Genetic Identification of Hybrid Walnuts ( Juglans × intermedia Carr.) in Hungary, the Hidden Potential for Future Breeding

: The question of the hybrid walnut ( Juglans × intermedia Carr.) is still under debate in the Central European region. There is not simply just an underutilization, rather, even the existence of these hybrid forms is not broadly accepted. On the contrary, there is an intensive cross-breeding activity in the western part of Europe resulting in commercially available hybrid clones. In Hungary, several individuals have been reported with intermediate morphology from different old black walnut plantations. Due to the lacking information, a preliminary study was conducted in order to prove the difference of these selected trees and to test the hybrid state. For this purpose, DNA ﬁngerprinting was used by applying 13 simple sequence repeat (SSR) markers for the identiﬁcation of 22 hybrid genotypes selected from one study plot. A comparative analysis with a reference sample set of the ‘parental’ species and other known hybrids was performed as well. The genetic analysis resulted distinct, unique genotypes for all of the samples. Based on the genetic pattern, the analyzed hybrid group was clearly distinguishable from the other two walnut groups. The result of this study also highlights the hidden potential in walnut breeding in the Central European region. Future concepts concerning hybrid walnut utilization in plantation forestry, agroforestry or as breeding material are also discussed.


Introduction
The genus Juglans is composed of four sections with 21 species worldwide, and comprises one of the most important nut species, Juglans regia L. (English or Persian or common walnut) [1]. Besides their definitely highest importance as nut fruit genetic resources, Juglans species also provide valuable timber, with the North-American Eastern black walnut (Juglans nigra L.) having the highest rank from this aspect. Juglans nigra has a central importance, especially in Europe, as an exotic tree species for highly valuable timber production, due to the superior wood quality and its remarkable light to dark brown colour with lilac traces [2]. Interestingly, black walnut also exhibits several varieties for nut production; however, it is considered only locally important in some parts of the US [3,4].
Hybridization is a relatively common phenomenon among Juglans species in the same section, and a rather rare one among species from different sections [5]. The question of hybridization is a key topic not only in the conventional walnut breeding [6], but also from the point of view of conservation biology strategies, as it is the case for the North-American butternut, Juglans cinerea L., for instance [7]. Furthermore, there is a well-defined group of hybrid walnuts that hold great potential for future walnut management in a broader context, namely the hybrids between the different North-American black walnut species markers, as this method proved to be a reliable tool for the identification of hybrids and the 'parental' species [14,19]. Moreover, a recommendation regarding the future application of the selected superior hybrid genotypes will be discussed.

Materials and Methods
The sample plot is approx. 1 hectare and located in north-west Hungary in a privately owned forest (47.36141267 N, 16.75932137 E, 182 m a.s.l.), near the Ablánc brook in a riparian mixed forest with partly planted origin. Regarding the tree composition, the forest compartment consists of a considerably high ratio of common ash (Fraxinus excelsior L., 38%) and pedunculate oak (Quercus robur L., 21%), mixed with hornbeam (Carpinus betulus L., 15%), sycamore (Acer pseudoplatanus L., 9%), Turkey oak (Quercus cerris L., 9%) and black walnut (Juglans nigra L., 2%), the latter standing within one block. The average age of the forest compartment is 67 years, the closure is 92%. The average height of black walnuts is 23 m, and the average diameter at breast height (DBH) is 29 cm based on the National Forestry Database. The black walnut block is close to its planned rotation age of 80 years.
Twenty-seven putative hybrids (trees with intermediate morphology) were signed individually and sampled for a leaf morphology description and a DNA-test. Plant material was collected in July 2021. The composed leaves were scanned and photo-documented, and the leaves for the DNA analysis were put in a deep freezer, at a temperature of −20 • C. Bark characteristics were also checked ( Figure A1 in Appendix A). Diameters of the putative hybrids were individually measured, as well as eight black walnut trees selected randomly from the plot as part of a broader stand assessment.
For the DNA analysis, the same method was used as previously for fingerprinting the common walnut assortment in Hungary [20], based on the original protocols [21][22][23]. In this study, 13 SSR markers were arranged in three marker sets as follows: set1: WGA89, WGA276, WGA202, WGA9, WGA69; set2: WGA72, JR 6160, WGA27, JR 1817, WGA 118; and set3: WGA01, WGA04, WGA321, WGA331, WGA349. For the comparative analysis, besides 28 hybrid genotypes (22 selected from the study plot, two other putative hybrids from Hungary and four known hybrid cultivars), the reference sample set of the two original species was included with 22 Juglans regia and 10 Juglans nigra samples (Table 1). Data analysis was conducted by GenAlEx 6.5, calculating the main marker indices, P ID , the genetic distance matrix, PCoA and AMOVA [24,25]. Moreover, an UPGMA dendrogram was constructed based on the genetic distance matrix by PAST [26], and STRUCTURE analysis was conducted with the following settings: 200,000 burn-in period, 500,000 MCMC replicates, estimating K from 1 to 6 with 10 repetitions using the admixture model and correlated allele frequencies [27]. STRUCTURE Harvester was used to decide the most likely K with the plateau criterion by Prichard et al. [28], and the deltaK method described by Evanno et al. [29].

