The Pennsylvania State University. The Graduate School. College of Agricultural Sciences A COMPARISON OF ANATOMICAL TRAITS RELATED TO THE

Transcription

1 The Pennsylvania State University The Graduate School College of Agricultural Sciences A COMPARISON OF ANATOMICAL TRAITS RELATED TO THE DEVELOPMENT OF BRITTLE BUD UNIONS IN APPLE A Thesis in Horticulture by Michael R. Basedow Copyright 2015 Michael R. Basedow Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science May 2015

2 The thesis of Michael R. Basedow was reviewed and approved* by the following: Robert M. Crassweller Professor of Horticulture Thesis Advisor Richard P. Marini Professor of Horticulture Head of the Department of Plant Science Kathleen M. Brown Professor of Plant Stress Biology Nicole R. Brown Associate Professor of Wood Chemistry *Signatures are on file in the Graduate School. ii

3 ABSTRACT Anatomical and histological traits related to the development of mechanically weak unions were evaluated between relatively weak ( Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ Geneva 41, Scilate (Envy )/ Geneva 41 ) and strong ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA, Cripps Pink cv.maslin/ M.9 NAKB T337, Scilate (Envy )/ M.9 NIC29 ) scion/rootstock combinations of apple. The objectives of these studies were to identify differences between these groups to determine the cause of the weak unions, and to determine if these methods would allow for the rapid screening of future potentially weak scion/rootstock combinations. Discolored sapwood was observed in the rootstock tissues of all of the combinations, and was investigated because of its association with wood decay. Trees were cut longitudinally and the discolored wood was photographed and quantified as a percentage of the total wood area using ImageJ image analysis software. Discolored sapwood did not differ significantly between weak and strong union combinations, but there was significantly more present in the most vigorous rootstocks compared to the dwarfing rootstocks. The thickness of the fiber cell walls was measured between the combinations. Tissues were hand sectioned from three areas of the tree (below, at, and above the union) and were stained with toluidine blue for microscopic study at 400x magnification. Cell wall thickness was measured using photomicrographs and Olympus CellSens software. Fiber cell wall thickness varied between some of the weak and strong combinations below and at the union. Few differences were observed between the Honeycrisp and Zestar! combinations in the scions, but the Cripps Pink combinations varied there and both combinations of Scilate were very thin. Xylem cells were divided into three categories of tissue based on their function within the wood. These categories included parenchymatous, fibrous, and conductive tissues. Xylem tissues from the most recent growth ring were sectioned and stained using toluidine blue. They were then examined at 200x magnification. The percentages of the three cell types were calculated between the combinations using ImageJ. Weak iii

4 combinations contained more parenchymatous tissue and slightly less fibrous tissue than some of the strong combinations, and may be a sign of localized incompatibility between the scions and rootstocks. Laser Ablation Tomography (LAT) and an iodine starch indicating test were evaluated for their potential use for observing localized incompatibility at the union. LAT was performed by sectioning union samples and ablating the wood material. Images were simultaneously taken, and images were combined using Avizo imaging software to virtually reconstruct the sample into a three dimensional model. Using LAT we were able to distinguish differences in the tissues of weak and strong scion/rootstock combinations, including large areas of swirling xylem tissue in the weak combination Honeycrisp / M.26 EMLA. However, abnormalities were also observed in some sections of the strong combinations. Iodine starch tests were conducted by cutting unions into longitudinal sections and then staining them with a 5% solution of iodine-potassium iodide. The iodine starch test was inconclusive, as the amount of tissues that stained varied greatly within each combination. Our results suggest anatomical differences can be determined between weak and strong scion/rootstock combinations, but these methods may not allow for the prediction of future weak unions due to their variability. iv

5 Table of Contents List of Tables... vi List of Figures... vii Acknowledgements... ix Chapter I. Introduction... 1 The role of the dwarfing rootstock in the modern orchard... 1 Graft union failure of clonal rootstocks... 2 Wood decaying fungi in apple... 3 The anatomical properties of dwarfing... 4 The anatomical properties of brittle wood... 6 Graft failure related to localized incompatibility... 7 Other symptoms commonly associated with incompatibility... 9 Potential biochemical factors of incompatibility Viral infection and incompatibility Microscopic examination of graft unions Other potential screening methods for incompatibility Hypotheses and objectives Literature Cited Chapter II. The presence of discolored wood in the xylem tissue of strong and brittle scion/rootstock combinations of apple Introduction Materials and Methods Results and Discussion Literature Cited Chapter III. Anatomical differences between the xylem of strong and brittle scion/rootstock combinations of apple Introduction Materials and Methods Results and Discussion Literature Cited Chapter IV. Potential methods for the rapid determination of localized incompatibility within the xylem of strong and brittle scion/rootstock combinations of apple Introduction Materials and Methods Results and Discussion Literature Cited Chapter V. Closing thoughts Literature Cited Appendix v

6 List of Tables Table 1.1. Combinations of scions and rootstocks used in this study, their relative strength, and their abbreviations used in this paper Table 2.1. Percent of discolored wood in the rootstock tissues of eight scion/rootstock combinations of apple Table 2.2. Percent of discolored wood in the rootstock tissues of eight scion/rootstock combinations by rootstock cultivar Table 3.1. Mean fiber cell wall thicknesses (µm) 7.0cm below, at, and 3.0cm above the unions of eight scion/rootstock combinations of apple Table 3.2. Percentages of wood tissues by combination in the unions of four scion/rootstock combinations of the apple scion cultivars Honeycrisp and Zestar! on rootstocks M.26 EMLA and M.7 EMLA. More parenchymatous tissue and less fibrous tissue were observed in the weak combination H Table 3.3. Percentages of wood tissues by combination at the unions of four scion/rootstock combinations of Cripps Pink and Scilate on the rootstocks G.41, M.9 NIC29 and M.9 NAKB T337. More parenchymatous tissue was observed in the weak combinations P41 and S Table 4.1. The percentages of wood tissues stained by I 2 KI vi

7 List of Figures Figure 2.1. Images showing both faces of the two central longitudinal sections of the combination Honeycrisp / M.7 EMLA with the scion at the top and the rootstock below; discolored wood tissue is present in the upper section of the rootstock where the upper portion of the rootstock was cut off during the budding process Figure 2.2. A transverse section of xylem tissue from the discolored wood of the combination Zestar! / M.7 EMLA with ray parenchyma cells (P) darkened with chemical extractives, and the vessel elements (V) plugged with gums Figure 2.3. The discolored wood in the combinations S41 (A) and H7 (B). In S41 the discolored wood (D) is confined to an area close to the initial wound, showing that it is well compartmented. In H7, the discolored wood extends longitudinally down the rootstock Figure 3.1. The effect of cultivar and rootstock on fiber cell wall thickness 3.0cm above the union for Honeycrisp and Zestar! Figure 3.2. The effect of cultivar and rootstock on fiber cell wall thickness at the union for Honeycrisp and Zestar! Figure 3.3. The effect of cultivar and rootstock on fiber cell wall thickness 7.0cm below the union for Honeycrisp and Zestar! Figure 3.4. Xylem tissues at the unions of H26 (A) and H7 (B). The xylem of H26 consists of many parenchyma cells (P), scattered areas of fiber cells, (F) and some vessels elements (V). The ray parenchyma (RP) cells are poorly defined. The tissue of H7 appears more uniform with many fiber cells. Vessel elements are large, and there are less parenchyma cells. The ray parenchyma cells also appear well organized into thin rays 59 Figure 4.1. A three dimensional model of a section of the union of H26 showing the longitudinal and radial planes of the section, with the rootstock on the lower left and scion on the upper right Figure 4.2. Transverse sections of wood from H26 (A) H7 (B) Z26 (C) and Z7 (D) with the scions on the left and rootstocks on the right. The wood tissue of H26 shows a large area of swirling xylem (SX) tissue within the subsequent year of growth. In H7, necrotic wood (N), callus tissue (Ca), and bark-like tissue can be seen. In Z26, an area of necrosis surrounded by callus tissue can also be observed. Z7 also shows a small section of bark-like necrotic tissue. Fragments of the callus tissue that initially bridged the gap between the rootstock and scion can be seen within the unions of H26 and H vii

8 Figure 4.3. Unions of H26 (A), H7 (B), Z26 (C) and Z7 (D) in longitudinal view with the rootstock on the left and the scion portions on the upper right. Swirling xylem (SX) is circled at the middle of the union extending towards the bark in H26. H7, Z26, and Z7 appear to have isolated areas of necrosis (N). Callus tissues (Ca) and empty spaces surrounding them between the rootstock and scion can be easily distinguished in H7 and Z26. The wood tended to split at this callus layer during the ablation process, producing these gaps. An additional small area of callus is seen in Z26. Open spaces further down the union of H26 and in Z26 (arrows) were very thin gaps also likely caused by the ablation process Figure 4.4. The unions of H26 (A) H7 (B), Z26 (C), and Z7 (D) with the scion on the upper left and the rootstock below. The images have had the brightness threshold adjusted, and the tissue stained by the iodine solution appears in red. The red shows the concentration of parenchyma tissue at the union. In the unions of H26 and Z26, the parenchyma tissues are in a line that corresponds to where the scion and rootstock join, and is fairly continuous across the union. The union of Z7 contains parenchyma, but a large section of the union consists of woody tissue viii

9 Acknowledgements Many people were involved in this study, and I would like to thank them all. I thank my advisor Rob Crassweller for his help through every step along the way. I thank Rich Marini for helping me with my statistical analysis, Kathleen Brown for helping me become a better writer, and Nicole Brown for her guidance in conducting my microscopy experiments. I thank the Eissenstat, Brown, Lynch, and Brown lab members, who helped me learn the techniques used throughout this paper. Finally, I thank my family and friends for their continued support. ix

10 Dedication To my grandparents, for taking me apple picking. x

11 Chapter I Introduction Recently, nurseries on the west coast of the United States have found that budded trees of certain scion/rootstock combinations are breaking at the union after being planted in the year following chip-budding (Manly, personal communication). When these trees do not break in the nursery, they are prone to breaking when planted in the orchard (Walsh, personal communication). In the following studies, different anatomical features of known weak and strong scion/rootstock combinations of apple [Malus xsylvestris (L.) Mill. var. domestica (Borkh.) Mansf.] were investigated to determine the cause of their structural weakness, with the hope that these features could then be used to predict future weak combinations. The role of the dwarfing rootstock in the modern orchard Prior to the adoption of dwarfing rootstocks, apple orchards in the United States primarily consisted of scion cultivars grafted onto seedling rootstocks. These sexually propagated rootstocks created large trees that were difficult to manage in an orchard setting. In the 1920 s the Malling series of rootstocks was brought to the United States, and many orchards still plant their trees on the dwarfing rootstock Malling 9 ( M.9 ) to this day. In addition to the original Malling Series and its derivatives, other breeding programs have released rootstocks with additional qualities valued by growers. The Budagovsky series was developed in the former USSR and is well known for its cold hardiness (Ferree and Carlson, 1987). The Cornell-Geneva series was developed specifically for North American growing conditions. The rootstocks are very cold hardy, and many are resistant to the diseases fire blight [Erwinia amylovora (Burr)] and Phytopthora root rot (Phytopthora cactorum) (Robinson et al., 1999). With these and many other rootstocks currently available, orchardists can choose rootstocks that are best suited for their specific orchard conditions and production goals. 1

12 Graft union failure of clonal rootstocks While clonally propagated rootstocks have proven useful, they are not without their limitations. One serious problem associated with some rootstocks is the tendency for the scion to break off at the graft union; this condition is commonly referred to as graft failure (Robinson et al. 2003). This damage is commonly observed after a heavy wind storm, but can be reduced by supporting the trees by tying them to a post or growing them on a trellis system. However, this will not guarantee protection, as damage can still occur when these precautions are taken (Robinson and Hoying, 2004). Many combinations of scion and rootstock are prone to this failure, including Honeycrisp grown on Malling 26 East Malling-Long Ashton ( M.26 EMLA ) (Privé et al. 2011). This combination has been particularly concerning because Honeycrisp is a popular cultivar that continues to be regularly planted. Honeycrisp scions had previously been planted on M.26 EMLA in northern climates because they are both known for their cold hardiness, but many trees have been lost to graft failure. To keep producing Honeycrisp, some growers began to plant the cultivar on the more vigorous rootstock Malling-Merton 106 ( MM.106 ) (Privé et al. 2011). This is not an ideal solution though, because MM.106 produces a larger tree that is difficult to manage in a high-density training system. The rootstock is also more susceptible to crown root rot (Phytopthora sp.) and performs poorly in wet soils (Cline, 2005; Privé et al., 2011; Wilson, 2000). An updated training system has recently been proposed to better manage Honeycrisp. Rather than using semi-dwarfing rootstocks like M.26 EMLA or MM.106, growers may be able to successfully manage Honeycrisp on dwarfing rootstocks in high-density training systems. Trees are tied to a support system immediately after planting, and can be grown at less than 0.6m apart to maximize planting densities. This system uses Honeycrisp on very dwarfing rootstocks like M.9 and Budagovsky 9 ( B.9 ), along with some recent dwarfing Geneva rootstock releases (Warner, 2014). Since the Budagovsky and Geneva series were bred for cold hardiness, this system may be a viable solution for northern growers. However, it is possible that some of these combinations may produce weak trees as well. 2

13 In addition to Honeycrisp / M.26 EMLA, many other scion/rootstock combinations are prone to developing brittle unions. A few of the Geneva releases are also prone to producing brittle graft unions with some scion cultivars (Robinson et al., 2003). Rootstock trials around the United States have found that the popular cultivar Gala is susceptible to graft failure when grown on Geneva 30 ( G.30 ). This rootstock was popular because it produces a tree similar in size to M.7 EMLA, but is more cold-tolerant and exhibits increased precocity (Robinson et al., 2003). Washington nurseries growing other Geneva rootstocks have encountered similar damage (Manly, personal communication). They noted that combinations of the scions Cripps Pink cv. Maslin and Scilate (Envy ) have proven exceptionally brittle when planted on G.41, a dwarfing rootstock similar in size to some clones of M.9 and B.9. Reports from Maryland (Walsh, personal communication) have also indicated that Cripps Pink / G.935 combinations are also prone to graft failure. As more scions and rootstocks are released by breeding programs, many combinations will need to be tested for brittle unions. Laboratory tests have been performed to determine the strength of some scion/rootstock combinations. Rehkugler et al. (1979) developed a method for measuring the force required to cause graft failure. They created a device that applied a measurable force to the graft unions of mature trees and found that Golden Delicious / M.9 could withstand only one third of the force that caused Golden Delicious to fail on seedling rootstock. This system could be used to evaluate other combinations, though it does not give any insight into the cause of the weakness between the scion and rootstock. In order to better manage the development of weak combinations, the cause of the weakness should be identified. Wood decaying fungi in apple Wood decay can substantially weaken the wood of apple trees, which can lead to tree losses in commercial orchards (Darbyshire et al., 1969). Propagation wounds associated with the budding process may allow for the colonization of the union by wood decaying fungi, which may weaken the trees and ultimately lead to union failure. Apple wood is susceptible to a number of wood-rotting fungi. When 200 apple trees were 3

14 examined across Washington, 44% of trees contained decayed wood. Microbial isolations were taken from these trees, and seventeen species representing fourteen genera of fungi were identified (Dilley and Covey, 1980). The development of wood decay is a complex process that begins when a tree is injured. Eide (1940) described how many decayed apple trees had originally experienced winter injury, and Dilley and Covey (1980) found that decay was often present in trees with unhealed wounds from winter or pruning injuries. Once the inner wood becomes exposed to the environment through wounds, the tree usually begins to induce changes to the wood in the area immediately surrounding the wound to reduce the possibility of microbial invasion. Parenchyma cells near the wound site produce defensive compounds that discourage microbial growth, and then die (Shigo and Hillis, 1973). New parenchyma cells that differentiate from the cambium have highly suberized cell walls that aid in blocking off the wound from future healthy wood. This response is part of a process known as compartmentalization (Shigo and Marx, 1977). If the tree is healthy, these barriers can effectively prevent the further invasion and spread of microorganisms (Shigo and Hillis, 1973). However, if the tree is stressed or disease pressure is high the tree may not be able to create these defenses effectively. Without these barriers in place pioneering microorganisms can invade the wounded tissue. These organisms include bacteria and nonhymenomycetous fungi. These break down the defenses established by the wood and may spread further into the healthy wood tissue. As the defenses are further broken down by microorganisms, the wood may then be vulnerable to wood decaying fungi (Shigo and Hillis, 1973). Hymenomycetous fungi are most commonly associated with wood decay, though other types of fungi have also been isolated (Dilley and Covey, 1980). The wood decay fungi feed on the lignin and cellulose of the xylem cell walls, reducing the strength of the tree (Schwarze, 2007). The anatomical properties of dwarfing In addition to wood decay, the anatomy of the tree itself may be useful in explaining why some combinations are more likely to produce a brittle union. A microscopic study of rootstock tissue was conducted to observe the anatomical 4

15 differences between dwarfing and vigorous rootstocks to better understand the mechanisms of dwarfing. Results of this work showed that the roots of dwarfing rootstocks contain a higher proportion of bark than wood tissue (Beakbane and Thompson 1947). Lockard and Schneider (1981) suggested that differences in the bark tissues of dwarfing rootstocks may be the most important feature in determining the overall size potential of the tree. They found that the combination Gravenstein / MM.111 grew less than control trees when it had a 10.0cm section of bark tissue from an M.26 rootstock inserted into its scion bark. The roots of dwarfing rootstocks contain a higher proportion of living cells than more vigorous rootstocks (Beakbane and Thompson, 1947). When scions of Cox s Orange Pippin were grafted onto the rootstocks M.7 and M.9, the resulting trees produced a higher proportion of parenchyma cells in roots of M.9 (71%) than in M.7 (57%). The roots of M.7 contained more fiber cells (32%) than M.9 (20%) (Beakbane and Thompson, 1947). Komarofski (1947) also found that M.9 had a tendency to produce large areas of parenchyma cells in its stems below the union. A later study (McKenzie, 1961) found that the proportions of different cell types in root xylem remained relatively constant between unworked rootstocks, and found that M.9 contained 40% ray parenchyma cells, while the vigorous rootstock M.16 contained only 20%. Interestingly, while these proportions did not change in unworked rootstocks, McKenzie found that these percentages changed up to 10% when different scions were grafted onto them. This variation suggests that scions may influence the production of xylem cells within the rootstock tissues. Dwarfing interstocks also impact scion cell dimensions. Doley (1974) observed fiber cells in scions of Cox s Orange Pippin on the rootstocks M.8 and M.9. He also investigated the same scion cultivar on MM.104 rootstocks that had an interstock of either MM.104 or M.20 inserted between the rootstock and the scion. He found that the walls of scion fiber cells were thinner when they were grafted to the very-dwarfing M.20 interstock. Studies have also investigated differences in vessel sizes between dwarfing and vigorous rootstocks. McKenzie (1961) found that xylem vessels in the roots of the dwarfing rootstock M.9 were smaller in diameter than those of the vigorous M.4. 5

16 Soumelidou et al. (1994a) found that the xylem vessels at the union of Bramley s Seedling were larger in diameter on M.9 than those on MM.106 during the union s early formation. However, the vessels of trees on M.9 were smaller in the subsequent year of growth. They attributed the early large vessel diameter of M.9 to a lack of auxin at the graft union, as they suggested auxin would lead to many small cells. They attributed the small vessel size in the subsequent year to an accumulation of auxin at the union, suggesting its flow was hindered at the union in M.9. The sweet cherry (Prunus avium L.) Rainier had significantly smaller xylem vessel lengths and diameters on the dwarfing rootstock Gisela 5 ( Gi.5 ) when compared to the same variety on the vigorous rootstocks Colt, F12/1, and Gi.6. However, Gi.5 grafted onto Gi.5 had the largest vessel cell diameters of all combinations (Olmstead et al., 2006). This combination was produced to show that the interaction of the rootstock and scion was likely to play a role in vessel size reduction, rather than from the mechanical injury of the grafting process alone. A reduction in vessel size from dwarfing rootstocks was also observed in studies of plum (Prunus domestica L.) (Gradinariu et al. 2011) and pear (Pyrus pyrifolia) (Zlati et al. 2011). The anatomical properties of brittle wood The anatomical composition of brittle wood has also been studied. The wood of apple tree limbs highly susceptible to breaking under heavy crop load contained irregular parenchyma tissue. The parenchyma cells were described as appearing halted in their development (Aaron and Clarke, 1949). Simons (1975) found that the wood of the strong cultivar Jonared had more fiber cells than the brittle Golden Delicious. He suggested that the high proportion of fiber cells could be responsible for the stem s relative mechanical strength and flexibility compared to the brittle cultivar. It is possible that the proportion of fiber to parenchyma cells within the wood can have profound effects on the strength of the graft union. Fibers are the primary cells associated with the mechanical support of the tree (Winandy and Rowell, 2013). They have a thick lignified secondary cell wall layer composed of cellulose strands that are embedded in a matrix of hemicelluloses and lignin. These cell wall properties provide the cells their strength and flexibility (Déjardin et al., 2010). Parenchyma cells are alive 6

