest that in vitro selection for growth traits, or tolerance to high temperatures or moisture stress, would be possible.
Like any other method of indirect selection, the value of in vitro selection is dependent on:
The extent to which in vitro and field traits are correlated. This also will incorporate juvenile- mature correlations where juvenile material is used for the in vitro test. It should be noted that many important characters operate at above the cellular level, e.g. resistance due to phenological features, and these won't be expressed in cultured cells. Other traits may be cellular, but not usually expressed in tissue cultures.
Heritabilities of both in vitro and field traits.
Ease (and thus cost) of measurement of both.
As discussed above, correlations between in vitro and field responses have been reported for many diseases of plants. Traditional methods of screening for disease resistance involve field and greenhouse trials. In many of these trials, the level of exposure to the pathogen is difficult to control, and heritability is low. In vitro methods offer advantages, and clearly there is some potential for broader application in crop breeding. For forest tree species, on the other hand, disease resistance is not a selection criterion of widespread importance. The application of in vitro selection for disease resistance is therefore limited, although there may be some applications beyond loblolly pine and the poplars.
Cold tolerance is an important selection criterion in some forest tree species, and exposure in the field is difficult to control. Other cold tolerance assays exist, however, which are fast and efficient and which do not involve in vitro selection. The leaf disk conductivity method, for example, has been shown to correlate well with whole plant performance in Eucalyptus (Tibbits et al. 1991).
Tolerance to salt and metals is important in some land rehabilitation programmes, and herbicide resistance may be of use in a few programmes. In general though, greenhouse trials are reasonably simple and effective for the assessment of these traits. Decontamination of material from seedling populations for in vitro screening, on the other hand, is likely to be laborious. Some application is likely though where selection by greenhouse testing of candidates or their progeny is difficult by virtue, for example, of difficulties in exposing plants evenly to the agents, or in obtaining progeny of field candidates.
The expense involved in decontaminating large seedling progenies for in vitro testing could be avoided if genetic variation could be reliably induced in culture - millions of variants in a cell culture could be rapidly screened. Applications of in vitro selection in plant breeding would be broader if the induction of somaclonal variation were demonstrated to be a reliable method for creating genetic variation, and if selection could be coupled with a somatic embryogenesis system.
Results of very limited studies conducted offer encouragement that further investigations of microspore selection for tolerance to low temperatures and other stresses, and the development of technologies for using the selections in breeding programmes, would be useful. As pointed out by van den Bulk (1991), a large portion of the genome is transcribed and translated during pollen development. A particularly interesting example has been the selection of Eucalyptus gunnii pollen surviving exposure to cold. This pollen was capable of germination and could be used in control pollinations (Boudet & Marien 1988).
For the selection criteria of major general importance in foresty, in particular vigour, stem form and wood quality, however, poor correlations with field responses will limit the usefulness of in vitro selection. In vitro selection is therefore likely to have very limited application in forest trees species - of possible interest in a few programmes where there is a disease resistance selection problem, but of no broad strategic value as a research objective.