How do rootstocks control shoot water relations?


Many woody crop plants such as grapevines and fruit trees are traditionally grown with scion varieties grafted onto rootstocks – the selection of an appropriate rootstock provides a powerful tool to manage the growth and fruiting of the scion. In the case of grapevine, the rootstock can also provide resistance to the root aphid known as Phylloxera, while in apples, the Malling and Malling–Merton series of rootstocks (M9, M27, MM109, etc.) have provided a critical crop management tool for controlling tree growth and for maximizing fruit production or quality with the latter rootstocks also providing resistance to woolly aphid (Hatton, 1935; Preston, 1966). In this issue of New Phytologist, Marguerit et al. (pp. 416–429) provide some important evidence on the genetic basis underlying the role of the rootstock in the control of some aspects of water relations in grapevines.

‘Unfortunately any direct effects of the rootstock on water relations tend to be confounded with the effects on vigour.’

The study by Marguerit et al. sought to investigate the localization of rootstock genes that control scion water relations with the hope that this could provide a basis for the selection of improved rootstocks because it is widely recognized that the rootstock can provide a powerful tool for manipulating the drought tolerance of grapevine scions. They used an F1 mapping population from an interspecific cross between two rootstocks and used this population to identify a number of quantitative trait loci (QTL) for traits related to the plant’s water relations. The traits studied by Marguerit et al. included: transpiration per unit leaf area, transpiration efficiency (as derived from biomass gain divided by water loss), intrinsic water use efficiency (WUE, as derived from carbon-13 (13C) discrimination) and total transpirable soil water. A particularly interesting feature of this study, however, is that they also looked at the acclimation of transpiration rate to drought and found evidence for genetic differences in the shape of the response curve relating transpiration to soil moisture deficit. Interestingly this result for grapevine contrasts with that from an earlier field study of apple rootstocks that did not detect any difference between rootstocks in the relationship between stomatal conductance and leaf water potential (Higgs & Jones, 1990), so it is worth considering the possible mechanisms underlying rootstock effects in more detail.

What do we know about the mechanism of rootstock effects?

The most obvious effects of rootstocks on scion behaviour include those on scion vigour and on scion water relations. Indeed, since the work on apple at East Malling Research Station (UK) by Beakbane and colleagues (e.g. Beakbane & Thompson, 1939; Beakbane, 1956), it has been thought that effects on water relations may underlie size control by rootstocks, though others have suggested a key role for endogenous growth regulators in altered signalling between the rootstock and the scion (see Fig. 1; Jones, 1986).

Figure 1.

An illustration of some of the main ways in which it is thought that rootstocks can affect shoot water relations and growth. These include direct chemical signalling through modification of xylem (or phloem) sap composition and hydraulic signalling resulting from alterations to xylem conductivity. Note that the hydraulic signalling loop also involves the feedback effects of altered canopy transpiration on soil water.

Beakbane showed that apples grafted onto dwarfing rootstocks tended to have smaller xylem vessels than those on invigorating rootstocks and that this might lead to an impaired hydraulic conductance with the resulting development of enhanced water deficits and hence reduced growth in trees grafted onto such rootstocks. This conclusion has been supported by much subsequent work showing that in both apple and peach rootstocks the dwarfing effect appears to be related to lowered water potentials during the afternoon, with these differences being associated with xylem vessel size (e.g. Atkinson et al., 2003; Basile et al., 2003; Tombesi et al., 2010). Interestingly Beakbane & Mujamder (1975) also found evidence that stomata tended to be smaller in dwarfing than in invigorating rootstocks, but whether this was a separate genetic effect or one that resulted from the differences in hydraulic conductivity and consequent water stress was not resolved. In many cases the reduction in conductance seems to be associated primarily with differences in the rootstock material and not with differences in the graft union or in the scion, at least for peach (Basile et al., 2003), though processes at the graft union can also be important, and the potential role of endogenous growth regulators is still unresolved (Jones, 1986; Soar et al., 2006).

Role of rootstocks in response to drought

As in apple and peach, the rootstock in grapevine is critical in determining the response of the scion to drought (Carbonneau, 1985; Soar et al., 2006). The interest in rootstock effects on water relations arises primarily because of the potential that the breeding or selection of improved rootstocks offers for improving water use efficiency and drought tolerance of fruit crops, including grapevines, while retaining the key quality features of the scion. Unfortunately any direct effects of the rootstock on water relations tend to be confounded with the effects on vigour. For example, at least in apple, there remains some question as to whether dwarfing rootstocks are better or worse than invigorating rootstocks under drought conditions. Because the more vigorous plants tend to use more water and hence dry the soil faster one might expect that this could be disadvantageous under drought conditions (see Landsberg & Jones, 1981; Higgs & Jones, 1990). However, there have been suggestions that more invigorating rootstocks can in fact confer drought tolerance on the whole plant (Preston et al., 1981), possibly because their greater vigour allows them to extract a greater volume of water from the soil (though it would be interesting to know whether this phenotype also occurs in pot experiments).

The balance between the direct rootstock effects on shoot growth and those effects resulting from the feedback effects on soil water content can be rather difficult to separate. The demonstration by Marguerit et al. that there are separate loci controlling effects on transpiration and differences in the sensitivity of transpiration (and hence presumably stomata) to soil moisture, however, makes a useful contribution to untangling the complexities of rootstock effects on scion water relations.

Next steps

A next logical step for this work could be to use refined QTLs in marker assisted selection of improved rootstocks, or even ultimately as an aid to identification of the genes that underlie the detected QTLs. Unfortunately the identification of these genes is extremely speculative at this stage because of the low mapping resolution of such analyses, which depend both on the number and spacing of microsatellite markers available and on the number of distinct genotypes in the population used. Marguerit et al. report that there were between 300 and 850 genes within the intervals covered by each of the QTLs identified in their study; therefore identification of the actual genes involved will need to await higher resolution mapping. Indeed it is probably not very helpful at this stage to speculate on the actual genes involved, not only because the resolution of the mapping is far too low, but also because of our uncertainty as to the actual physiological or biochemical mechanisms underlying the observed phenotypic differences. Even a simple effect on transpiration could result from genes regulating any one or more of a wide range of direct and indirect processes, while responses to soil drying are potentially even more complex. The genes involved in the control of transpiration, for example, might be any of those involved with the control of processes ranging from stomatal development or function, to any aspect of root–shoot signalling including ABA metabolism, to the control of xylem development and root or canopy architecture, through to the control of membrane transport and aquaporins. Indeed as in most such studies, very few of the detected QTLs were consistent from year to year or from treatment to treatment, providing further evidence for the complexity of the responses and the multigenic control of all the observed characters. A further potentially confounding aspect of the present study is that plants were pot-grown, thus limiting the feedbacks between root growth and water availability in a way that may not be relevant in the field.

The key requirement now is to build on this initial work to further refine our understanding of the genetic complexity underlying the drought response phenotypes. The work also needs to be extended to answer the critical questions relating to which rootstock phenotype actually has greatest effect on fruit yield and quality in different hydric environments. This will be critical before any useful generalizations or recommendations on rootstock choice or breeding strategies can be made.