17 at maturity, and they are primarily involved in the transport and storage of metabolic materials like carbohydrates and lipids. In addition to managing nutrients, parenchyma cells function in wound repair. Shortly after a tree is wounded, parenchyma can divide and differentiate into callus cells. These cells can then differentiate into cambium, which will then form other types of xylem cells (Myburg and Sederoff, 2001). Additionally, when trees are wounded axial parenchyma produce a boundary layer that helps to wall off the injury from the rest of the wood (Shigo and Marx, 1977). Brittle wood has been found in some dwarfing rootstocks, such as the brittle roots of the dwarfing rootstock B.9 (Ferree and Carlson, 1987). However, in a review article Webster (2004) asserts that while a dwarfing rootstock will occasionally prove structurally weak, the two characteristics are not directly linked, and many exceptions exist. While G.30 is similar in size to the semi-dwarfing rootstock M.7, G.30 occasionally produces trees with brittle graft unions when some cultivars are grafted to it (Robinson, et al., 2003). While M.9 produces a dwarf tree and is known for having weak roots, the graft union of the similarly sized G.41 appears to produce a weaker graft union than M.9 when certain cultivars are grown on it (Manly, personal communication). This suggests that graft failure may be caused primarily by an interaction between the specific scion and rootstock combination, rather than resulting from the growth habit of the rootstock or scion cultivar alone. Graft failure related to localized incompatibility While some scions and rootstocks may have inherently weaker wood caused by cultivar characteristics, it is likely that graft failure is exacerbated by an incompatibility of the tissues at the union. Mosse (1962) and Simons (1987) suggested that incompatibility is the primary cause of graft union failure. McCully (1983) described incompatibility in terms of seven steps that are necessary for the formation of a successful graft union. They are: 1. Formation of a necrotic zone at the interface of cut cells. 2. Extension of living cells from stock and scion into this necrotic zone. 3. Cell division to form callus. 4. Cohesion of stock and scion. 5. Differentiation of wound type vascular elements. 6. Differentiation of vascular cambium from callus cells. 7. Production of secondary xylem and phloem from reconstituted cambium. 7

18 Moore (1983) suggested that incompatibility can result when any of the steps in establishing a union are not completed, and believed that these events occur independently of one another. Incompatibility can generally be classified into two general forms: translocated and localized (Mosse, 1962). Translocated incompatibility exists when the factors leading to incompatibility are transferred from the rootstock to the scion (or vice-versa), without the need for direct tissue contact between the two. This can be confirmed when a mutually compatible cultivar is used as an interstock between the two and the resulting tree still proves incompatible. A reciprocal graft (switching the scion and the rootstock) does become compatible though, which suggests that the flow of the translocated incompatibility factors is directional. Other characteristics have been attributed to translocated incompatibility, including cellular necrosis at the union (Moore, 1983). Xylem tissue differentiates normally, but the phloem degenerates, preventing carbohydrate transport, which leads to their accumulation above the scion. The reduced transport of carbohydrates across the union causes trees suffering from translocated incompatibility to show symptoms shortly after grafting because the roots quickly become starved (Andrews and Serrano Marquez, 1993). Localized incompatibility is observed when cambial tissues of the scion and rootstock do not differentiate, leaving the xylem tissues partially discontinuous at the union. In localized incompatibility reciprocal grafts remain incompatible, but the inclusion of a mutually compatible interstock creates a compatible tree. This suggests that the direct contact of the stock and scion is necessary for localized incompatibility to occur. Rather than quickly declining, localized incompatibility is often marked by the tree s slow decline as vascular transport is slowed through the union (Andrews and Serrano Marquez, 1993; Mosse, 1962). However, during the subsequent growth of the union, the cambium may further lose its continuity. This causes the tree to produce unlignified ray parenchyma cells in place of well-differentiated xylem. This discontinuity may not happen until years after propagation, but the resulting lack of lignified wood is believed to end in graft failure (Mosse, 1962). These two classifications are general guidelines, and the same tree can exhibit both forms of incompatibility in varying degrees (Mosse, 1962). Other forms of 8

19 incompatibility have also been suggested. Delayed incompatibility was described by Moore (1983), but Andrews and Serrano Marquez (1993) do not recommend the use of this term because they believe defining incompatibility by timing is misleading. Although some trees are delayed in showing symptoms of incompatibility, in many cases the factors creating the incompatibility have long been present. They also noted that many cases of delayed incompatibility later proved to be caused by disease. They suggest that the delay is caused by the disease needing to spread from one graft partner to the other. Mosse (1962) also described incompatibility as an absence of normal vascular tissue at the graft union, replaced instead by incompletely lignified ray parenchyma tissue and described how trees generally break at this layer. Since this layer lacks interconnecting vessels and fibers, the break is smooth instead of jagged (Mosse, 1962). This definition is incomplete though, as it only applies to localized incompatibility. Andrews and Serrano Marquez (1993) suggested defining incompatibility as the failure of the graft combination to form a strong union and to remain healthy due to cellular, physiological intolerance resulting from metabolic, developmental, and/or anatomical differences. Other symptoms commonly associated with incompatibility While graft failure may be the most useful symptom for identifying localized incompatibility, many other symptoms of incompatibility have been observed (Andrews and Serrano Marquez, 1993). Internal symptoms generally precede the external manifestations. In their review article, Andrews and Serrano Marquez (1993) described how internal symptoms of incompatibility include: the degeneration of the phloem and cortex, atypical or missing axial parenchyma cells, increased peroxidase activity of the rootstock and scion, and increased deposition of lignin and polyphenols between the rootstock and scion. They continued by explaining that many of these features reduce vascular continuity, which reduces the transport of carbohydrates and nutrients across the union. This reduction in transport helps explain why carbohydrates and nutrients tend to accumulate on either side of the union. In addition to these differences, cellular necrosis is also commonly observed (Moore, 1983). 9

20 External symptoms of incompatibility include: late bud break, abnormal leaf morphology, early abscission of the leaves, reduced shoot growth, ill health, shoot dieback, and early death of the tree. An overgrowth of the union may also sometimes occur, though they noted this can also happen in compatible unions (Andrews and Serrano Marquez, 1993). They explained that inherent differences in the growth rates of the rootstock and scion can lead to union overgrowth, and that many of these trees still form compatible unions. They suggested if union failure is not observed, then two or more other symptoms should be observed together within the same tree to consider it incompatible. They recommended this because the other symptoms can be observed in compatible unions due to unrelated problems, such as poor environmental conditions or disease. Potential biochemical factors of incompatibility While the causes of incompatibility in apple are not yet fully understood, many biochemical factors are believed to be related to its symptoms, and many of these chemicals have been studied in other fruit trees to better understand their role in graft formation and incompatibility. Peroxidases are a group of enzymes associated with a number of processes in plant development, including auxin catabolism, lignification, and the synthesis of ethylene. Since many of these processes are involved in graft formation, they have been studied for their potential role in incompatibility (Güçlü and Koyuncu, 2012). A study by Feucht et al. (1983) investigated peroxidase activity in grafts of the sweet cherry scion Sam on the compatible sweet cherry rootstock F12/1 and the incompatible tart cherry (Prunus cerasus L.) rootstocks Weiroot 10 ( W 10 ), W 11, and W 13. They found that peroxidase activity was lower in incompatible unions, and that the cambial activity of these unions was greatly reduced. They suggested peroxidase may play a role in cambial cell division, and incompatible unions may not have adequate cambial division when peroxidase is limited. Another study of grafted sweet and tart cherry trees also found that compatible graft combinations produced higher levels of peroxidase at the graft union (Güçlü and Koyuncu, 2012). 10

21 However, some studies have suggested the production of peroxidase may inhibit compatibility. Copes (1978) found increased concentrations of peroxidase in Douglas fir [Pseudotsuga manziesii (Mayr) Franco] were closely associated with incompatibility. He believed increased concentrations of peroxidase may have been related to a hypersensitive response caused by a lack of recognition between the rootstock and scion cells. The lack of cellular recognition may have led to the increased suberization of the cell walls and cellular necrosis at the graft union. Incompatible unions of pepper (Capsicum anuum L.) scions grafted on tomato (Lycopersicum esculentum Mill.) rootstocks also produced more peroxidase than compatible grafts (Deloire and Hébant 1982). They suggested the increased production of peroxidase resulted in an increased rate of lignification at the union. They believed this lignification restricted the transport of water and nutrients across the graft union, which may have induced incompatibility between the rootstock and scion. Peroxidase activity was also higher in incompatible plum graft unions (Zarrouk et al., 2010), and they suggested that high peroxidase activity could reduce the differentiation of xylem and phloem and increase the amount of cellular degeneration within the union. While these studies appear to have conflicting results, Santamour (1988) hypothesized that incompatibility may be determined by the type of peroxidase that is produced, rather than by the amount. He suggested that differences in the types of peroxidase enzymes produced by the rootstock and scion may lead to different types of lignin being produced. He continued to explain how differences in the types of lignin produced by the cells could potentially lead to a lack of cellular recognition and differentiation, which could then lead to incompatibility. This has been shown experimentally (Gulen et al. 2002). Using combinations of pear (Pyrus communis L.) on quince (Cydonia oblonga L.) rootstocks, they studied isoperoxidase bands to determine if the isoperoxidases differed between compatible and incompatible combinations. The compatible scion Bartlett produced an isoperoxidase that was also present in the quince rootstock. The incompatible pear Buerre Hardy did not produce it, and they suggested the lack of the matching isoperoxidase may help to explain the incompatibility of Buerre Hardy. They suggested isoperoxidase profiling may be useful for screening graft combinations for incompatibility. 11

22 The transport of toxic compounds between the rootstock and scion has also been suggested as a possible cause of incompatibility. An incompatibility between pear (Pyrus communis L.) scions and quince (Cydonia oblonga Mill.) rootstocks could be caused by the quince producing the glycoside prunasin (Gur et al., 1968). They suggested when prunasin is transported from the quince to an incompatible pear, the prunasin breaks down and forms hydrocyanic acid. The acid then causes cellular necrosis near the union, which leads to incompatibility. Polyphenols are compounds that may damage tissues in graft unions, and have been associated with a reduction in the amount of differentiation of the callus tissues. Phenylalanine ammonia lyase (PAL) is one enzyme that is important for the biosynthesis of these phenolic compounds. Dos Santos Pereira et al. (2014) observed the gene expression and activity of PAL in the graft unions of peach scions (Prunus persica L. Batsch cv. Chimaritta ) grafted on two peach ( Capdeboscq and Tsukubal ) and one Japanese apricot (Prunus mume Sieb. et Zucc. cv. Umezeiro ) rootstocks. The incompatible combination Chimaritta / Umezeiro had the highest amount of PAL gene expression in the rootstock and scion tissues. PAL activity was also higher in the rootstock of the incompatible combination. They did not find increased PAL activity within the scion, as they believed post-transcriptional factors may have reduced its activity there. They suggested polyphenol levels may allow for incompatibility screening in future scion/rootstock combinations. Moore (1984) has suggested toxins may be present between many combinations of scions and rootstocks. He proposed compatible combinations may have other chemicals that neutralize these toxins, allowing the two to grow together. In incompatible combinations, these neutralizing chemicals may not be present in concentrations needed to prevent the toxins from affecting the tissues. Chemical recognition between cells of the rootstock and scion may play an important role in determining incompatibility. While callus material may form as a wound response regardless of recognition, having cellular recognition may allow for the development of plasmodesmata between the cell walls of the callus, ultimately connecting the tissues (Pina and Errea, 2005). 12

23 Plant growth regulators may be involved in incompatibility because they play many roles in scion and rootstock relations (Aloni et al., 2010). Auxin has long been known for its role in xylem differentiation (Roberts, 1969), and auxin plays a large role in graft formation of Arabidopsis thaliana (Yin et al., 2012). Digby and Wareing (1966) found that reducing the supply of indole-3-acetic acid (IAA) favored the production of phloem tissue instead of xylem tissue in hybrid poplar (Populus robusta). Lockard and Schneider (1981) suggested that this could help to explain the high bark-to-wood ratio commonly described in dwarfing rootstocks. Digby and Wareing (1966) found IAA promoted the elongation of xylem fiber cells, and Shininger (1971) found the cambium cells of cockleburs (Xanthium pennsylvanicum Wallr.) differentiated into xylem fibers after being treated with the auxin naphthaleneacetic acid (NAA). Soumelidou et al. (1994b) followed auxin transport in the rootstocks M.9 and MM.111. While the total amount of auxin uptake did not differ between the two, the rate of auxin transport was slower in M.9. The rate difference was determined by applying the auxin efflux inhibitor 2,3,4-triiodobenzoic acid (TIBA) to both rootstocks. TIBA reduced auxin transport in MM.111 dramatically, but only a negligible drop was observed in M.9. This finding suggested that auxin transport in M.9 may have already been restricted. The plant hormone abscisic acid (ABA) may influence auxin transport in grafted trees. Kamboj et al. (1997) found higher concentrations of ABA in the roots of dwarfing apple rootstocks resulted in lower rates of auxin transport in the tree limbs. While Jones (1986) suggested ABA was inconsistently found in trace amounts in apple, Kamboj et al. (1997) found ABA concentrations were highest in the most dwarfing rootstock cultivars. While auxin may be important for the differentiation of xylem cells, gibberellin (GA) may promote the specific differentiation of xylem fiber cells. Dayan et al. (2012) applied NAA and GA exogenously to the stems of tobacco (Nicotiana tabacum L.) and the addition of GA promoted the differentiation of fiber cells. GA also promoted the elongation of fiber cells within hybrid poplar, but this only occurred in the presence of IAA (Digby and Wareing, 1966). Jones (1986) found only inconsistent trace amounts of gibberellin in apple, making its role in apple xylem differentiation uncertain. 13

24 Cytokinins were shown to be important for the growth and development of the shoots of apple and cherry hybrids (Prunus avium L. x Prunus pseudocerasus Lindl.) (Jones, 1986). Cytokinins were shown to induce the differentiation of vascular elements in wounded stems (Aloni et al., 2010). Aloni (1982) found cytokinins played a role in the early differentiation of xylem fibers of sunflower (Helianthus annuus L.), as differentiation did not initially progress in the presence of auxin and gibberellin alone. However, later stages of differentiation were possible without cytokinins being present. Kamboj et al. (1999) observed the cytokinins zeatin and zeatin riboside in the grafts of dwarfing and vigorous apple rootstocks. Their concentrations were lowest in the xylem sap of the dwarfing rootstocks. In addition to having lower concentrations, dwarfing rootstocks may also vary in their sensitivity to cytokinins, as Jones (1986) found trees on M.9 required nearly twice as much cytokinin to achieve the same degree of shoot proliferation as other rootstocks. Viral infection and incompatibility Symptoms of incompatibility, including graft union failure, can also develop through viral infection of plant material. Tomato ringspot virus (TmRSV) is the causal agent of apple union necrosis and decline (Parish and Converse, 1981). This disease leads to anatomical features associated with incompatibility within the graft union. These include: the production of large areas of parenchyma tissue in place of fiber and vessel cells, an irregular orientation of the surrounding vascular tissues, starch accumulation above the union, and union failure (Tuttle and Gotlieb, 1985). Since these symptoms resemble incompatibility, the virus may cause the tree to induce similar biochemical changes between compatible scion/rootstock combinations that are usually found between incompatible combinations. This may further suggest that incompatibility may be caused by a lack of cellular recognition. 14

25 Microscopic examination of graft unions Anatomical studies of graft unions have identified many of the changes associated with incompatibility, beginning with the formation of a necrotic zone. Ermel et al. (1997) found necrotic zones in both compatible and incompatible combinations of pear and quince one day after budding, and noted this may have been related more to the budding process itself than a result of incompatibility. However, differences between compatible and incompatible combinations may become more apparent as the unions continue to develop. Simons and Chu (1985) found necrotic zones in the graft unions of the incompatible combination Sturdeespur Delicious / M.26. The necrotic regions were most prevalent in areas where cellular division had yet to occur, as they found less necrosis where callus cells were already dividing elsewhere in the same union. Ussahatanonta and Simons (1988) found necrotic tissue at the graft union when Golden Delicious was grafted onto rootstocks of varying vigor classes (seedling, MM.106, M.7, M.26 and M.9). They found the greatest amount of necrosis in the most dwarfing rootstocks, and noted that necrotic areas were still evident in the subsequent year s growth in M.26 and M.9. A lack of regular cambium differentiation was also observed. Ermel et al. (1997) found incompatible combinations of pear and quince had difficulty joining the cambia of stock and scion after budding. The cambia of the incompatible combinations had to join around patches of callus tissue, while cambia of compatible unions differentiated straight across the callus. Zarrouk et al. (2010) found similar problems in incompatible combinations of peach on cherry plum (Prunus cerasifera Ehrh.) and noted five-monthold incompatible grafts still lacked connected cambia. They also found callus cells appeared more uniform in compatible combinations. Errea et al. (1994) found poorly differentiated callus tissues in the graft unions of apricot (Prunus armeniaca L.) on cherry plum rootstocks. The lack of cambium differentiation may lead to poor xylem development (Mosse, 1962). Ussahatanonta and Simons (1988) found the xylem of Golden Delicious / M.26 still contained areas of undifferentiated callus parenchyma cells two years after budding. They noted the vascular tissues surrounding the callus tissue were in an irregular, swirling orientation. Irregularly oriented ray parenchyma cells were present 15

26 in the unions of Sturdeespur Delicious / M.26 (Simons and Chu, 1985) and Golden Delicious / M.9 (Ussahatanonta and Simons, 1988). Warmund et al. (1993) observed areas of parenchyma developing where normal vascular tissue should have differentiated in the union of Jonagold / Mark. They noted trees broke along the line of this tissue, though they did not say if the parenchyma was callus tissue or ray parenchyma cells. The amount of parenchyma cells relative to other xylem cells has also been documented. Herrero (1951) observed the relative proportion of xylem cell types in the scions of the peach Hale s Early on rootstocks of cherry plum Myrobalan B and the domestic plum Brompton. His data showed the compatible combinations of Hale s Early / Brompton produced more fiber cells and less ray parenchyma cells in their scion tissues than those on Myrobalan B. However, statistical analysis was not performed between these combinations. Since fiber cells provide much of the mechanical support to the tree (Winandy and Rowell, 2013), having large areas of disordered parenchyma tissues in place of fiber cells at the graft union may increase the risk of union failure (Mosse, 1962; Simons, 1987). Even if this parenchymatous tissue only persists in a small portion of the wood, it might be problematic for the tree. Niklas and Spatz (2012) discussed how imperfections within an object can increase the concentration of stress forces acting on it. They suggested that the fracture strength of an object results from its structural homogeneity because the forces of the stress are well distributed throughout the structure. When imperfections are present, the stress force converges and increases at the point of the imperfection, increasing the likelihood of fracture. In this case, a relatively small area of weak parenchyma tissue may break up the relative homogeneity of the wood tissue, increasing the likelihood of union failure. Other potential screening methods for incompatibility While microscopy studies have identified symptoms commonly associated with incompatibility, they often require fixing the sample tissues for sectioning and only provide a limited view of the union. This has led researchers to seek other methods for evaluating unions for incompatibility. 16

27 Ermel et al. (1999) used a number of histological traits to determine incompatibility in pear/quince combinations. They used 13 histological traits related to bark discontinuity, cambial disruption, or starch accumulation above the union. Some of the traits used included: the production of periderm rings, the appearance of bark at the union, differences in tissue staining by toluidine blue, the presence of lignified ring shaped meristems, and the accumulation of starch within the scion. While they suggested that analyzing many histological traits together could help to predict incompatibility, the individual measurements on their own were too variable to use as a sole method of evaluation. Warmund et al. (1993) used Magnetic Resonance Imaging (MRI) to evaluate the graft unions of the apple combination Jonagold / Mark. They were able to distinguish between unions with good and poor vascular connectivity, as well-connected unions appeared bright, while poorly connected tissues remained dark. They attributed these differences to the different concentrations of water within the vascular tissues. Milien et al. (2012) observed unions of two self-hybrids of Syrah on the rootstock 110 Richter. They observed two unions that differed in their development; one was described as a good graft, and the other as a bad graft. Using X-ray computed tomography (Ct X- ray) scans, they were able to observe graft unions in three dimensions, and were able to distinguish tissue-level differences between the two. They found the good graft had more vascular continuity than the bad graft, but were unable to make distinctions at the cellular level due to resolution limitations. Beyond determining incompatibility by the structure of the graft union, biochemical and genetic screening have also been tested. As mentioned in the biochemical discussion of incompatibility, studies (Dos Santos Pereira et al., 2014; Gulen et al., 2002; Zarrouk et al., 2010) found differing amounts of biochemical gene expression and activity between different scion/rootstock combinations, and believed these differences might allow researchers to quickly determine incompatibility in other graft combinations. These methods appear promising, but because the biochemical nature of incompatibility is still unclear, more research in these areas is necessary before definitive biochemical screening methods can be developed. 17

28 Hypotheses and objectives From the literature presented, the brittle wood commonly observed in the combinations of Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ G.41 and Scilate (Envy )/ G.41 may be caused by many factors. The brittle wood may be caused by an increased amount of wood decay caused by fungal infection during the budding procedure, as decay would reduce the strength of the cell walls (Schwarze, 2007). Some scion/rootstock combinations may be better able to defend against wood decaying fungi than others, which would explain why some combinations are more prone to producing weak unions. Additionally, since the thick lignified secondary cell walls of the fiber cells provide the trees their strength (Déjardin et al., 2010; Winandy and Rowell, 2013), the weak unions may have thinner fiber cell walls, or they may have fewer fiber cells and more parenchyma cells. An abundance of parenchyma cells may result from the rootstocks inherent tendency to produce more, such as how M.9 was found to produce more parenchyma than fiber cells within its roots (Beakbane and Thompson, 1947). This may be, to some extent, related to the dwarfing potential of the rootstock (Simons, 1987), which may be caused by hormonal differences. However, an overabundance of parenchyma, specifically at the union, may be due to an incompatibility between the rootstock and scion because this injury has been observed in specific combinations of scion and rootstock and is not limited to the most dwarfing rootstocks. The specificity of the union failures makes other potential causes of weakness, like adverse environmental conditions and poor propagation techniques (Andrews and Serrano Marquez, 1993), less likely. We also do not suspect apple union necrosis and decline to be the cause. TmRSV requires a few years after the initial inoculation for symptoms to develop (Rosenberger et al., 1989), and many nurseries use virus-free propagation materials (Baugher Jr., personal communication). This may make graft failure from apple union necrosis and decline in young nursery plants unlikely. Though M.26 EMLA has been reported as susceptible to TmRSV, MM.106 has been reported as being very susceptible (Ferree and Carlson, 1987), yet replanting Honeycrisp on MM.106 in Canada has been successful (Privé et al. 2011). If TmRSV was the primary 18

29 cause of Honeycrisp / M.26 EMLA graft failures, we would expect Honeycrisp to fail on MM.106 as well. In the following studies, possible histological and anatomical differences between brittle and strong bud unions of commercially important cultivars (Table 1.1) were investigated to determine the cause of the union failure commonly observed in the weak combinations. Our objectives were to determine these differences, and to determine if these features and methods could then be used for the early screening of future scion/rootstock combinations that would be prone to graft failure. The differences we chose to investigate included: the presence of discolored wood in the rootstock tissues, the thickness of the fiber cell walls, the relative proportions of different types of xylem cells in the subsequent year of growth at the graft union, and the presence of parenchymatous, necrotic, and swirling tissues at the union. While Zarrouk et al. (2010) and Ermel et al. (1997) had identified differences in the development of the cambium and callus cells shortly after trees were propagated, we chose to observe wood that grew an additional year after budding. We chose these because Mosse (1962) had suggested cambial discontinuity can develop at any time, and some studies (Simons and Chu, 1983; Ussahatanonta and Simons, 1988) found more apparent differences between apple combinations after the first year of union development. We also believe this was useful for our cell proportion study, as this gave the cambium more time to potentially reorient itself and to begin producing welldifferentiated xylem tissues. 19

30 Table 1.1. Combinations of scions and rootstocks used in this study, their relative strength, and their abbreviations used in this paper. Scion Rootstock Graft Strength Abbreviation Honeycrisp M.26 EMLA Brittle H26 Honeycrisp M.7 EMLA Strong H7 Zestar! M.26 EMLA Strong Z26 Zestar! M.7 EMLA Strong Z7 Cripps Pink G.41 Brittle P41 Cripps Pink M.9 NAKB T337 Strong P9 Scilate G.41 Brittle S41 Scilate M.9 NIC29 Strong S9 20

31 Literature Cited Aaron, I. and W.S. Clarke Jr Breakage of apple trees. Proc. Amer. Soc. Hort. Sci. 54: Aloni, R Role of cytokinin in differentiation of secondary xylem fibers. Plant Physiol. 70: Aloni, B., R. Cohen, L. Karni, H. Aktas, and M. Edelstein Hormonal signaling in rootstock-scion interactions. Sci. Horti. 127: Andrews, P.K., and C. Serrano Marquez Graft incompatibility. Hort. Rev. 15: Baugher Jr. J Adams County Nursery. Personal communication. Beakbane, A.B. and E.C. Thompson, Anatomical studies of stems and roots of hardy fruit trees. IV. The root structure of some new clonal apple rootstocks budded with Cox's Orange Pippin. J. Pomol. 23: Cline, J.A Commercial production of Honeycrisp apples in Ontario. OMAFRA Factsheet Copes, D Isoenzyme activities differ in compatible and incompatible Douglas fir graft unions. For. Sci. 24: Darbyshire, B., G.C. Wade, and K.C. Marshall In vitro studies of the role of nitrogen and sugars on the susceptibility of apple wood to decay by Trametes versicolor. Phytopathol. 59: Dayan, J., N. Voronin, F. Gong, T. Sun, P. Hedden, H. Fromm, and R. Aloni Leafinduced gibberellin signaling is essential for internode elongation, cambial activity, and fiber differentiation in tobacco stems. The Plant Cell 24: Déjardin, A., F. Laurans, D. Arnaud, C. Breton, G. Pilate, and J Leplé Wood formation in Angiosperms. Comptes Rendus Biologies 333: Deloire, A. and C. Hébant Peroxidase activity and lignification at the interface between stock and scion of compatible and incompatible grafts of Capsicum on Lycopersicum. Ann. Bot. 49: Digby, J. and P.F. Wareing The effect of applied growth hormones on cambial division and the differentiation of cambial derivatives. Ann. Bot. 30: Dilley, M.A. and R.P. Covey Jr Survey of wood decay and associated hymenomycetes in central Washington apple orchards. Plant Dis. 64: Doley, D Effects of rootstocks and interstock on cell dimensions in scion stems of apple (Malus pumila Mill.) New Phytol. 73:

32 Dos Santos Pereira, I., R. Da Silva Messias, Â. Diniz Campos, P. Errea, L.E. Corrêa Antunes, J.C. Fachinello, and A. Pina Growth characteristics and phenylalanine ammonia-lyase activity in peach grafted on different Prunus spp.. Biol. Plant. 58: Eide, C.J Wood decay in apple trees in Minnesota. Phytopathol. 30: Ermel, F.F., J.L. Poëssel, M. Faurobert, and A.M. Catesson Early scion/stock junction in compatible and incompatible pear/pear and pear/quince grafts: a histocytological study. Ann. Bot. 79: Ermel, F.F., J. Kervella, A.M. Catesson, and J.L. Poëssel Localized incompatibility in pear/quince (Pyrus communis/cydonia oblonga) combinations: multivariate analysis of histological data from 5-month-old grafts. Tree Phys. 19: Errea, P., A. Felipe, and M. Herrero Graft establishment between compatible and incompatible Prunus spp.. J. Exp. Bot. 45: Ferree, D.C. and R.F. Carlson Apple rootstocks, p In: R.C. Rom and R.F. Carlson (eds.) Rootstocks for fruit crops. John Wiley and Sons. New York, NY. Feucht,W., P.P.S. Schmid, and E. Christ Compatibility in prunus avium/prunus cerasus grafts during initial phase. II. reduction of cell number and peroxidases in the rootstock cambium. Sci. Hort. 23: Gradinariu, G., C. Zlati, M. Istrate, and L. Draghia Improvement of grafting union in plum by studying graft anatomy. Acta Hort. 923: Güçlü, S.F. and F. Koyuncu A method for prediction of graft incompatibility in sweet cherry. Notulae Botanicae Horti Agrobotanici Cluj-Napoca 40: Gulen, H., R. Arora, A. Kuden, S.L. Krebs, and J. Postman Peroxidase isozyme profiles in compatible and incompatible pear-quince graft combinations. J. Amer. Soc. Hort. Sci. 127: Gur, A., R.M. Samish, and E. Lifschitz The role of the cyanogenic glycoside of the quince in the incompatibility between pear cultivars and quince rootstocks. Hort. Res. 8: Herrero, J Studies of compatible and incompatible graft combinations with specific reference to hardy fruit trees. J. Hort. Sci. 26: Jones, O.P Endogenous growth regulators and rootstock/scion interactions in apple and cherry trees. Acta Hort. 179: Kamboj J.S., P.S. Blake, J.D. Quinlan, and A.D. Webster Recent advances in studies on the dwarfing mechanism of apple rootstocks. Acta Hort. 451:

33 Kamboj, J.S., P.S. Blake, J.D. Quinlan and D.A. Baker Identification and quantitation by GC-MS of zeatin and zeatin riboside in xylem sap from rootstock and scion of grafted apple trees. Plant Growth Regulat. 28: Komarofski, B The wood anatomy of certain apple stock and scion varieties and correlative structural changes in the trunk induced by budding. Palestine J. Bot. 6: Lockard, R.G, and G.W. Schneider Stock and scion growth relationships and the dwarfing mechanism in apple. Hort. Rev. 3: Manly, N Willow Drive Nursery, Personal communication. McCully, M.E Structural aspects of graft development. p In: R. Moore (ed.). Vegetative compatibility responses in plants. Baylor Univ. Press, Waco, TX. McKenzie, D.W Rootstock-scion interaction in apples with special reference to root anatomy. J. Hort. Sci. 36: Milien, M., A.S. Renault-Spilmont, S.J. Cookson, A. Sarrazin, and J. Verdell Visualization of the 3D structure of the graft union of grapevine using x-ray tomography. Scientia Hort. 144: Moore, R Physiological aspects of graft formation, p In: R. Moore (ed.). Vegetative compatibility responses in plants. Baylor Univ. Press, Waco, Tx. Moore, R A model for graft compatibility-incompatibility in higher plants. Amer. J. Bot. 71: Mosse, B Graft incompatibility in fruit trees: with particular references to its underlying causes. Tech. Comm. 28. Comm. Bur. Hort. Plant Crops, East Malling, England. Myburg, A.A. and R.S. Sederoff Xylem structure and function. Encyclopedia of life sciences. Nature publishing group. London, UK. Niklas, K.J. and H.C. Spatz Plant Physics. 1 st ed. Univ. of Chicago Press, Chicago, IL. Olmstead, M.A., N.S. Lang, F.W. Ewers, and S.A. Owens Xylem vessel anatomy of sweet cherries grafted onto dwarfing and nondwarfing rootstocks. J. Amer. Soc. Hort. Sci. 131: Parish, C.L., and R.H. Converse Tomato ringspot virus associated with apple union necrosis and decline in western United States. Plant Dis. 65: Pina, A. and P. Errea A review of new advances in mechanism of graft compatibility-incompatibility. Sci. Hort. 106:1-11. Privé, J.P., A. LeBlanc, and C.G. Embree Preliminary evaluation of supported and free standing 'Honeycrisp' trees on 24 apple rootstocks. Acta Hort. 903:

34 Rehkugler, G.E., J.N. Cummins, and E.D. Markwardt Rupture strength of unions of Golden Delicious apple with Malling 8, Malling 9, and vigorous rootstocks. J. Amer. Soc. Hort. Sci. 104: Roberts, L.W The initiation of xylem differentiation. Bot. Rev. 35: Robinson, T.L., J.N. Cummins, W.C. Johnson, S.A. Hoying, H.S. Aldwinckle, and J.L. Norelli Orchard performance of fire blight-resistant Geneva apple rootstock. Acta Hort. 489: Robinson, T.L., G. Fazio, T. Holleran, H. Aldwinckle The Geneva series of rootstocks from Cornell: Performance, disease resistance, and commercialization. Acta Hort. 622: Robinson, T.L. and S.A. Hoying Performance of elite Cornell Geneva apple rootstocks in long-term orchard trials on growers farms. Acta Hort. 658: Rosenberger, D.A., J.N. Cummins, and D. Gonsalves Evidence that tomato ringspot virus causes apple union necrosis and decline: symptom development in inoculated apple trees. Plant Dis. 73: Santamour Jr., F.S Graft compatibility in woody plants: An expanded perspective. J. Environ. Hort. 6: Schwarze, F.W.M.R Wood decay under the microscope. Fungal Bio. Rev. 21: Shigo, A.L. and W.E. Hillis Heartwood, discolored wood, and microorganisms in living trees. Ann. Rev. Phytopathol. 11: Shigo, A.L. and H.G. Marx Compartmentalization of decay in trees. USDA Forest Service Ag. Info. Bul. No Shininger, T.L The regulation of cambial division and secondary xylem differentiation in xanthium by auxins and giberellins. Plant Physiol. 47: Simons, R.K Variation of wood tissues from two cultivars of apples. Fruit Var. J. 29(2):34-35 Simons, R.K. and M.C. Chu Graft union characteristics of M.26 apple rootstock combined with 'Red Delicious' strains-morphological and anatomical development. Scientia Hort. 25: Simons, R.K Compatibility and stock-scion interactions as related to dwarfing, p In: R.C. Rom and R.F. Carlson (eds.) Rootstocks for fruit crops. John Wily and Sons. New York, NY. Soumelidou, K., N.H. Battey, P. John, and J.R. Barnett. 1994a. The anatomy of the developing bud union and its relationship to dwarfing in apple. Ann. Bot. 74:

35 Soumelidou, K., D.A. Morris, N.H. Battey, J.R. Barnett, and P. John. 1994b. Auxin transport capacity in relation to the dwarfing effect of apple rootstocks. J. Hort. Sci. 69: Tuttle, M.A., and A.R. Gotlieb Graft union histology and distribution of tomato ringspot virus in infected McIntosh/Malling Merton 106 apple trees. Phytopathol.75: Ussahatanonta, S. and R.K. Simons Graft union development of the Golden Delicious apple when combined with various dwarfing rootstocks. Fruit Var. J. 42: Walsh, C University of Maryland. Personal communication. Warmund, M.R., B.H. Barritt, J.M. Brown, K.L. Schaffer, and B.R. Jeong Detection of vascular discontinuity in bud unions of Jonagold apple on mark rootstock with magnetic resonance imaging. J. Amer. Soc. Hort. Sci. 118: Warner, G Choose the right rootstock for honeycrisp. Good Fruit Grower. April Webster, A.D Vigour mechanisms in dwarfing rootstocks for temperate fruit trees. Acta Hort. 658: Wilson, K Apple Rootstocks. OMAFRA Factsheet Winandy, J.E. and R.M. Rowell Chemistry of wood strength. p In: R.M. Rowell (ed.). Handbook of wood chemistry and wood composites. CRC Press. Taylor & Francis Group. Boca Raton, FL. Yin, H., B. Yan, J. Sun, P. Jia, Z. Zhang, X. Yan, J. Chai, Z. Ren, G. Zheng, and H. Liu Graft-union development: a delicate process that involves cell-cell communication between scion and stock for local auxin accumulation. J. Exp. Bot. 63: Zarrouk, O., P. Testillano, M.C. Risueno, M.A. Moreno, and Y. Gogorcena Changes in cell/tissue organization and peroxidase activity as markers for early detection of graft incompatibility in peach/plum combinations. J. Amer. Soc. Hort. Sci. 135:9-17. Zlati, C., G. Gradinariu, M. Istrate, and L. Draghia Histological investigation on graft formation in pear/quince (Pyrus communis/cydonia oblonga) combinations. Acta Hort. 923:

36 Chapter II The presence of discolored wood in the xylem tissue of strong and brittle scion/rootstock combinations of apple Abstract. Percentages of discolored wood in the unions of five strong ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA, Cripps Pink cv.maslin/ M.9 NAKB T337, Scilate (Envy )/ M.9 NIC29 ) and three weak ( Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ G.41, Scilate (Envy )/ G.41 ) chip-budded combinations of apple were examined to determine if discolored wood was related to the union strength of different scion/rootstock combinations. Samples were sectioned longitudinally to expose the xylem tissue at the unions. Four images of each tree were measured visually, and the amount of discolored wood was measured as a percentage of the total union wood area using ImageJ image analysis software. Few differences were observed between the weak and strong scion/rootstock combinations, however differences were observed between rootstock vigor classes, as the most vigorous rootstock M.7 EMLA produced the most discolored wood, while the least vigorous G.41 produced the least. These differences in the amount of discolored wood produced may be explained by inherent differences between rootstock cultivars, such as their ability to form vessel occlusions, or a difference in the amount of parenchyma cells present within their xylem. Introduction Apple [Malus xsylvestris (L.) Mill. var. domestica (Borkh.) Mansf.] nurseries on the west coast of the United States have recently found that certain scion/rootstock combinations of young budded trees are prone to breaking at the union (Manly, personal communication). If these trees do not break at the nursery, they may be prone to breaking later when planted in the orchard (Walsh, personal communication). Union failure was observed previously in other unrelated combinations of scion and rootstock (Privé et al. 2011). Many factors are believed to cause this damage, including adverse environmental conditions, disease, and poor propagation practices (Andrews and Serrano Marquez, 1993). In this study, the amount of discolored wood at the union, believed to be related to the development of wood decay (Shigo and Hillis, 1973), was observed to determine if it may be the cause of the weakness observed in the scion/rootstock combinations Cripps Pink cv.maslin/ Geneva 41 ( G.41 ), Scilate (Envy )/ G.41, and Honeycrisp / Malling 26 East Malling-Long Ashton ( M.26 EMLA). 26

37 Many types of xylem cells make up the wood of trees. Some cell types commonly observed in the xylem of angiosperms include fiber cells, vessel elements, and parenchyma cells. These cells differ in their morphology and growth characteristics, allowing them to serve different roles within the xylem tissue. The vessel elements are hollow, and are stacked into long tubes that conduct water throughout the tree. The fiber cells are long and have thick cell walls that allow them to aid in the structural support of the tree (Winandy and Rowell, 2013). The fiber cells and vessel elements develop a secondary cell wall layer as they grow. Once this wall layer is completed, the cells die and are non-living at maturity (Déjardin et al., 2010). Parenchyma cells store starch and lipids within the woody tissue, and move these and other assimilates either within the xylem (axial parenchyma) or between the xylem and phloem (ray parenchyma) of the wood. They also function in repairing wounds within the wood of the tree by rapidly dividing and differentiating into new cells (McCully, 1983). The parenchyma cells also produce a secondary cell wall, but they remain living after this layer forms (Déjardin et al., 2010). Wood that still contains living parenchyma cells is commonly referred to as sapwood. Sapwood consists of the younger functioning xylem tissue of the tree, and its role is to support the weight of the tree, conduct water, heal wounds, and move and store nutrients (Déjardin et al., 2010; Shigo and Hillis, 1973; Taylor et al., 2002). Heartwood is another wood classification, and it is created when parenchyma within a section of xylem die. It forms from older xylem tissues and commonly occurs in the oldest growth rings of trees. The amount of heartwood present within a tree is variable and depends on a number of conditions, including the age of the tree and environmental conditions. Not all tree species will form heartwood, and variation in the amount produced can exist between trees of the same species (Shigo and Hillis, 1973). The formation of heartwood begins as parenchyma cells within the wood age. Many processes are involved in the transition from sapwood to heartwood, and sometimes a distinct transition zone can be found between the two. During the transition the parenchyma switch from aerobic to anaerobic respiration. As anaerobic respiration begins, the starch, cytoplasm, and nuclei of the parenchyma begin to degrade (Taylor et al., 2002). As these contents degrade, the energy from the starch and the other metabolic 27

38 compounds within the parenchyma are used to produce chemical extractives that will then fill these cells (Nair, 1988). These extractives are found primarily in the lumen of the cells, but they can also be found in the pit borders between cells and deposited on cell walls (Shigo and Hillis, 1973). In addition to forming extractives, the parenchyma cells also occupy the lumens of the vessel elements by producing tyloses or gums (Shigo and Hillis, 1973). Tyloses are formed when parenchyma adjacent to the vessel elements grow through the pits to fill the element, while gums are created by chemicals that are secreted by the parenchyma cells (Bamber, 1987). Though these vessel occlusions can sometimes be observed in sapwood, they are more abundant in heartwood. Once these changes occur, the final phase of heartwood formation occurs when the parenchyma cells die (Shigo and Hillis, 1973; Taylor et al., 2002). Since vessels become plugged by occlusions, heartwood is incapable of conduction. The primary functions of heartwood are to allow for the resorption of cell nutrients as the parenchyma die and to provide structural support to the tree. Heartwood may also help to protect the tree from decay, since many of the chemicals within the parenchyma and the occlusions of the vessels inhibit fungal growth within the wood (Bamber, 1987; Blanchette, 1979; Taylor et al., 2002). Baker (1933) found discolored wood in the stems of apple and believed it was normal heartwood. He suggested the heartwood had formed by the desiccation of the wood after being exposed to the air through pruning wounds. A later study by Nilsson et al. (2002) investigated the formation of heartwood by mechanically injuring the trunks of Scots pine (Pinus sylvestris L.). They found that wounding creates discolored wood tissue similar in composition to heartwood. However, when Hart (1968) observed the xylem of Osage orange [Maclura pomifera (Raf.) Schneid.] and black locust (Robinia pseudoacacia L.), he described differences between what he considered to be true heartwood and discolored sapwood caused by mechanical injury. He suggested that chemical differences exist between the two types of tissue, even though they are similar in appearance. He found that heartwood and discolored sapwood differ in many characteristics (color, water content, frequency of amorphous deposits, solubility in water and 1% NaOH, ash content, and ph), and suggested that the two should be regarded as different tissues. 28

39 While chemical differences exist, heartwood and discolored sapwood have a few similar qualities (Hart, 1968). Both contain non-living parenchyma cells, as evidenced by the lack of a color change when triphenyltetrazolium chloride is applied. Both tissues lack starch granules, and both contain plugged vessel elements. Shigo and Hillis (1973) also described that the formation of discolored wood is similar to heartwood, with the exception that heartwood is formed by internal factors, and discolored sapwood by external. While these two tissues can be difficult to differentiate, healthy apple wood does not ordinarily contain heartwood, and any discolored wood present is likely to be discolored sapwood caused by wounds from pruning, cold injury, or other injuries caused through improper management (Rosenberger, 2007). Since discolored sapwood results from wounds to the tree, researchers have questioned whether it forms from wounding injury alone, or if microorganisms are involved. To determine this, Sucoff et al. (1966) observed discoloration in the wood of mechanically wounded aspen (Populus tremuloides Michx). In one of their experiments, some trees were bored and injected with differing concentrations of biocides to prevent microbial infection. These included streptomycin, the fungicide semesan, and formaldehyde. Wood samples were examined microscopically twenty days after the treatments were applied. Very few bacteria or fungi were found within the wood of the treated trees, suggesting any observed changes would have been caused by wounding alone. No significant decreases in wood discoloration were observed in the treated trees. They also observed that higher concentrations of biocides led to more discoloration, which led them to suggest that discoloration may be caused by wounding alone. Discolored sapwood is commonly observed in decaying trees (Dilley and Covey, 1980). Apple wood is commonly decayed by a number of species of wood-rotting fungi. When 200 trees were examined in Washington, 52.5% of trees contained discolored sapwood, while 44% had undergone some form of decay (Dilley and Covey Jr, 1980). Seventeen species representing fourteen genera of fungi were isolated from these trees (however, some species have since been renamed). Of the 200 trees observed, 190 contained the wood-rotting golden needle mushroom Flammulina velutipes. The second most commonly identified fungus (24 of 200 trees) was Trametes versicolor (referred to in the paper as Coriolus versicolor). This fungal pathogen has been shown to cause limb 29

40 dieback and papery bark disease in apple (Dilley and Covey, 1981). This suggests that discolored sapwood may be a precursor to decay in the wood of apple. Since discolored sapwood and decay occurred together in wounded wood, the role that discolored sapwood plays in the formation of decay has been studied. Shigo and Hillis (1973) described how the decomposition of wood can be divided into three stages. The first stage involves the production of discolored sapwood when a tree is wounded. They suggested that the color changes are caused by the production of chemical extractives by the parenchyma cells and the oxidation of the wood upon contact with the air. In the second stage, the wood tissue is invaded by primary invading microorganisms, which mostly consist of nonhymenomycetous fungi and bacteria. These organisms continue to change the color of the wood as they interact with the remaining living parenchyma cells. In stage three, the wood is invaded by wood rotting fungi. These fungi decay the tree by removing the lignin and cellulose from the walls of the xylem cells, which may ultimately decrease the strength of the tree (Schwarze, 2007). Researchers also studied whether discolored sapwood serves as a defense to decay, or if it favors its production. Shigo and Hillis (1973) stated that when healthy trees are wounded, the decay process often stops at stage one, suggesting discolored wood can serve as an effective deterrent to microbial decay in many cases. Shortle and Cowling (1978) also indicated that discolored sapwood initially serves as a defense against the growth of wood decaying fungi. Blanchette (1979) observed the progression of microorganisms in the xylem of apple and found the wood could not be adequately decayed by wood decay fungi until the discolored sapwood had been decolorized first by the primary invading microorganisms. This decolorization process included the removal of the chemical extractives toxic to decay fungi, the removal of the vessel gums, and the breakdown of pit membranes between the cells of the wood. Discolored sapwood may resist microbial decay because it has similar chemical extractives of those found in heartwood (Scheffer and Cowling, 1966). The decay resistance of discolored wood has been tested in a few species of trees. White oak (Quercus alba L.), black locust (Robinia pseudoacacia L.), and Osage orange [Maclura pomifera (Raf.) Schneid] produce discolored sapwood that is more resistant to decay than their regular sapwood (Hart and Johnson, 1970). However, they also found that 30

41 resistance was variable between and within other species. They found that the discolored sapwood of silver maple (Acer saccharinum L.), black walnut (Juglans nigrum L.) and European beech (Fagus sylvatica L.) had no increase in decay resistance compared to their regular sapwood. In addition to chemical defenses, vessel occlusions serve as physical barriers to prevent the spread of decay (Blanchette, 1979). These vessel occlusions help to compartmentalize wound injuries to keep it from spreading into uninjured wood tissue. Shigo and Marx (1977) explained this in their Compartmentalization of Decay in Trees (CODIT) model. They described how a tree can be thought of as a cylinder of wood. When a tree is injured, the injured section of the cylinder can be blocked off from the rest of the healthy wood by sealing off the wound with four walls that form from xylem cells. One of these walls is created by occlusions of the xylem vessel elements. This wall is the weakest for compartmentalizing injury, as the occlusions may not fill the vessel quickly (Shigo and Marx, 1977), and microbes may break them down (Blanchette, 1979). The second wall is formed by the last few layers of fiber cells that divided in the previous annual ring. The third wall is formed from ray parenchyma cells. These ray cells create both of the compartment walls in the radial plane of the tree. The fourth wall has the strongest resistance to infection and consists of cells that divide from the actively growing cambium. This layer consists mostly of parenchyma cells with suberized cell walls, and is referred to as the boundary layer (Schwarze et al., 2000). While discolored sapwood may help to defend against the microbial decay of injured wood within a tree, a large area of discoloration may suggest that the tree has not effectively compartmentalized its injury after it was wounded (Shigo and Hillis, 1973). While Shigo and Hillis (1973) explained that a large area of discolored sapwood does not necessarily mean a large area of decay will follow, Shigo (1965) had stated that decay is likely to follow discolored wood tissues. If the occlusions are slow to form or are too weak to compartmentalize wounding injury, they may be less resistant to decay as well. If more wood is susceptible to decay, the mechanical strength of the tree may be further compromised. The main objectives of this study were to determine the cause of weak unions in our study combinations and to evaluate if our methods could be used to predict the 31

42 likelihood of future scion/rootstock combinations producing brittle unions. The wound caused by the removal of the upper portion of the rootstock during the budding procedure may be an area where discolored sapwood and decay is likely to form, and our initial observations made in the spring of 2014 confirmed this. Since discolored sapwood may lead to wood decay that could ultimately reduce the mechanical strength of the tree (Schwarze, 2007), we predicted that discolored wood could relate to the production of weak scion/rootstock combinations. Since this characteristic is easily observed, it might possibly be used for a rapid test of union strength. Our hypothesis was that larger areas of discolored wood would be present within the wound sites of the weak combinations ( Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ G.41, and Scilate (Envy )/ G.41 ), than within strong combinations ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA, Cripps Pink cv.maslin/ Malling 9 NAKB T337 ( M.9 NAKB T337 ) and Scilate (Envy )/ M.9 NIC29 ). Materials and Methods This study used 64 trees. In Feb. 2014, finished chip-budded apple trees were received from Willow Drive Nursery, Ephrata, WA. These were budded in 2012, and included six trees each of Cripps Pink on the rootstocks G.41 and M.9 NAKB T337 ( P41 and P9 ) and Scilate on the rootstocks G.41 and M.9 NIC29 ( S41 and S9 ). In Apr. 2014, additional chip-budded trees were received from Adams County Nursery, Aspers, PA. These included ten trees each of the cultivars Honeycrisp and Zestar! on the rootstocks M.26 EMLA and M.7 EMLA ( H26, H7, Z26, Z7 ). All trees were kept at 6 C until sampling. Beginning in May 2014, trees were cut using a circular saw to 10.0cm in length from 7.0cm below to 3.0cm above the union, and then sectioned to mm thick longitudinal sections using a band saw. Each tree yielded five to seven sections. The two sections closest to the center of each tree were photographed using a Canon Powershot A4000 camera (Canon Inc, Tokyo, Japan). Each face of the two sections was photographed, yielding four images per tree (Figure 2.1), with the exception of one sample of H26, where only one section (two images) was used. 32

43 Figure 2.1. Images showing both faces of the two central longitudinal sections of the combination Honeycrisp / M.7 EMLA with the scion at the top and the rootstock below; discolored wood tissue is present in the upper section of the rootstock where the upper portion of the rootstock was cut off during the budding process. To observe anatomical features related to discolored sapwood, a 12.0mm 2 sample was taken from the discolored zone of wood from one union of Z7. Discolored wood was examined at 400x magnification using an Olympus CX-41 compound microscope (Olympus Inc, Tokyo, Japan). The micrograph was taken using an Olympus DP-72 digital camera connected to the microscope. Olympus Cellsens Standard software was used for image capture. Along with the evident color change of the tissues, darkened parenchyma and gums within the vessels were present (Figure 2.2). These features were not found in the regular sapwood, which showed that physical differences existed between these tissues beyond the change in color. 33

44 Figure 2.2. A transverse section of xylem tissue from the discolored wood of the combination Zestar! / M.7 EMLA with ray parenchyma cells (P) darkened with chemical extractives, and the vessel elements (V) plugged with gums. P V Using ImageJ image analysis software (National Institutes of Health, Bethesda, Maryland) (Rasband, 2014), the total area of wood in each photograph of the longitudinal sections was traced manually. The discolored wood was then traced, and the percentage of discolored wood to the total wood area was recorded. The width of the sections varied, so percentages of the total wood area were taken to normalize these differences. The percentage of discolored wood from each of the segments was averaged into a tree average. Averages were analyzed by an Analysis of Variance (ANOVA) test. ANOVA was performed because although the number of samples between combinations varied, the assumption of homogeneity of variances was still met. Since the percentages met the requirements for normality and homogeneity of variances, no transformations of the data were used. Mean separation was performed by the Games-Howell Method using Minitab 17 (Minitab Inc. State College, PA). The Games-Howell Method was used for mean separation because it accounts for unequal sample sizes and variances by adjusting the error calculations for each comparison based off of the original sample size and variance of each group being compared (Rusticus and Lovato, 2014). 34

45 The percentage of discolored wood was also compared by rootstock cultivar without regard to which scion was budded to it because the discolored wood was found only in rootstock wood tissues. Analysis followed the procedures used for comparing weak and strong scion/rootstock combinations. Results and Discussion The ANOVA showed that the percent of discolored sapwood varied significantly between the scion/rootstock combinations (p-value.05) (Table 2.1) Table 2.1. Percent of discolored wood in the rootstock tissues of eight scion/rootstock combinations of apple. Combination Abbreviation Strength Percent Discolored Honeycrisp H26 Weak 7.49 ab Z / M.26 EMLA Honeycrisp H7 Strong 9.62 a / M.7 EMLA Zestar! Z26 Strong 7.02 abc / M.26 EMLA Zestar! Z7 Strong 9.65 a / M.7 EMLA Cripps Pink P41 Weak 2.98 d / G.41 Cripps Pink P9 Strong 4.67 bcd / M.9 NAKB T337 Scilate S41 Weak 3.73 cd / G.41 Scilate S9 Strong 5.56 abcd / M.9 NIC29 ANOVA p-value <.001 Z Means within columns for percentage of tissue of each combination followed by common letters do not differ at the 5% level, mean separation by Games-Howell test. While significant differences were observed, the weak combination H26 did not differ significantly from the strong combinations H7, Z26, or Z7, nor did any of the weak and strong Cripps Pink or Scilate combinations differ from each other. Additionally, the weak combination H26 had significantly more discolored wood than the weak combinations P41 and S41. These findings suggest that the amount of discolored wood present at the union may not reliably indicate the presence of a weak or 35

46 strong union, and that the weak unions of H26, P41, and S41 may be caused by other factors. While the amount of discolored sapwood appeared to be unrelated to the relative strength of the union, the amount produced by rootstock regardless of scion cultivar differed significantly (p-value.05) (Table 2.2). The lack of a scion influence may suggest that there is little scion/rootstock interaction in the formation of this discoloration. This seems appropriate, as the damaged tissue developed only in the woody tissue of the rootstocks. While it may be coincidental, the amount of discolored wood produced at the union from propagation wounds appeared to be positively correlated to the potential vigor of the rootstock, as the amount of discolored wood produced increased with increased vigor potential. The potential relative vigor of each rootstock has been well documented (Fazio et al., 2005; Robinson et al., 2003; Webster and Wertheim, 2003). Trees on the semi-vigorous rootstock M.7 EMLA produced 2.37% more discolored wood than the semi-dwarfing M.26 EMLA (p-value =.019), 4.96% more than the very dwarfing M.9 NAKB T337 (p-value =.009), and 6.27% more than the very dwarfing G.41 (p-value <.001). M.7 EMLA also produced 4.07% more discolored wood than M.9 NIC29; however this difference was not significant. While it may seem incorrect for M.9 NIC29 to not significantly differ from M.7 EMLA when M.26 EMLA does, this was likely caused by the Games-Howell method, because this method adjusts standard errors based on the individual variances and degrees of freedom of each sample. Since M.9 NIC29 was represented by fewer samples and had a larger variance than M.26 EMLA, the confidence interval for the comparison of M.9 NIC29 to EMLA 7 was much larger. In addition to these differences, EMLA 26 also contained 3.90% more discolored wood than G.41 (p-value <.001). 36

47 Table 2.2. Percent of discolored wood in the rootstock tissues of eight scion/rootstock combinations by rootstock cultivar. Rootstocks Vigor Percent Discolored M.7 EMLA Semi-vigorous 9.63a z M.26 EMLA Semi-dwarfing 7.26b M.9 NIC29 Dwarfing 5.56abc M.9 NAKB T337 Very dwarfing 4.67bc G.41 Very dwarfing 3.36c ANOVA p-value <.001 Z Means within columns for percentage of tissue of each combination followed by common letters do not differ at the 5% level, mean separation by Games-Howell Method. Many factors could have contributed to the differences we observed between the amounts of discolored wood produced between the rootstocks. Since we were unable to determine the presence of organisms within the wood, it is unclear if the discoloration was caused solely by mechanical injury or from microbial decay. Since we are uncertain, we will offer potential explanations for both circumstances. If the discoloration was caused by the wounding injury alone, the very dwarfing rootstocks G.41 may not have produced as much discolored sapwood in their response to the injury, or they may have been able to compartmentalize the injury more effectively. The former explanation seems less likely, because the G.41 rootstock combinations did produce discolored wood after it was injured. However, this discolored wood was confined to a small area near the initial injury, which suggests that G.41 was better able to compartmentalize its injury. Unlike G.41, the discolored areas in M.7 EMLA rootstocks tended to extend further down the length of the rootstocks (Figure 2.3). 37

48 Figure 2.3. The discolored wood in the combinations S41 (A) and H7 (B). In S41 the discolored wood (D) is confined to an area close to the initial wound, showing that it is well compartmented. In H7, the discolored wood extends longitudinally down the rootstock. A B D D The longitudinal spread of discolored sapwood should generally be prevented by the first wall of the CODIT model, which is formed by occlusions within the vessel elements (Shigo and Marx, 1977). The differences in the ability of these walls to compartmentalize the injury may be due to G.41 being able to form stronger vessel occlusions, or they may produce these occlusions faster than M.7 EMLA, which would reduce the amount of time the internal tissues are exposed to the external environment. G.41 may also produce fewer vessel elements, because this would reduce the amount of occlusions necessary to create the first wall of the compartment. Additionally, G.41 rootstocks may also produce more parenchyma cells, because these cells provide the materials that are necessary for creating the vessel occlusions that would prevent the longitudinal spread of discoloration (Bamber, 1987). These features could be evaluated in a microscopic study of the discolored rootstock tissues. If the discoloration was caused by the microbial decay of the wood, the rootstocks with less decay may be inherently more resistant to microbial colonization and subsequent decay. The Geneva series of rootstocks were specifically bred to be disease 38

49 resistant to fire blight [(Erwinia amylovora (Burr)] and Phytopthora root rot (Phytophora cactorum) (Robinson et al. 1999), so it is possible that they also have an increased resistance to wood rotting fungi. Increased disease resistance could come from the parenchyma cells being able to create more chemical extractives to prevent the spread of decay organisms (Schwarze et al. 2000). It is also possible that the xylem of these rootstocks may be better defended against decay by containing a higher percentage of parenchyma cells. Schwarze et al. (2000) explained that trees containing more parenchyma cells are often better equipped to reduce the spread of fungal decay because they can create more extractives. A difference in the percentage in parenchyma cells may also help to explain why the amount of discolored wood produced appeared to be related to rootstock vigor. Komarofski (1947) found evidence that the xylem in the stems of the dwarfing rootstock M.9 contained a higher percentage of parenchyma than more vigorous cultivars, and suggested it was related to their vigor. Additionally, if the rootstocks were better able to compartmentalize wound injuries, they may be better defended against the spread of decay organisms as well. Blanchette (1979) suggested that vessel occlusions usually need to be broken down before decay can progress rapidly through the wood tissues, which may relate to the amount of vessels these rootstocks contain as well. Other differences between our samples not related to rootstock cultivar differences should also be considered. The less vigorous rootstocks may have had smaller diameters when they were budded, which may have possibly decreased the surface area of the wood exposed during budding. This could have reduced the amount of wood being injured, and would also allow a smaller entry point for microorganisms. While not investigated in this study, the method of propagation may also have had an effect on the development of discolored sapwood. Since discolored wood is likely formed by the production of extractives and the oxidation of the xylem by air exposure after wounding (Shigo and Hillis, 1973), it may be less prevalent on bench grafted trees as less surfaces of the wood are exposed once the tissues of the stock and scion have been joined together. This may already be in practical application, as some nurseries now produce brittle combinations only through bench grafting in an effort to produce stronger 39

50 trees (Callahan, 2014). However, since we found little evidence suggesting that brittle combinations produce more discolored sapwood, bench grafted trees may have other benefits that allow them to produce stronger unions. Ecological factors could have also caused differences between our samples. Though all of the trees in this study were nursery grown, the Cripps Pink and Scilate trees were grown in Washington, while the Honeycrisp and Zestar! were grown in Delaware. These regions may have very different disease complexes and climates, which could both greatly influence their likelihood of being invaded by wood-rotting fungi. Our study has found a difference in the amount of discolored wood present in the wood of vigorous and dwarfing rootstocks. While it did not relate to the strength of the unions, this may prove to be a novel method for determining the vigor potential of young graft unions. Future studies of the discolored wood in apple unions should include more tree samples to determine if the trends we observed remain consistent. These studies should also evaluate more combinations of scions and rootstocks to determine if there is a true link between vigor class and the presence of discolored wood. Isolations should also be performed to detect the presence of wood-rotting fungi, as this will make it easier to determine if the discolored wood is caused by microbial decay or from wound discoloration alone. Future studies should also compare the discolored wood of chipbudded and bench grafted trees to determine if propagation method impacts the formation of the discoloration. 40

51 Literature Cited Andrews, P.K., and C. Serrano Marquez Graft incompatibility. Hort. Rev. 15: Baker, C. E. 1933: Water conductivity in the apple tree as affected by pruning and drought. Univ. of Ill., Urbana-Campaign, Ill. PhD Diss. Abstr. 18. Bamber, R.K Sapwood and heartwood. Tec. Pub. No. 2. For. Commission New South Wales. Beecroft, NSW, Australia. Blanchette, R.A A study of progressive stages of discoloration and decay in Malus using scanning electron microscopy. Can. J. For. Res. 9: Callahan, T Adams County Nursery. Personal communication. Déjardin, A., F. Laurans, D. Arnaud, C. Breton, G. Pilate, and J Leplé Wood formation in Angiosperms. Comptes Rendus Biologies. 333: Dilley, M.A. and R.P. Covey Jr Survey of wood decay and associated hymenomycetes in central Washington apple orchards. Plant Dis. 64: Dilley, M.A. and R.P. Covey Jr Coriolus versicolor infection of young apple trees in Washington state. Plant Dis. 65(3):280. Fazio, G., H.S. Aldwinckle, J. Cummins, and T.L. Robinson Geneva 41: a new fire blight resistant, dwarf apple rootstock. Hortscience. 40:1027. Hart, J.H Morphological and chemical differences between sapwood, discolored sapwood, and heartwood in black locust and osage orange. For. Sci. 14(3): Hart, J.H. and K.C. Johnson Production of decay-resistant sapwood in response to injury. Wood Sci. and Tech. 4: Komarofski, B The wood anatomy of certain apple stock and scion varieties and correlative structural changes in the trunk induced by budding. Palestine J. Bot. 6: Manly, N Willow Drive Nursery, Personal communication. McCully, M.E Structural aspects of graft development. p In: R. Moore (Editor). Vegetative Compatibility Responses in Plants. Baylor Univ. Press, Waco, TX. Nair, M.N.B Wood anatomy and heartwood formation in neem. Bot. J. Linnean Soc. 97: Nilsson, M., S. Wikman, and L. Eklund Induction of discolored wood in scots pine (Pinus sylvestris). Tree Phys. 22: Privé, J.P., A. LeBlanc, and C.G. Embree Preliminary evaluation of supported and free standing 'Honeycrisp' trees on 24 apple rootstocks. Acta Hort. 903:

52 Rasband, W.S ImageJ. U. S. National Institutes of Health. Bethesda, Maryland, USA. Robinson, T.L., J.N. Cummins, S.A. Hoying, W.C. Johson, H.S. Aldwinckle, and J.L. Norelli Orchard performance of fire blight-resistant Geneva apple rootstocks. Acta Hort. 489: Robinson, T.L., G. Fazio, T. Holleran, H. Aldwinckle The Geneva series of rootstocks from Cornell: Performance, disease resistance, and commercialization. Acta Hort. 622: Rosenberger, D.A Canker problems in apple orchards. New York Fruit Qrtly. 15:9-12. Rusticus, S.A., and C.Y. Lovato Impact of sample size and variability on the power and type I error rates of equivalence tests: A simulation study. Practical Assessment, Research, and Evaluation. 19(11):1-10. Scheffer, T.C., and E.B. Cowling Natural resistance of wood to microbial deterioration. Annu. Rev. Phytopathol. 4: Schwarze, F.W.M.R., J. Engels, and C. Mattheck Host-fungus interactions: Development and prognosis of wood decay in the sapwood, p In: F.W.M.R. Schwarze, F.W.M.R Wood decay under the microscope. Fung. Bio. Rev. 21: Shigo, A.L The pattern of decays and discolorations in northern hardwoods. Phytopathol. 55: Shigo, A.L. and W.E. Hillis Heartwood, discolored wood, and microorganisms in living trees. Annu. Rev. Phytopathol. 11: Shigo, A.L. and H.G. Marx Compartmentalization of decay in trees. USDA For. Serv. Ag. Info. Bul. No Shortle, W.C. and E.B. Cowling Interaction of live sapwood and fungi commonly found in discolored and decayed wood. Phytopathol. 68: Sucoff, E., H. Ratsch, and D.D. Hook Early development of wound-initiated discoloration in Populous tremuloides michx. Can. J. Bot. 45: Taylor, A.M., B.L. Gartner, and J.J. Morrell Heartwood formation and natural durability- a review. Wood Fiber Sci. 34(4): Walsh, C University of Maryland. Personal communication. Webster, A.D and S.J. Wertheim Apple rootstocks, p In: D.C. Ferree and I.J. Warrington (eds.) Apples: Botany, production and uses. CABI Publishing. Wallingford, UK. 42

53 Winandy, J.E. and R.M. Rowell Chemistry of wood strength. p In: R.M. Rowell (ed.) Handbook of wood chemistry and wood composites. CRC Press. Taylor & Francis Group. Boca Raton, FL. 43

54 Chapter III Anatomical differences between the wood tissues of strong and brittle scion/rootstock combinations of apple Abstract. Anatomical differences were observed in the xylem tissues of three brittle ( Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ G.41, Scilate (Envy )/ G.41 ) and five strong ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA, Cripps Pink cv.maslin/ M.9 NAKB T337, Scilate (Envy )/ M.9 NIC29 ) apple union combinations. Transverse sections of secondary xylem were examined at 400x and 200x magnification. At 400x magnification, the thickness of the fiber cell walls was determined within the rootstock, union, and scion tissues of each tree. Fiber cell wall thickness varied between some of the weak and strong combinations below and at the union. Few differences were observed between the Honeycrisp and Zestar! combinations in the scions, but the Cripps Pink combinations varied there and both combinations of Scilate were thin at the scion. The thickness of the fiber cell walls at each location may be dependent on the overall scion/rootstock combination. At 200x magnification, sections were evaluated for the presence of highly parenchymatous wood, a trait commonly associated with localized incompatibility. The weak scion/rootstock combinations contained higher percentages of parenchymatous tissue than some of the strong combinations. These characteristics of the weak combinations may make them more prone to failure at the union. Introduction The formation of a mechanically weak graft union is a significant problem of some scion/rootstock combinations of apple [Malus xsylvestris (L.) Mill. var. domestica (Borkh.) Mansf.]. These weak unions are prone to breaking during heavy wind storms, which can cause substantial losses to orchards and nurseries. Weak unions have been observed in many popular scion/rootstock combinations, including Honeycrisp / Malling 26 East Malling-Long Ashton ( M.26 EMLA ) (Privé et al., 2011), Gala / Geneva 30 ( G.30 ) (Robinson, et al., 2003), and Cripps Pink cv.maslin/ G.41 (Manly, personal communication). In this study, anatomical traits of chip-budded apple trees were studied to determine if anatomical differences were responsible for the production of weak scion/rootstock combinations. The development of weak wood may be caused by differences in the anatomy of the wood of the trees (Andrews and Serrano Marquez, 1993). Angiosperm woody tissue 44

55 consists of many types of cells that serve specific functions within the tree. Among these are fiber cells, which provide mechanical support to the wood. Fiber cells contain thick secondary cell walls composed of cellulosic microfibrils that are held in place by a matrix of hemicellulose and lignin. These physical properties of the cell wall allow the fiber cells to be strong and flexible (Déjardin et al., 2010), making them the most important for ensuring the mechanical stability of the tree (Winandy and Rowell, 2013). The effects that different rootstocks can have on the thickness of the fiber cell walls of apple have previously been researched. Doley (1974b) found that the walls of fiber cells within the scions of Cox s Orange Pippin were significantly thinner when they were grown on Malling-Merton 104 ( MM.104 ) with an interstock of the very dwarfing Malling 20 ( M.20 ) rootstock. Doley compared this interstock with trees grown on the less dwarfing rootstocks M.8, M.9, and MM.104 with an interstock of MM.104. This study showed scion cell dimensions could be affected by the rootstock/interstock cultivar it is propagated to. Although Doley (1974b) found these differences among cell wall thicknesses, he believed genetic differences between the rootstock cultivars did not contribute much to the variability he observed. Instead, he believed the largest source of variation was caused by the position of the fiber cell within the annual growth ring of the wood. Parenchyma are another type of xylem cell. These cells make up the living component of woody tissue. They are primarily used for the storage of starch and lipids and for transporting these and other assimilates either throughout the xylem (axial parenchyma cells) or between the xylem and phloem (ray parenchyma cells). They are also involved in wound healing within the wood by differentiating into callus tissues (McCully, 1983). Parenchyma cells also have lignified secondary cell walls, but they are generally thinner than the walls of the fiber cells. Vessels are another type of xylem cell. They are hollow and relatively large in diameter compared to the fiber and parenchyma cells. Their primary function is to transport water throughout the tree. Along with the fiber cells, vessels are non-living when they reach maturity (Déjardin et al., 2010). Past research at the East Malling Research Station included anatomical studies of the xylem tissues within the roots of the Malling rootstock series. When scions of the cultivar Cox s Orange Pippin were grafted onto the rootstocks M.7 and M.9, the 45

56 resulting trees produced a higher percentage of fiber cells on the vigorous M.7 (32%) than on the dwarfing M.9 (20%). In addition, the dwarfing rootstock M.9 produced more parenchyma (71% in M.9 and 57% in M.7) (Beakbane and Thompson, 1947). These findings suggested the proportions of xylem cell types could vary among different rootstock cultivars, and that these proportions were related to rootstock vigor. Komarofski (1947) found M.9 rootstocks had the tendency of producing large areas of parenchyma cells in the stem below the union as well. Later work by McKenzie (1961) found the proportions of xylem cells remained relatively constant within the roots of the same cultivar when the rootstock was not grafted to a scion. However, cell proportions were variable when other cultivars were grafted to them. This suggested the scion may also influence the cell characteristics of the rootstock tissues. Other studies also found differences in the proportions of fiber and parenchyma cells between trees that have brittle or strong wood. Aaron and Clarke (1949) observed apple tree limbs that were frequently breaking under heavy crop load. Trees with the most broken limbs contained areas of highly parenchymatous tissue dispersed throughout the wood. Simons (1975) found the strong wood in the cultivar Jonared had more fiber cells in its trunk than Golden Delicious and suggested having a higher percentage of fiber cells might contribute to the strength of the wood of Jonared. While these studies may suggest dwarfing rootstocks and brittle wood both contain higher proportions of parenchyma cells, Webster (2004) did not believe a direct relationship between dwarfing and brittleness exists. Rootstocks like G.30 have helped to support his claim. Although G.30 produces a semi-dwarf tree similar in size to M.7, some cultivars like Gala are more likely to exhibit graft failure on G.30 (Robinson et al., 2003). Incompatibility may play a key role in the formation of some brittle graft combinations (Simons, 1987). Incompatibility has been defined by Andrews and Serrano Marquez (1993) as the failure of a graft combination to form a strong union and to remain healthy due to cellular, physiological intolerance resulting from metabolic, developmental, and/or anatomical differences. Incompatibility can generally be split into two types, localized and translocated (Mosse, 1962). These two forms differ in 46

57 many aspects, including the likelihood of producing unions with the potential to break at the union. Graft failure is commonly attributed to localized incompatibility. Incompatibility is considered localized when the direct cellular contact of the rootstock and scion leads to the incompatibility. This type of incompatibility can be determined when a mutually compatible interstock is grafted between the rootstock and scion. If the incompatibility is localized, the inclusion of an interstock should produce a compatible tree since the tissues of the rootstock and scion are no longer in direct contact. This differs from translocated incompatibility, because the inclusion of an interstock in that type still produces an incompatible tree. Localized incompatibility also commonly leads to a discontinuity of the vascular system at the graft union. In translocated incompatibility the xylem differentiates normally, but the phloem at the union degenerates. Localized incompatibility often results in a slow decline of the tree until the union suddenly breaks, while symptoms of translocated incompatibility often develop quickly due to the degeneration of the phloem. While these descriptions are useful, these terms are not absolute, and tissues of the same tree can exhibit symptoms of both types of incompatibility in varying degrees (Mosse, 1962). Delayed incompatibility is another type that has also been described (Moore, 1983). However, in their review of incompatibility Andrews and Serrano Marquez (1993) believed this term should not be used. They suggested while the physical symptoms of incompatibility may not be immediately apparent, the factors causing the incompatibility are likely to have long been present within the tree. They continued to state how many cases of delayed incompatibility have since been discovered to be related to disease. Additionally, they explained cases of delayed incompatibility can be classified under the categories of translocated or localized incompatibility, making the third definition unnecessary. In locally incompatible scion/rootstock combinations, fibers and vessels are replaced by incompletely lignified ray cells (Mosse, 1962). Parenchyma in the form of callus tissues are essential for bridging the initial lack of continuity between the rootstock and scion (McCully, 1983), however their proliferation should soon give way to differentiated xylem cells that can better support the weight of the tree. If differentiation 47

58 is further delayed and parenchyma proliferation continues, the lack of fiber and vessel connections between the rootstock and scion may ultimately lead to the failure of the graft union. In addition to parenchyma cells being weaker than fibers, the parenchyma cells found in incompatible unions also tend to be poorly organized. This may cause subsequent vascular tissue to develop in irregular orientations, further weakening the graft union (Mosse, 1962). While new growth may lack organization immediately after grafting, proper xylem orientation is generally achieved within a year or two after grafting in compatible scion/rootstock combinations (Simons and Chu, 1985; Soumelidou et al., 1994). The parenchymatous tissue described by Mosse (1962) has been observed in past microscopic studies of apple unions. Callus parenchyma tissue and ray parenchyma cells were present in the unions of Sturdeespur Delicious / M.26 (Simons and Chu, 1985). The tissue was described as being poorly organized, which led to the disorganization of the subsequent vascular tissue. Ussahatanonta and Simons (1988) also observed abnormally oriented ray parenchyma cells in unions of Golden Delicious / M.9. They noted these trees were relatively weaker than those produced on MM.106, M.7A, and M.26 rootstocks. Soumelidou et al. (1994) found two-year-old bud unions of the weak combination Bramley s Seedling / M.9 had rings of parenchyma tissue followed by bands of thin-walled fiber cells (cell thickness was not quantified, nor did they distinguish whether the parenchyma were callus parenchyma or ray cells). They observed irregular distributions of parenchyma and thin fiber cells less frequently when Bramley s Seedling was grown on the more vigorous rootstock MM.106. Warmund et al. (1993) found parenchyma tissue present in the unions of Jonagold / Mark combinations. Though they did not describe what kind of parenchyma tissue was present, they noted trees grown in the field tended to break along the line of this tissue. Large areas of parenchyma have been found in the unions of other weak fruit tree combinations, and incompatibility was also suggested as the likely cause. Herrero (1951) documented the proportions of different cell types within the scions of combinations of peach (Prunus persica L. Batch cv. Hale s Early ) on compatible plum (Prunus domestica L. cv. Brompton ) and incompatible cherry plum 48

59 (Prunus cerasifera Ehrh. cv. Myrobalan B ) rootstocks. He had split the tissue area into four cell categories, including vessels, fibers, wood parenchyma, and ray parenchyma. He found the compatible combinations contained 7.2% more fiber cells in the scion and 6.8% more in the rootstock than the incompatible combinations. The compatible combinations also contained 5.3% less ray cells in the scion and 1.4% less in the rootstock. These differences were not analyzed statistically. Errea et al. (1994) found poorly differentiated callus tissues between the incompatible bud union combinations of apricot (Prunus armeniaca L. cvs. Luizet and Moniqui ) on the cherry plum rootstock Mirobolán. They believed these would disrupt the development of the cambium, which could then lead to an overproduction of ray parenchyma. Ermel et al. (1997) found the cambia of the incompatible pear (Pyrus communis L.)/quince (Cydonia oblonga Mill.) combinations Epine du Mas clone P 1885 / Old Home Farmingdale 333 clone P 2420 had to develop around callus tissue, while the cambia of compatible combinations joined in a straight line. Zarrouk et al. (2006) described areas of unlignified tissues within the unions of many incompatible combinations of peach and nectarine scions Catherina, Tebana, and Summergrand scions on damson plum (Prunus insititia L. cv. PM 95 AD ) rootstocks. A later study by Zarrouk et al. (2010) found that five-month-old incompatible peach and cherry plum combinations lacked cambial connections, while they were already established in compatible combinations. The objective of this research was to use light microscopy to observe and quantify differences in the wood anatomy of known brittle ( Honeycrisp / M.26 EMLA, Cripps Pink cv.maslin/ G41, Scilate (Envy )/ G41 ) and strong ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA, Cripps Pink cv.maslin/ M.9 NAKB T337, Scilate (Envy )/ M.9 NIC29 ) scion/rootstock combinations of apple to discover the possible causes of their union failure, and to determine if these differences could be used to predict future weak combinations. Since fiber cells derive their strength from their lignified cell walls, we observed the fiber cell wall thicknesses within the rootstocks, unions, and scions. We predicted that strong combinations would have thicker fiber cell walls at all three tissue locations. We also investigated the potential 49

60 roles that the scions and rootstocks have on influencing the cell wall thickness of the fiber cells at each location. In addition to the fiber cell wall study, we also determined the percentages of xylem cell tissues that developed in the subsequent year s growth of the wood to determine if the weak scion/rootstock combinations had more parenchymatous tissue. More parenchyma tissue may suggest an incompatibility between the budding partners, as this anatomical feature has previously been associated with incompatibility and the failure of the graft union (Andrews and Serrano Marquez, 1993; Mosse, 1962). Since Herrero (1951) had observed differences in the relative proportions of cell types between compatible and incompatible combinations of peach and cherry plum, we expected that this may be a useful tool for determining incompatibility in apple and warranted further testing. Materials and Methods Experiment 3.1. Determination of fiber cell wall thickness. This experiment consisted of 48 trees, with six replications of eight different scion/rootstock combinations. In Feb. 2014, finished chip-budded trees were received from Willow Drive Nursery, Ephrata, WA. Trees included Cripps Pink cv.maslin on the rootstocks G.41 and M.9 NAKB T337 (abbreviated as P41 and P9, respectively) and Scilate (Envy ) on the rootstocks G.41 and M.9 NIC29 ( S41 and S9 ). In Apr. 2014, additional finished chip-budded trees were received from Adams County Nursery, Aspers, PA. These included combinations of Honeycrisp and Zestar! on the rootstocks M.26 EMLA and M.7 EMLA ( H26, H7, Z26, and Z7 ). All trees were budded in the All trees were kept at 6 C until sampling. After being cut into 10.0cm long sections by a circular saw, and 4.0mm thick blocks by a band saw, the samples were placed in water for three to seven days to soften the wood tissue for hand sectioning. Two sets of trees were kept in 70% ethanol for 38 and 27 days before being moved into water for five and six days, respectively. These trees had originally been intended for another study. After softening, samples were then cut transversely into three blocks representing different heights of the tree: 7.0cm below the union (rootstock), at the union, and 3.0cm above the union (scion). The phloem 50

61 tissue was removed from the outer edge of these blocks to facilitate hand sectioning of the xylem. The three blocks from each tree were then hand sectioned using single-edged razor blades. Four sections approximately 12.0mm 2 were removed from each of the three blocks of the tree and were placed in two drops of distilled water on glass microscope slides. Sections were then stained with 1% toluidine blue for one minute. Samples were then rinsed with water and cover slips were applied. Sections were examined at 400x magnification with an Olympus CX-41 compound microscope (Olympus Inc., Tokyo, Japan). Photomicrographs were taken using an Olympus DP-72 digital camera connected to the microscope. Olympus Cellsens Standard software was used for image capture and data gathering. Though four sections were taken from each block, only the clearest section was used for data collection. Fiber cell wall thickness was measured from radial fiber cell walls. Walls were measured from the middle lamella to the lumen of the cell. Fifty fiber cells were counted from each of the three areas of the tree (below, at, and above the union) 150 cells in total were sampled per tree. The fifty fiber cell wall measurements were averaged for each block of the tree. Averages were then transformed to control for unequal variances using the Box-Cox transformation function on Minitab 17 (Minitab Inc., State College, PA). Transformed data were analyzed by Analysis of Variance (ANOVA) tests and mean comparisons were performed using Fisher s Least Significant Difference (LSD) tests on Minitab 17. The original averages of the raw data from the blocks are reported. Interactions between the rootstocks and scions of the Honeycrisp and Zestar! combinations were evaluated graphically on Microsoft Excel (Microsoft Corp. Redmond, WA). This analysis was not performed on the Cripps Pink and Scilate combinations because different clones of M.9 rootstocks were used. Experiment 3.2. Proportions of xylem cell types in Honeycrisp and Zestar!. In this study, six trees of the four combinations of the Honeycrisp and Zestar! combinations described in experiment 3.1 were also used (24 in total). After trees were sectioned and stained as previously described, an additional photomicrograph of every 51

62 tree was taken from the union at 200x magnification using the same microscope/camera/software system previously described in experiment 3.1. The xylem cells were divided into three tissue types based on their function within the wood: fibrous tissue, parenchymatous tissue, and conductive tissue. Fibertracheids and libriform fibers serve similar functions by providing mechanical support to the wood (Carlquist, 2001). The parenchymatous tissue included ray, axial, and callus parenchyma, as they are used for storage and transport of metabolic materials and in wound repair. The conductive tissue included vessel elements, which transport water throughout the tree (Déjardin et al., 2010). Percentages of the three types of tissue were determined using ImageJ image analysis software (National Institutes of Health, Bethesda, Maryland) (Rasband, 2014). The parenchymatous and conductive cells were traced manually, while fibrous cells were estimated by subtracting the two former measurements from the total area of the photomicrograph. Percentages were not transformed because the assumptions of normality and equal variances were met. Percentages were analyzed by ANOVA and mean separation was performed by Fisher s LSD test using Minitab 17. Experiment 3.3. Proportions of xylem cell types in Cripps Pink and Scilate. For this experiment, nine trees were used. Tree samples were obtained from the Cripps Pink and Scilate combinations described in experiment 3.1. Three replications of S41 and P41, two replications of P9, and one sample of S9 were used. Sampling procedures followed those previously described in experiment 3.2. The mean percentage of each tissue type of the combinations was recorded. Further analysis was not conducted because of power limitations associated with low sample numbers. Results and Discussion Experiment 3.1. ANOVA showed that fiber cell wall thickness differed significantly (p-value.05) between graft combinations below, at, and above the unions. Significant differences were observed between some of the strong and weak scion/rootstock combinations of the same cultivar (Table 3.1). 52

63 Table 3.1. Mean fiber cell wall thicknesses (µm) 7.0cm below, at, and 3.0cm above the unions of eight scion/rootstock combinations of apple. Graft Combination Abbreviation Strength 7cm Below At Union 3cm Above Honeycrisp H26 Weak 3.47 bc z 3.72 bc 3.95 a / M.26 EMLA Honeycrisp H7 Strong 3.91 a 4.10 a 3.98 a / M.7 EMLA Zestar! Z26 Strong 3.52 bc 3.59 bcd 3.79 a / M.26 EMLA Zestar! Z7 Strong 3.70 ab 3.84 ab 3.78 a / M.7 EMLA Cripps Pink P41 Weak 3.21 c 3.38 cd 3.33 b / G.41 Cripps Pink P9 Strong 3.73 ab 3.71 bc 4.04 a / M.9 NAKB T337 Scilate S41 Weak 3.40 bc 3.31 d 3.33 b / G.41 Scilate S9 Strong 3.90 a 3.45 cd 3.34 b / M.9 NIC29 ANOVA p-value < <.001 z Means within columns followed by common letters do not significantly differ at the.05 level. Mean separation by Fisher s LSD p Below the unions, the fiber cell walls of the weak combination P41 were significantly thinner than those of P9 (p-value =.003), the walls of the weak combination S41 were significantly thinner than S9 (p-value =.002), and the walls of H26 were thinner than H7 (p-value =.002). These findings support our hypothesis, as we had expected to find thinner cell walls in weak combinations. However, while P41 cell walls were thinner than S9 (p-value <.001), neither S41 nor H26 differed from P9. Additionally, the walls of H26 and S41 did not differ from the walls of Z26 or Z7, nor did the walls of P41 differ from Z26. At the graft union, the cell walls of H26 were also significantly thinner than those of H7 (p-value =.027). However, in contrast to our hypothesis the walls of H26 were slightly thicker at the union than the strong combinations S9, P9, and Z26. The walls of P41 did not differ from Z26, P9, or S9, nor did the cell walls of S41 differ from Z26 or S9. This suggests that cell wall thickness may not be a reliable indicator of union strength. 53

64 Few differences were observed between Honeycrisp and Zestar! combinations within the scion tissues. These observations suggest the overall scion/rootstock combination may not have much effect on fiber cell wall thickness above the union in these combinations. However, in the Cripps Pink combinations, fiber cell walls within the scion tissues were significantly thinner in P41 compared to P9 (p-value <.001). These results suggest that the G.41 rootstock may have reduced fiber cell wall thickness of the Cripps Pink scions, similar to the M.20 interstock in the study by Doley (1974b). With the exception of both of the Scilate combinations, fiber cell walls tended to be thinnest below the union and thickest above. The observation that Scilate produced thinner cell walls at the union and scion regardless of rootstock differences may suggest that the fiber cell walls of Scilate may be inherently thinner than the other scion cultivars, making it a particularly weak cultivar. Overall, fiber cell wall thicknesses throughout the length of the tree appeared to depend on interactions between the rootstock cultivar, the scion cultivar, and the fiber cell s location within the tree. However, few clear trends were observed amongst these factors throughout the different combinations, suggesting that all of these factors interact in differently between combinations. To further explain these differences, another way of interpreting the data from the Honeycrisp and Zestar! combinations is to examine possible interactions between the rootstocks and scions from the three different regions. Since two different clones of M.9 were used, interactions could not be evaluated for Cripps Pink and Scilate. Above the union Honeycrisp produced slightly thicker cell walls than Zestar! regardless of which rootstock it was budded to (Figure 3.1). A similar trend was observed at the union (Figure 3.2). This trend was also observed in trees on M.7 EMLA below the union, as H7 produced thicker fiber cell walls than Z7. However, in the M.26 EMLA combinations the cell wall thickness varied very little between H26 and Z26 (Figure 3.3). The statistical interaction between these factors suggests that cell wall thickness of fiber cells below the union may depend on the scion/rootstock combination rather than rootstock alone. 54

65 Wall Thickness (µm) Wall Thickness (µm) Figure 3.1. The effect of cultivar and rootstock on fiber cell wall thickness 3.0cm above the union for Honeycrisp and Zestar! EMLA 26 EMLA Honeycrisp Zestar! Figure 3.2. The effect of cultivar and rootstock on fiber cell wall thickness at the union for Honeycrisp and Zestar! EMLA 26 EMLA Honeycrisp Zestar! 55

66 Wall Thickness (µm) Figure 3.3. The effect of cultivar and rootstock on fiber cell wall thickness 7.0cm below the union for Honeycrisp and Zestar! EMLA 26 EMLA Honeycrisp Zestar! Overall, only some of the weak and strong combinations differed within the rootstock and the union, and no differences were observed in the scions of the Honeycrisp, Zestar!, and Scilate combinations. These observations contradict our hypothesis, and suggest that other factors may be more important in predicting graft failure in some scion/rootstock combinations. Additional variation between our measurements unrelated to rootstock cultivar differences could have been introduced in a few ways. Doley (1974a) described how cell position within the growth ring provided the greatest single source of variation within his study of cell wall thicknesses. In our study, cell walls were measured from the edge of the growth ring to control for this variation. However, since we did not use a formal measure of growth ring position, differences in ring position may account for some variation within our samples. Additionally, sampling method can be problematic when measuring fiber cells (Carlquist, 2001; Doley, 1974a). Since fibers are tapered at their ends, it would not be sufficient to sample the cells at random, as doing so could lead to sampling cells near their ends where they may have very different dimensions. The ideal 56

67 way to measure fiber cell dimensions may be to individually isolate whole cells (Doley, 1974a). For this study cells were sampled by selectively choosing cells that appeared larger, as cells that were much smaller were likely to be the ends of the cells. While this measuring strategy is not ideal, it is likely to have provided a more accurate representation of the true average than what would have been achieved by sampling at random. Additionally, our sampling method may have given consistent results, as wall thicknesses were very similar above the union between most of the same cultivars of scion. Future work should be conducted to further examine the cell wall characteristics of weak scion/rootstock combinations to confirm the results we have reported. Though difficult to prepare, the measuring of individual fiber cells would provide the most accurate results and other cell dimensions, such as length, could be quantified. The addition of these other cell characteristics would allow for a more complete investigation of these tissues. Experiment 3.2 and Experiment 3.3. Qualitative differences in the distribution of tissue types were observed between weak and strong scion/rootstock combinations of Honeycrisp and Zestar! (Figure 3.4). Xylem of the weak combination H26 appeared disordered with large areas of parenchymatous tissue. This appeared to be accompanied by a reduction in the amount of fibrous and conductive tissues. The xylem tissues of P41 and S41 appeared to have more parenchyma cells than strong combinations, but appeared well-ordered compared to H26. Percentages of fibrous and parenchymatous wood tissues varied significantly (pvalue.05) between some of the weak and strong scion/rootstock combinations of Honeycrisp and Zestar! (Table 3.2). H26 contained 24.83% more parenchyma tissue than Z7 and 14.64% more than Z26. Mean comparisons by Fisher s LSD showed that these differences were statistically significant (p-values = <.001 and =.018, respectively). While H26 did not have significantly more parenchymatous tissue than H7, the mean for H26 was 9.84% greater than H7. Future studies with larger sample sizes may allow us to determine if a significant difference in the amount of parenchymatous tissue 57

68 truly exists between these two combinations. Additionally, although H7 is not prone to producing a brittle union, it also contained 14.98% more parenchymatous tissue than Z7 (p-value =.016). The increased production of parenchymatous tissue in H26 may result from the decreased differentiation of fibrous and conductive tissues. H26 contained 19.57% less fibrous tissue than Z7 (p-value =.002) and contained slightly less fibrous tissue than H7 and Z26 (7.49% and 9.79%, respectively). This decrease in the amount of fibrous tissues may contribute to the decreased strength of H26 unions. Though H26 did not have significantly less fibrous tissue than H7 or Z26, future studies with larger sample sizes may help to determine if the differences observed between these combinations are significant. Similar to the parenchymatous tissues, Z7 also contained 12.09% more fibrous tissue than H7 (p-value =.043). Though also not significantly different, H26 contained 4.85% less conductive tissue than Z26 and 5.26% less than Z7. If less conductive tissue is also present in the phloem, a reduction in the amount of sieve tube elements may help to explain why Honeycrisp is prone to zonal leaf chlorosis (Chen and Cheng, 2004), as the lack of conductive tissue could reduce the transport of starch from the leaves to the carbon sinks. Though not analyzed statistically, xylem tissue proportions also appeared to vary between Cripps Pink and Scilate combinations (Table 3.3). The weak combination P41 contained 11.38% more parenchymatous and 11.47% less fibrous tissue than P9, and 13.58% more parenchymatous and16.28% less fibrous tissue than S9. S41 contained 8.88% more parenchymatous tissue and 15.85% less fibrous tissue than S9, and 6.68% more parenchymatous and 11.04% less fibrous tissue than P9. Compared to the Honeycrisp and Zestar! combinations, P41 contained 1.26% more fibrous tissue and 6.63% less parenchymatous tissue than H26, and S41 contained 1.69% more fibrous tissue and 11.33% less parenchymatous tissue than H26. H26 also contained more parenchymatous and less fibrous tissue than both of the M.9 combinations, even though M.9 tends to produce a more dwarfing tree than M.26. This reinforces the belief that graft failure is not directly related to rootstock vigor (Webster, 2004). Contrary to our hypothesis, P41 and S41 showed relatively few differences from the proportions of the strong combination H7. 58

69 Figure 3.4. Xylem tissues at the unions of H26 (A) and H7 (B). The xylem of H26 consists of many parenchyma cells (P), scattered areas of fiber cells, (F) and some vessels elements (V). The ray parenchyma (RP) cells are poorly defined. The tissue of H7 appears more uniform with many fiber cells. Vessel elements are large, and there are less parenchyma cells. The ray parenchyma cells also appear well organized into thin rays. A B 59

70 Table 3.2 Percentages of wood tissues by combination in the unions of four scion/rootstock combinations of the apple scion cultivars Honeycrisp and Zestar! on rootstocks M.26 EMLA and M.7 EMLA. More parenchymatous tissue and less fibrous tissue were observed in the weak combination H26. Graft Combination Abbreviation Parenchymatous Fibrous Conductive Honeycrisp H a z 46.08b 6.80 / M.26 EMLA Honeycrisp H ab 53.57b 9.16 / M.7 EMLA Zestar! Z bc 55.87ab / M.26 EMLA Zestar! Z c 65.65a / M.7 EMLA ANOVA p-value z Mean separation within columns followed by common letters do not significantly differ at the.05 level. Mean separation by Fisher s LSD p Table 3.3 Percentages of wood tissues by combination at the unions of four scion/rootstock combinations of Cripps Pink and Scilate on the rootstocks G.41, M.9 NIC29 and M.9 NAKB T337. More parenchymatous tissue was observed in the weak combinations P41 and S41. Graft Combination Abbreviations Parenchymatous Fibrous Conductive Cripps Pink / G.41 Cripps Pink / M.9 NAKB T337 Scilate / G41 Scilate / M.9 NIC29 P P S S Our findings from experiment 3.2 suggest the weak unions of H26 may be caused by a localized incompatibility between the scion/rootstock combinations. This belief is supported by the increased amount of poorly differentiated parenchymatous tissues at their unions, which has previously been reported as a key feature of localized incompatibility (Mosse, 1962). Much like the work by Simons and Chu (1985) and Ussahatanonta and Simons (1988), we found that H26 appeared to have areas of poorly 60

71 differentiated parenchyma tissues a year after propagation, while the strong unions appeared to form well-differentiated xylem tissues. Though not statistically significant from all the strong combinations in this study, the decrease in the differentiation of fibrous and conductive tissues in the weak combination may lead to the characteristic smooth break commonly associated with union failure (Andrews and Serrano Marquez, 1993; Mosse, 1962). The weak unions of S41 and P41 may also be due to incompatibility, though the tissues appeared more ordered and had less parenchymatous and more fibrous tissue than H26. These results suggest their differences may be caused from being a very dwarfing cultivar. Komarofksi (1947) and McKenzie (1961) found xylem cell proportions varied by rootstock vigor within the stems and roots, and this may explain the increased production of parenchyma in the G.41 samples. However, because G.41 tends to break more readily when certain scion cultivars are grown on it, the cause of union failure in the Cripps Pink and Scilate combinations may be related to incompatibility or another scion/rootstock interaction affecting the production of xylem tissue, and requires further testing. Our findings are somewhat in agreement with the work done by Herrero (1951), as incompatible combinations of peach/cherry plum also contained more parenchyma and less fiber cells. However, the peach/cherry plum combinations also exhibited other external signs of incompatibility like reduced growth, which suggests they may have suffered from translocated incompatibility. He also observed plum/cherry plum and pear/quince combinations that may have been suffering from localized incompatibility, but found few differences in their proportions of parenchyma and fiber cells. Many possible explanations may help to account for these discrepancies. Herrero (1951) made his peach/cherry plum combination measurements in the scion tissue, rather than from the tissues located at the union. The pear/quince and plum/cherry plum combinations in his study were planted on interstocks, and his measurements were taken from tissues in the scion, interstock, rootstock, and the roots, rather than from directly at the union. These differences in sampling method could help to explain the differences between our results. 61

72 Since Honeycrisp / M.26 EMLA is prone to graft failure, this combination may primarily suffer from localized incompatibility. However, it is possible that these trees may suffer from both forms of incompatibility, as Mosse (1962) had said it was possible to find both forms within the same tree. Honeycrisp is known to exhibit many growth problems, including zonal leaf chlorosis (Chen and Cheng, 2004). It is possible that these may be caused by a translocated incompatibility. Since Honeycrisp develops these problems on many rootstocks they may be cultivar-related problems, or Honeycrisp may have a translocated incompatibility with many rootstocks and both forms of incompatibility with M.26 EMLA. As Doley (1974a) described fiber cell walls, Mosse (1962) described how the anatomical features of graft unions can be very variable, even within the same union. This may help to explain why relatively few differences in tissue proportions were observed in the G.41 combinations compared to H26, even though their external symptoms suggest a localized incompatibility. Methods that allow for the investigation of the tissues of the entire union would help to greatly further our understanding of the development of the vascular tissues at the scion/rootstock interface. Results from such a test may allow rootstock breeders to develop a numeric threshold for the percentage of parenchymatous tissue where graft failure is most likely to occur. This numeric threshold would potentially allow for future brittle scion/rootstock combinations to be screened out before being planted out in orchards, or allow for specific training guidelines for the weaker combinations to prevent their unions from breaking in the field. Though we may have observed differences in the anatomy of young scion/rootstock combinations related to localized incompatibility in this study, vascular discontinuity may sometimes not occur until many years following propagation (Mosse, 1962). Since the exact mechanisms of incompatibility are still unknown in apple, genetic and biomolecular studies may ultimately provide the means for the early detection of incompatibility. 62

73 Literature Cited Aaron, I. and W.S. Clarke Jr Breakage of apple trees. Proc. Amer. Soc. Hort. Sci. 54: Andrews, P.K., and C. Serrano Marquez Graft incompatibility. Hort. Rev. 15: Beakbane, A.B. and E.C. Thompson Anatomical studies of stems and roots of hardy fruit trees. IV. The root structure of some new clonal apple rootstocks budded with Cox's Orange Pippin. J. Pomol. 23: Carlquist, S Imperforate tracheary elements. p In: S. Carlquist (Editor). Comparative wood anatomy: systematic, ecological, and evolutionary aspects of dicotyledon wood. Springer-Verlag, Berlin, Germany. Chen, L.S. and L. Cheng CO 2 assimilation, carbohydrate metabolism, xanthophyll cycle, and the antioxidant system of Honeycrisp apple leaves with zonal chlorosis. J. Amer. Soc. Hort. Sci. 129: Déjardin, A., F. Laurans, D. Arnaud, C. Breton, G. Pilate, and J Leplé Wood formation in Angiosperms. Comptes Rendus Biologies. 333: Doley, D. 1974a. Alternatives to the assessment of earlywood and latewood in dicotyledonous trees: A study of structural variation in growth rings of apple (Malus Pumila Mill.). New Phytol. 73: Doley, D. 1974b. Effects of rootstocks and interstock on cell dimensions in scion stems of apple (Malus pumila Mill.) New Phytol. 73: Ermel, F.F., J.L. Poëssel, M. Faurobert, and A.M. Catesson Early scion/stock junction in compatible and incompatible pear/pear and pear/quince grafts: a histocytological study. Ann. Bot. 79: Errea, P., A. Felipe, and M. Herrero Graft establishment between compatible and incompatible Prunus spp.. J. Exp. Bot. 45: Herrero, J Studies of compatible and incompatible graft combinations with specific reference to hardy fruit trees. J. Hort. Sci. 26: Komarofski, B The wood anatomy of certain apple stock and scion varieties and correlative structural changes in the trunk induced by budding. Palestine J. Bot. 6: Manly, N Willow Drive Nursery, Personal communication. McCully, M.E Structural aspects of graft development. p In: R. Moore (ed.). Vegetative compatibility responses in plants. Baylor Univ. Press, Waco, TX. McKenzie, D.W Rootstock-scion interaction in apples with special reference to root anatomy. J. Hort. Sci. 36:

74 Moore, R Physiological aspects of graft formation, p In: R. Moore (ed.). Vegetative compatibility responses in plants. Baylor Univ. Press, Waco, TX. Mosse, B Graft incompatibility in fruit trees: with particular references to its underlying causes. Tech. Comm. 28. Comm. Bur. Hort. Plant Crops, East Malling, UK. Privé, J.P., A. LeBlanc, and C.G. Embree Preliminary evaluation of supported and free standing 'Honeycrisp' trees on 24 apple rootstocks. Acta Hort. 903: Rasband, W.S ImageJ. U. S. National Institutes of Health. Bethesda, Maryland, USA. Robinson, T.L., G. Fazio, T. Holleran, H. Aldwinckle The Geneva series of rootstocks from Cornell: Performance, disease resistance, and commercialization. Acta Hort. 622: Simons, R.K Variation of wood tissues from two cultivars of apples. Fruit Var. J. 29 (2): Simons, R.K. and M.C. Chu Graft union characteristics of M.26 apple rootstock combined with 'Red Delicious' strains-morphological and anatomical development. Scientia Hort. 25: Simons, R.K Compatibility and stock-scion interactions as related to dwarfing, p In: R.C. Rom and R.F. Carlson (eds.) Rootstocks for fruit crops. John Wily and Sons. New York, NY. Soumelidou, K., N.H. Battey, P. John, and J.R. Barnett The anatomy of the developing bud union and its relationship to dwarfing in apple. Ann. Bot. 74: Ussahatanonta, S. and R.K. Simons Graft union development of the Golden Delicious apple when combined with various dwarfing rootstocks. Fruit Var. J. 42: Webster, A.D Vigour mechanisms in dwarfing rootstocks for temperate fruit trees. Acta Hort. 658: Warmund, M.R., B.H. Barritt, J.M. Brown, K.L. Schaffer, and B.R. Jeong Detection of vascular discontinuity in bud unions of Jonagold apple on mark rootstock with magnetic resonance imaging. J. Amer. Soc. Hort. Sci. 118: Winandy, J.E. and R.M. Rowell Chemistry of wood strength. p In: R.M. Rowell (ed.) Handbook of wood chemistry and wood composites. CRC Press. Taylor & Francis Group. Boca Raton, FL. Zarrouk, O., Y. Gogorcena, and M.A. Moreno Graft compatibility between peach cultivars and Prunus rootstocks. HortScience. 41:

75 Zarrouk, O., P.S. Testillano, M.C. Risueño, M.A. Moreno, and Y. Gogorcena Changes in cell/tissue organization and peroxidase activity as markers for early detection of graft incompatibility in peach/plum combinations. J. Amer. Soc. Hort. Sci. 135:

76 Chapter IV Potential methods for the rapid determination of localized incompatibility between apple scion/rootstock combinations Abstract. Brittle unions are a common problem of certain scion/rootstock combinations of apple. These include commercially important cultivars like Honeycrisp and Cripps Pink. Brittle unions are often associated with a localized incompatibility between the rootstock and scion, which is marked by a proliferation of ray parenchyma tissue where the rootstock and scion meet. This characteristic can be observed using microscopy, however woody tissue is difficult to section and provides a limited view of the overall tree histology. In this study, we evaluated two methods, Laser Ablation Tomography and an iodine starch test, to determine their potential use for the rapid detection of parenchyma tissue within the union. The results from the LAT experiment showed that qualitative differences between graft combinations could be observed. These differences included the presence of poorly differentiated regions of parenchyma and irregular orientations of the vascular tissues within the weak combinations. However, because of the system s inability to ablate woody samples of more than 4.0mm thick or reach a level of magnification sufficient to differentiate tissues at the cellular level while maintaining a wide field of view, the LAT process may not be well suited to rapidly quantify whole-union histological differences at this time. The results of the starch test were inconclusive. While some weak combinations showed a large amount of staining at the union, the percent of wood stained was extremely variable between samples of the same combination. This may have been caused by variable camera settings and from some of our sample trees being discolored from prolonged storage in ethanol. Additionally, starch concentration may vary by the time of year, which may further prevent its use as a test for localized incompatibility. Introduction The formation of a mechanically weak graft union is a significant problem of some scion/rootstock combinations of apple [Malus xsylvestris (L.) Mill. var. domestica (Borkh.) Mansf.]. These weak unions are prone to breaking during heavy wind storms, which can cause substantial losses to orchards and nurseries. Weak unions have been observed in many popular scion/rootstock combinations, including Honeycrisp / Malling 26 East Malling-Long Ashton ( M.26 EMLA ) (Privé et al., 2011), Gala / Geneva 30 ( G.30 ) (Robinson, et al., 2003), and Cripps Pink cv.maslin/ G.41 (Manly, personal communication). Many factors can lead to graft union failure. These include environmental conditions such as wind and cold stress, disease, and poor propagation techniques. While 66

77 all of these factors can contribute to union failure, brittle unions are commonly associated with an incompatibility between the scion and rootstock cultivars. Graft failure caused by incompatibility usually results in a characteristic clean break at the union because of a lack of vascular cohesion between the tissues of the graft partners (Andrews and Serrano Marquez, 1993). In addition to graft failure, other physical symptoms have been associated with incompatibility. Swelling of the union is sometimes observed, but this may not be a reliable indicator. The rootstock M.9 is prone to swelling at the union; however this is likely to be an inherent feature of the rootstock cultivar that is unlikely to be caused by incompatibility (Mosse, 1962). A difference in the diameter of the rootstock and scion has also been attributed to incompatibility, but much like swelling this feature is considered to be an unreliable indicator. Andrews and Serrano Marquez (1993) explained that some rootstocks and scions have inherently different growth rates leading to an overgrowth of one over the other, but many of these still go on to form compatible unions. They also discussed how other symptoms, like the reduced growth and premature death of the tree, have also been suggested as symptoms. However, they argued that these can be caused by a number of abnormalities unrelated to incompatibility, and suggested that graft failure is the most reliable symptom of localized incompatibility. Incompatibility can generally be classified into two forms: translocated and localized (Mosse, 1962). These forms differ in a number of characteristics, including their likelihood to produce a brittle union. The most distinguishing characteristic of translocated incompatibility is that the factor causing the incompatibility is translocated from one graft partner to the other. This factor is transferred even if the rootstock and scion tissues are not in direct contact, because the resulting tree remains incompatible even when a mutually compatible interstock is grafted between the two. The translocated factor is also directional, as reciprocal grafts are compatible. In addition to having a translocated factor, the transport of carbohydrates from the scion to the rootstock across the union is hindered in this type of incompatibility, as starch often accumulates directly above the union of these combinations. While the xylem tissues remain continuous at the union, the phloem often degenerates, and this is likely related to the reduced transport of 67

78 carbohydrates across the union. In this type of incompatibility, the growth of the tree quickly declines because the roots are starved (Andrews and Serrano Marquez, 1993). In localized incompatibility, the direct tissue contact of the rootstock and scion causes vascular discontinuity within the xylem at the union. Unlike translocated incompatibility, a compatible tree forms when an interstock is inserted between the rootstock and scion of locally incompatible unions. However, reciprocal grafts remain incompatible, suggesting that the incompatibility is not directional (Andrews and Serrano Marquez, 1993; Mosse, 1962). Due to the lack of xylem continuity at the union, this type of incompatibility usually leads to graft union failure (Mosse, 1962). At some point during the development of the union, the cambia of the stock and scion become discontinuous. This results in a lack of normal xylem differentiation, as poorly lignified ray parenchyma cells proliferate in the region in place of vessel elements and fiber cells. This shift from normal differentiation can happen at any time during the subsequent growth of the tree and may not be evident for up to ten years after propagation. Mosse (1962) also noted that locally incompatible unions exhibit changes in their vascular orientation, which may be explained by the need for the xylem to grow around the regions of parenchyma cells. In addition to the eventual failure of the union, the tree slowly dies due to the reduced transport of water and nutrients across the union into the scion. A third type of incompatibility was described and is referred to as delayed incompatibility (McCully, 1983). It has been described as an incompatibility that takes many years to develop. However, Mosse (1962) and Andrews and Serrano Marquez (1993) suggested that this term can be misleading and should not be used. Andrews and Serrano Marquez (1993) explained that although a tree may appear to gradually decline, the factors leading to the decline were likely present in the tree long before the incompatibility became apparent. They also described how many cases of delayed incompatibility were later shown to be caused by disease. Although translocated incompatibility generally causes trees to decline rapidly and localized incompatibly causes a slow decline, these characteristics are not definite, and the same tree can exhibit both forms of incompatibility in varying degrees (Mosse, 1962). 68

79 Large areas of ray parenchyma cells associated with localized incompatibility may be detrimental to the strength of the tree. Parenchyma cells are the living cells of secondary xylem tissues and they are primarily associated with the storage and transport of nutrients within the tree. They also aid in wound healing, as parenchyma cells divide into callus parenchyma when a tree is injured (McCully, 1983). While callus tissue is necessary for wound healing, having too much of it at the union may lead to an irregular orientation of the xylem (Simons and Chu, 1985). Fiber cells are another common cell type found in xylem tissue. As fiber cells mature, they form a lignin rich secondary cell wall and then die. These secondary walls consist of strands of cellulous that are held in place by a matrix of hemicelluloses and lignin. These physical properties of the cell walls give fiber cells their strength and elasticity (Déjardin et al., 2010) allowing the fiber cells to provide most of the mechanical support to the tree (Winandy and Rowell, 2013). An abundance of callus tissue and ray parenchyma cells has been observed in previous anatomical studies of apple unions. Simons and Chu (1985) found callus parenchyma cells present at the base of the scion in the union of Red Delicious / M.26. They noted that this led to an uneven development of the subsequent vascular tissues and that the callus tissue and ray parenchyma cells continued to develop in the subsequent year s growth. Ussahatanonta and Simons (1988) found swirling xylem tissue interspersed with necrotic parenchyma tissue in the unions of Golden Delicious / M.9. They also observed ray cells developing with an abnormal orientation, and noted that the Golden Delicious / M.9 unions were relatively weak compared to the other trees in the study, which included Golden Delicious scions on seedling, Malling-Merton 106 ( MM.106), M.26, and M.7A rootstocks. Soumelidou et al. (1994) observed bands of parenchyma cells followed by bands of thin-walled fiber cells in the budded unions of Bramley s Seedling / M.9, though they did not indicate if these were callus tissue or ray parenchyma cells. Signs of incompatibility were observed in the unions of other clonally propagated tree fruit species. Ermel et al. (1997) found poor cambial connections in incompatible combinations of pear (Pyrus communis L. cv. Epine du Mas Clone P 1885 ) scions on quince (Cydonia oblonga Mill. cv. Old Home x Farmingdale OHF 333 Clone P 2420 ) rootstocks. Zarrouk et al. (2010) found that incompatible five-month-old 69

80 combinations of peach (Prunus persica L. cv. Summergrand ) on cherry plum (Prunus cerasifera Ehrh. cv. Myrobalan 3-1 ) lacked connected cambia. Zarrouk et al. (2006) found unlignified tissues within incompatible combinations of the peach and nectarine cultivars Catherina, Tebana, and Summergrand on Damson plum (Prunus insititia cv. PM 95 AD ) rootstocks. Errea et al. (1994) found parenchyma tissue within the unions of incompatible combinations of apricot (Prunus armeniaca L. cv. Moniqui ) scions and rootstocks of the cherry plum Mirobolán 605. Herrero (1951) measured the proportions of cell types in the scions of incompatible combinations of the peach Hale s Early on cherry plum Myrobalan B, and found that the wood contained more ray parenchyma and less fiber cells than the wood of the compatible combination Hale s Early and domestic plum (Prunus domestica L. cv. Brompton ). While much of the previous work performed on incompatible graft unions showed an overproduction of parenchyma cells and irregular orientations of vascular tissues, these studies were performed using either light microscopy or scanning electron microscopy. These methods often took a long time to prepare the samples for sectioning and provided a limited view of the tissue distribution, as only a single layer of tissue could be observed at a time. To better understand the distribution of the parenchyma tissue within the entire union, a rapid method for observing the tissues of the entire union is required. Researchers observed the structure of the graft union without the aid of microscopy. Warmund et al. (1993) observed the graft unions of Jonagold / Mark using Magnetic Resonance Imaging (MRI) and were able to detect vascular discontinuity. Areas of vascular continuity appeared bright in MRI pictures, while poorly connected areas were dark. They attributed these changes to differences in signal intensities between the tissues, which they suggested were related to the tissues water content. While the MRI technique did not allow them to determine the cause of the discontinuity, upon sectioning and microscopic examination they found that the trees possessed areas of irregular parenchyma tissue within their unions. They noted that some trees within their field plantings broke along the line of this parenchyma tissue. Milien et al. (2012) used X-ray computed tomography (Ct X-ray) scans to observe the anatomy of two graft unions of grape (Vitis vinifera L.) on the rootstock

81 Richter. The scionwood was from two self-hybrids of Syrah. The two differed in their development, and were designated as either a good graft or a bad graft. The study found that the Ct X-ray scans could be used to analyze the unions in three dimensions (3D). Differences between the grafts could be distinguished at the tissue level because the xylem, phloem, and pith were identifiable. They were also able to observe the orientation of the vascular tissues. The bad graft produced less xylem tissue after the grafting process than the good graft and the wood tissue of the bad graft was also less dense. The good graft had more vessel continuity, but further vessel analysis was limited because vessels could not be observed individually because of limitations in the resolution of the 3D image rendering. Laser ablation tomography (LAT) is another tool that has been used to examine 3D plant structures. LAT was used to observe anatomical traits of the roots of maize (Zea mays L.) (Chimungu et al., in press) and to visualize the unions of apple trees (Harrison, personal communication). LAT utilizes a laser beam to ablate sections of a sample in sub-micrometer increments. As the sample is ablated, a camera photographs the remaining tissue surface while the sample is simultaneously positioned forward as it rests on a computer controlled stage. The method is destructive, but sequential photographs are taken and imaging software allows for a reconstructed 3D view of the sample. Since LAT has been used to show differences at the cellular level, it may possibly be used to compare the union tissues of weak ( Honeycrisp / M.26 EMLA ) and strong ( Honeycrisp / M.7 EMLA, Zestar! / M.26 EMLA, Zestar! / M.7 EMLA ) apple unions. Our objectives were to determine if we could observe differences between these combinations and to determine if this method would allow us to rapidly evaluate future potentially weak combinations. In addition to LAT, we evaluated the possibility of using iodine potassium iodide (I 2 KI) solution as an indicator of localized incompatibility at the graft union. Iodine solutions have been used in the past to observe starch concentrations within graft unions because the accumulation of starch above the union is a common symptom of translocated incompatibility (Ermel et. al, 1999; Herrero, 1951). Herrero (1951) found that starch tended to accumulate above the graft union in incompatible combinations of peach when Hale s Early scions were grafted on cherry 71

82 plum Myrobalan B rootstocks. He noted that these trees also suffered from external symptoms commonly associated with translocated incompatibility, including decreased tree growth and early defoliation. Ermel et al. (1999) observed starch in the unions of pear and quince. Compatible combinations included scions of the pear Passe-Crassane Clone P 27 on the pear rootstock Old Home x Farmingdale 333 Clone P 2420 and the quince rootstock East Malling C Clone C 348. Scions of the pear cultivar Epine du Mas Clone P 1885 were also compatible on the pear rootstock, but incompatible on the quince. They found few differences in the relative amount of starch between incompatible and compatible unions, but noted that incompatible unions had more starch in the scions than in the rootstocks. Herrero (1951) also observed starch levels in combinations of plum scions on Myrobalan B rootstocks with plum interstocks. The plums Victoria and President were compatible with the rootstock, while Oullins Golden Gage and Reeves Seedling were incompatible. Combinations of pear/quince were also evaluated. Buerré Hardy was the compatible pear scion, and Williams Bon Chrétien and Clapp s Favorite were incompatible on the quince rootstock Quince B. Each scion was grafted on an interstock of the three scion cultivars, and was then grafted on the rootstock, producing nine total combinations. These trees may have had localized incompatibilities, because no symptoms other than vascular disruption were observed. The incompatible combinations had lower concentrations of starch in the interstocks relative to their rootstocks, but he noted that starch concentrations fluctuated throughout the year in all of the combinations. We hypothesize that the unions of the weak combinations in our study will have a greater percentage of parenchyma cells directly at the union of the rootstock and scion. Since parenchyma cells are known to store and transport starch (Déjardin et al., 2010), they should stain heavily with the addition of the iodine solution. Since our trees are prone to union failure, they are likely to primarily suffer from localized incompatibility, though we cannot rule out translocated incompatibility. In which case, we would expect starch to stain above the union as well. Along with the Honeycrisp and Zestar! combinations previously described, the weak combination Scilate (Envy )/ G.41 was 72

83 compared to the strong combinations Scilate / M.9 NIC29 and Cripps Pink cv.maslin/ M.9 NAKB T337. Materials and Methods Experiment 4.1. Laser ablation tomography. For this study, 16 finished chip-budded nursery trees were received in Apr from Adams County Nursery, Aspers, PA. These included four trees each of the cultivars Honeycrisp and Zestar! on the rootstocks M.26 EMLA and M.7 EMLA ( H26, H7, Z26, and Z7 ). Trees were budded in Trees were then kept at 6 C until sampling. Using a circular saw, trees were transversely cut to 10.0cm long sections of wood. Trees were then cut longitudinally to mm in thickness using a band saw. Sections were then cut to a width of 2.5cm to fit within the path of the laser beam. Sections were stored in 70% ethanol for at least one week before ablating, as ethanol aided in the fluorescence of the xylem tissue. Samples were ablated using an AVIA mm pulsed laser (Coherent Inc., Santa Clara, CA). Samples were ablated at 100.0µm intervals to either 2.5cm or 3.0cm in length from top to bottom, starting at the height of the tree where the rootstock and scion met. Images were taken in 100.0µm increments for all but one sample, as one sample of H26 was photographed at every µm. Images were captured using a Canon T3i camera (Canon Inc., Tokyo, Japan) with a Canon MP-E 65mm 5x micro lens. For our study, the lens was reduced to 1x zoom to capture a greater field of view of the union. Images were stacked to create 3D models of the sections using Avizo imaging software, (FEI Company, Hillsboro, OR). Samples were visually inspected for the development of ray cells and callus parenchyma tissue, irregularly oriented xylem, and areas of necrosis. Experiment 4.2. Graft union starch accumulation. This study used 24 trees. Four trees each of the combinations H26, H7, Z26 and Z7 were compared. In addition to these 16, three trees each of the cultivar Scilate budded on the rootstocks G.41 and M.9 NIC29 were evaluated ( S41 and S9 ), along with two trees of Cripps Pink budded on the rootstock M.9 NAKB T337 (P9). The 73

84 Cripps Pink and Scilate samples were received in Feb from Willow Drive Nursery, Ephrata, WA, and were also budded in All trees were cut into longitudinal sections as described in experiment 4.1. Trees were then stored in water and stained with Lugol s iodine solution (5% iodine potassium iodide) (VWR International, Radnor, PA) shortly after or were stored in 70% ethanol for long term storage before staining. Sections were then stained with Lugol s solution by applying enough solution to moisten the entire surface of the sample. Samples were allowed to react for two minutes. Sections were then photographed with a Canon A4000 Powershot camera (Canon Inc. Tokyo, Japan). Images were analyzed using ImageJ image analysis software (National Institutes of Health, Bethesda, Maryland) (Rasband, 2014). Images were cropped to include only the xylem tissue of each section, and this area of xylem was recorded. The color brightness was then adjusted using brightness threshold values between one and 70. The brightness threshold separates pixels based on their brightness. The first sample of H26 was used to choose these numeric parameters. This range included the most heavily stained areas, without including the completely black background, which had a brightness of 0. The value of 70 was chosen as the upper limit to objectively standardize the upper limit of stained wood that would be included in our calculations. This allowed us to set an objective level of darkness that would represent the stained parenchyma tissue. The area of the selected tissue was then recorded. Data were recorded as the percent of stained tissue to total wood tissue of the union to correct for differences in tree size. While the original trees had been cut to 10.0cm before longitudinal sectioning, the actual tree sections used for staining were variable in height and width because the bud unions did not always line up with the center of the tree. Means and standard deviations of the percent by combination were calculated. Due to non-normality and very small sample sizes, means were compared using the non-parametric Kruskall-Wallis test in Minitab 17 (Minitab Inc., State College, PA). Data were also transformed using a Box Cox transformation to meet the assumptions of normality and homogeneity of variances, and an analysis of variance (ANOVA) test was conducted on the transformed data points. Post-hoc comparisons 74

85 were not performed, as neither the Kruskall-Wallis nor the ANOVA returned significant differences between combinations (p-values =.696 and.708, respectively). Results and Discussion Experiment 4.1. We were able to produce 3D models of the graft unions using LAT; however they were limited to the thickness of the sections so a whole tree model could not be created (Figure 4.1). Sections were no more than 4.0mm thick, as this was the maximum thickness that we could effectively ablate. Figure 4.1. A three dimensional model of a section of the union of H26 showing the longitudinal and radial planes of the section, with the rootstock on the lower left and scion on the upper right. Differences in the amount of parenchyma tissue present and the orientation of the xylem tissues were examined in the transverse images (Figure 4.2) and in the simulated longitudinal planes (Figure 4.3) of the combinations. 75

86 Figure 4.2. Transverse sections of wood from H26 (A) H7 (B) Z26 (C) and Z7 (D) with the scions on the left and rootstocks on the right. The wood tissue of H26 shows a large area of swirling xylem (SX) tissue within the subsequent year of growth. In H7, necrotic wood (N), callus tissue (Ca), and bark-like tissue can be seen. In Z26, an area of necrosis surrounded by callus tissue can also be observed. Z7 also shows a small section of bark-like necrotic tissue. Fragments of the callus tissue that initially bridged the gap between the rootstock and scion can be seen within the unions of H26 and H7. A SX Ca B Bark-like tissue Ca N Ca C N Ca D N 76

87 Figure 4.3. Unions of H26 (A), H7 (B), Z26 (C) and Z7 (D) in longitudinal view with the rootstock on the left and the scion portions on the upper right. Swirling xylem (SX) is circled at the middle of the union extending towards the bark in H26. H7, Z26, and Z7 appear to have isolated areas of necrosis (N). Callus tissues (Ca) and empty spaces surrounding them between the rootstock and scion can be easily distinguished in H7 and Z26. The wood tended to split at this callus layer during the ablation process, producing these gaps. An additional small area of callus is seen in Z26. Open spaces further down the union of H26 and in Z26 (arrows) were very thin gaps also likely caused by the ablation process. A B Ca SX N C D Ca Ca N N 77

88 Callus parenchyma tissue was present in all combinations between the rootstock and scion. Swirling tissue was commonly observed in the scion adjacent to the union and in areas of callus parenchyma proliferation. A very large section of swirling xylem extended into the following season s growth in one sample of H26 (Figure 4.2A), this has been described in other studies of the union, and may be a sign of weakness (Simons and Chu, 1985). This swirling tissue may have been accompanied by a large amount of poorly arranged ray parenchyma (Ussahatanonta and Simons, 1988) and may be a sign of localized incompatibility (Mosse, 1962). However, because we did not have cellular resolution, this could not be confirmed in this sample. Other irregularities could be observed from our images. In H7, Z26, and Z7, one sample of each contained a large area of necrotic tissue. In H7, the tissue around this necrotic wood consisted mostly of callus tissue, and it extended towards the outer growth of the union. H7 also appeared to have a few large areas of parenchyma tissue as well. Tissue that resembled bark was also present (Figure 4.2B and Figure 4.3B), though it is difficult to determine where this tissue derived from. Mosse (1962) described how bark tissues from the rootstock and scion could fill in parenchymatous regions within the unions of incompatible grafts. This inclusion of the bark tissues is caused by the reorientation of the cambium cells within the union. However, this tissue may also have been remnant bark tissue from the budding process that the callus subsequently grew around. In either event, this material may disrupt the continuity of the xylem, potentially weakening the union. In Z26, the vascular system had a small region of callus disrupting the xylem at the union, though normal xylem growth soon began to differentiate from it (Figure 4.3C). A region of necrotic wood tissue surrounded by wound callus was also observed further down the union as well (Figure 4.2C). However, we do not believe this was related to incompatibility as this tissue was found below the union in the subsequent year s xylem growth (Figure 4.3C). This suggests the necrotic tissue may have been caused by a wound independent of the budding process. A sample of Z7 also had a necrotic zone where new wood tissue was growing around what appeared to be remnant bark material (Figure 4.2D). Unlike the union of H7, the presence of this material in Z7 created few disturbances to the xylem around it, because only a small area of callus developed 78

89 (Figure 4.2D). This union may have had less parenchyma around it because it was a smaller area of necrotic tissue, or it may suggest that Z7 was inherently better at reestablishing the proper orientation and differentiation of xylem than H7. Since we found very little parenchyma tissue around it, we could further rule out the likelihood that this tissue was related to incompatibility. The wood of Z7 appeared to have the least amount of irregular parenchyma tissue and swirling vascular tissues. In this combination callus tissue was found adjacent to where the bud was inserted, and rarely extended into the subsequent year s growth. This further suggests that Z7 may be better adapted to reestablishing normal vascular differentiation shortly after the budding process. In addition to the irregular parenchyma tissue and areas of necrotic tissues present in the wood of H7, Z26, and Z7, empty spaces between the rootstock and scion where the callus tissue had formed were commonly observed (Figure 4.3). These large gaps occurred between the rootstock and scion tissues where the original callus bridge formed, however we do not believe these gaps were present in the actively growing trees. Instead, these may have been artifacts of the ablation process. When trees are cut to be budded, a few layers of cells are injured and subsequently die, which produces a necrotic layer that is then surrounded and bridged by the proliferating callus cells (Soumelidou et al., 1994). Since samples were treated in 70% ethanol to promote florescence during ablation, the wood may have dried out and split along this necrotic layer. It is unclear how much these necrotic layers may have been a point of weakness within the actively growing trees, though from our observations they did not seem to be present in greater numbers in any one of the combinations, so they may not relate well to the overall weakness observed in H26. In terms of the previous descriptions of incompatibility provided by Mosse (1962) and Andrews and Serrano Marquez (1993), we were able to find a large area of swirling xylem tissue within the wood of one sample of H26, a combination known to experience graft failure in field plantings. Swirling xylem is sometimes accompanied by poorly arranged ray parenchyma (Ussahatanonta and Simons, 1988), which would suggest that the weak combination H26 may suffer from localized incompatibility. While the other three combinations are not prone to failure in the field, we found areas of 79

90 swirling xylem, parenchyma, and necrotic areas within their unions as well. While these may have been caused by external factors, such as propagation error or non-related sources of wounding, these findings suggest that it may be difficult to predict tree strength using the histological traits of irregular parenchyma, necrosis, and swirling xylem tissues alone, and warrants a quantitative examination of the full union. Additionally, Mosse (1962) had suggested that a proliferation of ray parenchyma could happen at any time in the life of the tree. This suggests that while we may not find evidence of anomalies in the wood of young nursery trees, they may still prove incompatible as they age. LAT is a promising technology for the visualization of plant structures. While LAT has been used to observe the morphological features of herbaceous plant roots, the system is currently not designed to handle the large woody stems that would be necessary to create a full 3D model of an apple union. The laser is most effective when ablating material no more than mm in thickness, which limits its ability to handle larger pieces of woody tissue. In addition, at the 1x zoom we were unable to have more than 2.5cm in width in our field of view. Being at a low level of zoom, we were also unable to achieve the resolution necessary to differentiate samples at the individual cell level. The time it took to ablate a sample was also a limiting factor. Each sample took between seven to nine hours to ablate 3.0cm of sample, depending on its thickness. In that amount of time, the sample would dry out and become distorted, making it difficult to achieve adequate image quality. It may be possible to ablate green tissue faster; however there will still be a size limitation. Finally, the destructive nature of LAT sampling makes it impossible to observe the development of the same union over time. CT X-ray scans may prove more useful for studying tissue-level differences of young graft unions, as this system allows for the modelling of the full width of the graft union (Milien et al., 2012). Experiment 4.2 In a few of our samples, we were able to observe differences that matched our hypotheses. In two of the four samples of H26, the union stained very dark, showing an extensive amount of starch that may represent poorly differentiated parenchyma tissue developing at the union (Figure 4.4A). This was also observed in a sample of Z26 80

91 (Figure 4.4C). Although trees can suffer from both forms of incompatibility simultaneously (Mosse, 1962), our trees may have suffered primarily from localized incompatibility. The accumulation of starch above the union was seen in previous studies of trees showing other symptoms of translocated incompatibility, including a degeneration of the phloem tissue and a lack of xylem discontinuity (Ermel et al., 1999; Herrero, 1951). However, we observed the most intense wood staining in the tissues at the union and rootstock, suggesting that starch transport was not blocked at the union. Figure 4.4. The unions of H26 (A) H7 (B), Z26 (C), and Z7 (D) with the scion on the upper left and the rootstock below. The images have had the brightness threshold adjusted, and the tissue stained by the iodine solution appears in red. The red shows the concentration of parenchyma tissue at the union. In the unions of H26 and Z26, the parenchyma tissues are in a line that corresponds to where the scion and rootstock join, and is fairly continuous across the union. The union of Z7 contains parenchyma, but a large section of the union consists of woody tissue. A B C D 81

92 The amount of wood tissue stained between combinations was extremely variable (Table 4.1). In our three samples of S41, the percentage of stained wood tissues varied between 30% and.04%. The Kruskall-Wallis test resulted in a p-value =.695, while the ANOVA on the transformed values resulted in a p-value =.708. Table 4.1. The percentages of wood tissues stained by I 2 KI. Combination Mean Percent Standard Deviation 'Honeycrisp'/'M.26 EMLA 'Honeycrisp'/'M.7 EMLA 'Zestar!'/'M.26 EMLA' 'Zestar!'/'M.7 EMLA' 'Cripps Pink'/'M.9 NAKB T337' 'Scilate'/'G.41' 'Scilate'/'M.9 NIC29' Kruskall-Wallis p-value.695 ANOVA p-value.708 While the amount of starch at the union may be highly variable, the inconsistencies we observed could have been caused by a number of factors related to our experimental procedures. In addition to having very low sample numbers, our samples often lacked uniformity in size and shape. They may have also not captured the exact center of the bud insertion point because the bud did not always grow straight out from the rootstock. To correct for these problems, uniform samples from the center of the union need to be taken. However, getting uniform sections may be difficult because of these differences in how the scion grows out from the union. In addition to size differences, wood samples that had been stored in ethanol appeared much darker than freshly cut sections. It is likely that the brightness threshold could not separate differences between staining caused by the I 2 KI or from storage in ethanol. The camera used for imaging the sections may also have been problematic. The camera had automatic settings, which allowed the camera to change settings between each image depending on the amount of light present. This may have changed how dark 82

93 each image appeared. A camera with manual settings should have been used, and the camera should have been kept at a fixed distance from the samples in a room with consistent light levels to reduce the likelihood of there being differences in brightness caused by external sources of variation. Since our findings were variable within each combination, and Herrero (1951) found that starch concentrations were variable in graft unions depending on the time of year, measuring the total amount of wood tissue stained may not prove useful for rapidly determining incompatibility in apple scion/rootstock combinations. Other histological features may prove more useful, however this may require analyzing many histological features together (Ermel et al, 1999) because few histological disorders are unique to incompatibility alone (Andrews and Serrano Marquez, 1993). While methods for detecting anatomical anomalies are improving, the anatomical features of incompatibility can be variable and may not develop until the tree is already in the orchard (Mosse, 1962). To prevent the large scale propagation of brittle apple combinations, a better understanding of the genetic and biochemical causes of incompatibility is necessary. 83

94 Literature Cited Andrews, P.K., and C. Serrano Marquez Graft incompatibility. Hort. Rev. 15: Chimungu, J.G., M.F.A. Maliro, P.C. Nalivata, G. Kanyama-Phiri, K.M. Brown and J.P. Lynch. In press. Utility of root cortical aerenchyma under water limited conditions in tropical maize (Zea mays L.). Field Crops Res. Déjardin, A., F. Laurans, D. Arnaud, C. Breton, G. Pilate, and J Leplé Wood formation in angiosperms. Comptes Rendus Biologies. 333: Ermel, F.F., J.L. Poëssel, M. Faurobert, and A.M. Catesson Early scion/stock junction in compatible and incompatible pear/pear and pear/quince grafts: a histocytological study. Ann. Bot. 79: Ermel, F.F., J. Kervella, A.M. Catesson, and J.L. Poëssel Localized incompatibility in pear/quince (Pyrus communis/cydonia oblonga) combinations: multivariate analysis of histological data from 5-month-old grafts. Tree Phys. 19: Errea, P., A. Felipe, and M. Herrero Graft establishment between compatible and incompatible Prunus spp.. J. Exp. Bot. 45: Harrison, N East Malling Research, Personal communication. Herrero, J Studies of compatible and incompatible graft combinations with specific reference to hardy fruit trees. J. Hort. Sci. 26: Manly, N Willow Drive Nursery, Personal communication. McCully, M.E Structural aspects of graft development. p In: R. Moore (ed.). Vegetative compatibility responses in plants. Baylor Univ. Press, Waco, TX. Milien, M., A.S. Renault-Spilmont, S.J. Cookson, A. Sarrazin, and J. Verdell Visualization of the 3D structure of the graft union of grapevine using x-ray tomography. Scientia Hort. 144: Mosse, B Graft incompatibility in fruit trees: with particular references to its underlying causes. Tech. Comm. 28. Comm. Bur. Hort. Plant Crops, East Malling, England. Privé, J.P., A. LeBlanc, and C.G. Embree Preliminary evaluation of supported and free standing 'Honeycrisp' trees on 24 apple rootstocks. Acta Hort. 903: Rasband, W.S ImageJ. U. S. National Institutes of Health. Bethesda, Maryland, USA. Robinson, T.L., G. Fazio, T. Holleran, and H. Aldwinckle The Geneva series of rootstocks from Cornell: Performance, disease resistance, and commercialization. Acta Hort. 622:

95 Simons, R.K. and M.C. Chu Graft union characteristics of M.26 apple rootstock combined with 'Red Delicious' strains-morphological and anatomical development. Scientia Hort. 25: Soumelidou, K., N.H. Battey, P. John, and J.R. Barnett The anatomy of the developing bud union and its relationship to dwarfing in apple. Ann. Of Bot. 74: Ussahatanonta, S. and R.K. Simons Graft union development of the Golden Delicious apple when combined with various dwarfing rootstocks. Fruit Var. J. 42: Warmund, M.R., B.H. Barritt, J.M. Brown, K.L. Schaffer, and B.R. Jeong Detection of vascular discontinuity in bud unions of Jonagold apple on mark rootstock with magnetic resonance imaging. J. Amer. Soc. Hort. Sci. 118: Winandy, J.E. and R.M. Rowell Chemistry of wood strength. p In: R.M. Rowell (ed.) Handbook of wood chemistry and wood composites. CRC Press. Taylor & Francis Group. Boca Raton, FL. Zarrouk, O., Y. Gogorcena, and M.A. Moreno Graft compatibility between peach cultivars and Prunus rootstocks. HortScience. 41: Zarrouk, O., P. Testillano, M.C. Risueno, M.A. Moreno, and Y. Gogorcena Changes in cell/tissue organization and peroxidase activity as markers for early detection of graft incompatibility in peach/plum combinations. J. Amer. Soc. Hort. Sci. 135:

96 Chapter V Closing thoughts From these studies, we may have narrowed down the cause of the weak unions commonly observed in H26 to localized incompatibility, and the weak unions of P41 and S41 to either localized incompatibility or another scion/rootstock interaction affecting the proportion of the vascular tissues. Although H26 contained a large area of discolored wood within its rootstock, the strong combinations of H7 and Z7 contained more. The weak G.41 combinations also contained the least discolored wood, further suggesting that this did not impact the overall union strength. Fiber cell wall thickness varied between some of the weak and strong combinations below and at the union, though this test appears to be an unreliable indicator of union strength. Interestingly, few differences were observed between the Honeycrisp and Zestar! combinations in the scions, but the Cripps Pink combinations varied there and both combinations of Scilate were very thin at the scion. In the Honeycrisp and Zestar! combinations, we found fiber cell wall thickness may be somewhat dependent on the overall scion/rootstock combination rather than from the rootstock or scion alone, though this interaction between the two may not necessarily be related to incompatibility. The xylem that we observed and quantified in the subsequent year s growth of H26 consisted of more parenchyma and fewer fiber cells than most of the strong combinations, in this may be due to localized incompatibility. This trend was also observed in the weak G.41 combinations; however these did not appear significantly different than some of the strong combinations. The increased production of parenchyma may be partially due to rootstock vigor differences, (Beakbane and Thompson, 1947); however it may also be related to localized incompatibility, as many studies (Mosse, 1962; Simons, 1987; Andrews and Serrano Marquez, 1993) have described an abundance of parenchyma as its most common symptom. While our findings were in agreement with Herrero s (1951) observation of more parenchyma and less fiber cells in translocatably incompatible combinations of peach/cherry plum, this inconsistency may 86

97 be explained by differences in where the unions were sampled. In addition, our starch test showed little evidence of starch accumulating above the union, which is a common symptom of translocated incompatibility. The cause of this incompatibility remains unknown. Hormonal factors may have reduced the amount of xylem differentiation. Auxin plays a large role in xylem differentiation, and cytokinin also plays a role in the differentiation of xylem after wounding (Aloni et al., 2010). It is possible that these hormones could not cross the graft union in these combinations, resulting in less vascular differentiation. In addition to hormonal imbalance, various other factors may be associated with the overproduction of parenchyma. As other studies have shown (Ermel et al., 1997; Zarrouk et al., 2010), callus cells may develop poorly between the rootstock and scion shortly after wounding. This may lead to a poorly formed cambium that can ultimately result in poor differentiation later in the development of the wood, which Mosse (1962) had suggested as a cause of the sudden development of poorly lignified ray parenchyma cells years later in the wood. As Santamour (1988) suggested, a lack of cellular recognition may be caused by differences in the types of peroxidase being produced between the rootstock and scion (Zarrouk et al. 2010). This may cause differences in the types of lignin being produced (Santamour, 1988), which may disrupt the formation of cell walls and plasmodesmata between the cells (Pina and Errea, 2005). This lack of recognition may also be related to the development of destructive polyphenols, which could further disrupt the organization of vascular tissues in the wood of the weak unions (Dos Santo Pereira et al., 2014). These studies suggested that these differences may help to explain incompatibility, and may also be a biochemical and genetic means for the early detection of incompatibility. However, since the exact mechanisms are still uncertain, more biochemical work is necessary. While our findings suggest the structural weaknesses of our study combinations may have been caused by incompatibility, these methods may not prove useful as screening methods for determining strength in future scion/rootstock combinations. The wood tissues of unions are very variable (Doley, 1974, Mosse, 1962), so observing small sections of the union may not be useful for determining their strength. Our efforts to 87

98 observe the full union were unsuccessful using laser ablation tomography, because our samples were too large and woody to be effectively ablated to yield a three dimensional model of the entire union. Differences in callus cells may be distinguishable between compatible and incompatible combinations shortly after grafting (Ermel et al., 1997; Zarrouk et al., 2010). If newly budded trees are small enough to be effectively ablated, this technology may allow researchers to observe the cambial development of the unions. These observations may allow for screening of incompatibility, however, determining incompatibility on a single histological trait alone would be difficult due to the large amount of variation often observed (Ermel et al., 1999). In addition to incompatibility, the G.41 rootstocks may also simply contain more parenchyma in their wood, as very dwarfing varieties tend to produce larger percentages of parenchyma in their roots (Beakbane and Thompson, 1947). While this may further reduce their mechanical strength, having more parenchyma may provide them with other benefits, such as increased disease resistance. Having more parenchyma may allow these rootstocks to produce more chemical extractives and vessel gums during microbial invasion, which could help defend them against many bacterial and fungal diseases (Shigo and Hillis, 1973). Though some of these rootstocks are known to produce weak unions with certain scion cultivars, they should not be completely dismissed by fruit growers. Modern highdensity training systems rely on training trees to a wire-and-trellis system immediately after planting. While theses combinations may always be of higher risk, these failures may be reduced through proper training. Since these incompatibilities are scion cultivar specific, these rootstocks can still be used with many other scions. Additionally, these weak rootstocks may provide the grower with many benefits, such as cold hardiness and resistance to major apple [Malus xsylvestris (L.) Mill. var. domestica (Borkh.) Mansf.] diseases like fire blight [Erwinia amylovora (Burr)] and phytopthora root rot (Phytopthora cactorum) (Robinson et al., 1999). Every rootstock commercially available has its disadvantages, and these should all be considered when planting an orchard. 88

99 Literature Cited Aloni, B., R. Cohen, L. Karni, H. Aktas, and M. Edelstein Hormonal signaling in rootstock-scion interactions. Sci. Horti. 127: Andrews, P.K., and C. Serrano Marquez Graft incompatibility. Hort. Rev. 15: Beakbane, A.B. and E.C. Thompson Anatomical studies of stems and roots of hardy fruit trees. IV. The root structure of some new clonal apple rootstocks budded with Cox's Orange Pippin. J. Pomol. 23: Doley, D Effects of rootstocks and interstock on cell dimensions in scion stems of apple (Malus pumila Mill.) New Phytol. 73: Dos Santos Pereira, I., R. Da Silva Messias, Â. Diniz Campos, P. Errea, L.E. Corrêa Antunes, J.C. Fachinello, and A. Pina Growth characteristics and phenylalanine ammonia-lyase activity in peach grafted on different Prunus spp.. Biol.Plant. 58: Ermel, F.F., J.L. Poëssel, M. Faurobert, and A.M. Catesson Early scion/stock junction in compatible and incompatible pear/pear and pear/quince grafts: a histocytological study. Ann. Bot. 79: Ermel, F.F., J. Kervella, A.M. Catesson, and J.L. Poëssel Localized incompatibility in pear/quince (Pyrus communis/cydonia oblonga) combinations: multivariate analysis of histological data from 5-month-old grafts. Tree Phys. 19: Herrero, J Studies of compatible and incompatible graft combinations with specific reference to hardy fruit trees. J. Hort. Sci. 26: Mosse, B Graft incompatibility in fruit trees: with particular references to its underlying causes. Tech. Comm. 28. Comm. Bur. Hort. Plant Crops, East Malling, England. Pina, A. and P. Errea A review of new advances in mechanism of graft compatibility-incompatibility. Sci. Hort. 106:1-11. Robinson, T.L., J.N. Cummins, W.C. Johnson, S.A. Hoying, H.S. Aldwinckle, and J.L. Norelli Orchard performance of fire blight-resistant Geneva apple rootstock. Acta Hort. 489: Santamour Jr., F.S Graft compatibility in woody plants: An expanded perspective. J. Environ. Hort. 6: Shigo, A.L. and W.E. Hillis Heartwood, discolored wood, and microorganisms in living trees. Ann. Rev. Phytopathol. 11: Simons, R.K Compatibility and stock-scion interactions as related to dwarfing, p In: R.C. Rom and R.F. Carlson (eds.) Rootstocks for fruit crops. John Wily and Sons. New York, NY. 89

100 Zarrouk, O., P. Testillano, M.C. Risueno, M.A. Moreno, and Y. Gogorcena Changes in cell/tissue organization and peroxidase activity as markers for early detection of graft incompatibility in peach/plum combinations. J. Amer. Soc. Hort. Sci. 135:

101 Appendix Wood discoloration in Honeycrisp / M.26 EMLA. Discoloration appears to enter the callus tissue where the bud union had been inserted. 91

102 Wood discoloration in Zestar! / M.26 EMLA. Discoloration also appears in a full line between the rootstock and scion. Wood discoloration in Honeycrisp / M.7 EMLA. 92

103 Wood discoloration in Zestar! / M.7 EMLA, extending vertically. Wood discoloration in Cripps Pink / G.41. A large section of callus tissue separates the scion from the rootstock. 93

104 Wood discoloration in Scilate / G.41. Wood discoloration in Cripps Pink / M.9 NAKB T337. A necrotic zone also appears between the rootstock and scion tissues. 94

105 Wood discoloration in Scilate/ M.9 NIC29. Discoloration extends through the entire length of the rootstock. Wood discoloration in Zestar! / M.7 EMLA 100x magnification. 95

106 Honeycrisp / M.26 EMLA wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C Zestar! / M.26 EMLA wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C Honeycrisp / M.7 EMLA wood stained with toluidine blue. 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C 96

107 Zestar! / M.7 EMLA wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C Cripps Pink / G.41 wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C Scilate / G.41 wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C 97

108 Cripps Pink / M.9 NAKB T337 wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C Scilate/ M.9 NIC 29 wood stained with toluidine blue 7.0cm below (A), at (B), and 3.0cm above (C) the union. 400x magnification. A B C 98

109 Honeycrisp / M.26 EMLA wood stained with toluidine blue at the union, 200x magnification. 99

110 Zestar! / M.26 EMLA wood stained with toluidine blue at the union, 200x magnification. Color differences were caused by differences in microscope lighting and time after staining. 100

111 Honeycrisp / M.7 EMLA wood stained with toluidine blue at the union, 200x magnification. Color differences were caused by differences in microscope lighting and time after staining. 101

112 Zestar! / M.7 EMLA wood stained with toluidine blue at the union, 200x magnification. 102

113 Cripps Pink / G.41 wood stained with toluidine blue at the union, 200x magnification. 103

114 Scilate / G.41 wood stained with toluidine blue at the union, 200x magnification. 104

115 Cripps Pink / M.9 NAKB T337 wood stained with toluidine blue at the union, 200x magnification. Scilate / M.9 NIC29 wood stained with toluidine blue at the union, 200x magnification. 105

116 Transverse images from the four samples of Honeycrisp / M.26 EMLA produced by LAT. The scion is on the left and the rootstock on the right. Images are from the lower region of the scion/rootstock interface (see image below). 106

117 Transverse images from the four samples of Zestar! / M.26 EMLA produced by LAT. The scion is on the left and the rootstock on the right. Images are from lower regions of the scion/rootstock interface. 107

118 Transverse images from the four samples of Honeycrisp / M.7 EMLA produced by LAT. The scion is on the left and the rootstock on the right. Images are from lower regions of the scion/rootstock interface. 108

119 Transverse images from the four samples of Zestar! / M.7 EMLA produced by LAT. The scion is on the left and the rootstock on the right. Images are from lower regions of the scion/rootstock interface. 109

120 Iodine staining of four samples of Honeycrisp / M.26 EMLA. Samples tended to stain most heavily directly at the union between the rootstock and scion. 110

121 Iodine staining of four samples of Zestar! / M.26 EMLA. While dark staining was also on the paper towel, samples were traced to only calculate the area stained on the woody tissue. 111

122 Iodine staining of four samples of Honeycrisp / M.7 EMLA. 112

123 Iodine staining of four samples of Zestar! / M.7 EMLA. 113

124 Iodine staining of three samples of Scilate / G

125 Iodine staining of three samples of Scilate / M.9 NIC29 115

126 Iodine staining of two samples of Cripps Pink / M.9 NAKB T