Results
The applied 13 markers were all polymorphic, and it can be concluded that the method was appropriate for genetic fingerprinting as the P ID value (in other words, the probability of random matching) was very low (1.1 × 10 −9 for the JR (J. regia) group, 7.5 × 10 −15 for the JN (J. nigra) group, and 4.4 × 10 −14 for the HYB (J. × intermedia) group). The main genetic indices are presented in Table 2. The genetic analysis resulted in distinct, unique genotypes for all samples. Based on the allelic parameters, the highest genetic diversity can be seen in the HYB and the JN groups, while the analyzed JR sample set shows lower variability. Private alleles occur in all the three groups, including the hybrid group, with a moderately high number. To prove the distinctiveness of the two species groups and the hybrids, the molecular variance was checked by AMOVA, as well. AMOVA proved that 19% of the total variance came from the variance among groups (with p = 0.0001 confidence interval and based on standard permutation after 9999 repetitions). The pairwise F st values, representing the ratio of molecular variance among groups, are the following: 0.292 between the J. regia and J. nigra groups; 0.184 between the hybrids and the J. regia group; and 0.113 between the hybrids and the J. nigra group.
For the visualization of the genetic relatedness of the analyzed samples, a principal coordinate analysis (PCoA) was conducted based on the genetic distance matrix. While the two parental species' groups are completely separated, putative hybrids are situated between them, but closer to the black walnut group (Figure 1). The UPGMA dendrogram supported this grouping, common walnut samples separated within one distinct cluster, while black walnut samples in another cluster together with a subgroup of the putative hybrids selected from the study plot ( Figure A2 in Appendix A). On the other hand, the four known hybrid clones, as well as one included hybrid genotype, found also in Hungary, are grouped with the common walnut samples in this analysis, indicating a higher genetic relatedness between these groups by the clustering approach. For the visualization of the genetic relatedness of the analyzed samples, a principal coordinate analysis (PCoA) was conducted based on the genetic distance matrix. While the two parental species' groups are completely separated, putative hybrids are situated between them, but closer to the black walnut group (Figure 1). The UPGMA dendrogram supported this grouping, common walnut samples separated within one distinct cluster, while black walnut samples in another cluster together with a subgroup of the putative hybrids selected from the study plot ( Figure A2 in Appendix A). On the other hand, the four known hybrid clones, as well as one included hybrid genotype, found also in Hungary, are grouped with the common walnut samples in this analysis, indicating a higher genetic relatedness between these groups by the clustering approach. Based on the STRUCTURE analysis, the existence of four clusters has the highest probability based on the mean of the estimated Ln probabilities, while the Evanno method also supports the existence of K = 4 clusters (Figure 2, for the supported deltaK see Figure  A3 in Appendix A). Interestingly, the separation of the analyzed groups is apparent, where the hybrids also form two distinct, unique clusters. The K = 4 pattern refers partly to a sub-clustering within the Juglans nigra group as well. Based on the STRUCTURE analysis, the existence of four clusters has the highest probability based on the mean of the estimated Ln probabilities, while the Evanno method also supports the existence of K = 4 clusters (Figure 2, for the supported deltaK see Figure A3 in Appendix A). Interestingly, the separation of the analyzed groups is apparent, where the hybrids also form two distinct, unique clusters. The K = 4 pattern refers partly to a sub-clustering within the Juglans nigra group as well.

Discussion
The hint of a former hybridization event was detected and studied in an old black walnut stand in Hungary. Twenty-two putative hybrid trees with intermediate morphology were selected for deeper investigation. The selected hybrids represent half of the walnut stand, and 18 trees show even superior diameter growth and height compared with the pure black walnut individuals. Out of these, four outstanding trees with DBH above 45 cm were designated for tree breeding purposes as selected plus trees. A detailed forestry assessment is in progress, with a complex survey of tree height, crown structure and trunk quality. Observations of phenological traits (leafing period, flowering features, fruitification, etc.) have been started as well, to provide a full description of the selected plus trees.
As the spontaneous occurrence of hybrids in such a high number is not a common phenomenon, in this preliminary study, a genetic analysis was conducted to confirm the hybrid state of the observed trees. Based on the genetic analysis, it is clearly visible that the two parental species' groups are completely separated, while putative hybrids sampled from the study site form a distinct group. The applied two ordination methods (PCoA, UPGMA clustering) divide the analyzed hybrids into two subgroups. Both approaches set the study plot's hybrid subgroup (with one additional hybrid from Kámon, Hungary) closer to the black walnut group, while the subgroup of the other hybrids (including the known hybrid cultivars) is located between the parental species (in the case of PCoA), or even closer to the J. regia group (based on the UPGMA).
On the other hand, the Bayesian analysis by STRUCTURE shows a different genetic pattern along the analyzed data set, where the hybrids of the study site form a completely distinct group as if the samples of a third species had been evaluated. On the contrary, the genotypes of the reference hybrids follow a different pattern. From the latter result, the question would arise whether the putative hybrids are real hybrids at all, or if they are rather the progenies of a third Juglans species. As the origin of the propagation material is unknown, this scenario should also be considered. The use of full-or half-sib progenies for forestation was not uncommon in Hungary, especially for exotic tree species-mainly conifers, but also in some rare indigenous broadleaved species-where the propagation material was not available. In these cases, seed collection from solitary trees in parks or botanical gardens used to be a general practice in forestry. In Hungary, two other walnut species can be found besides J. regia and J. nigra. One is the North American butternut (J. cinerea L.), which belongs to the Cardiocaryon section. Another possible species would be the Manchurian walnut (Juglans mandshurica Maxim.) also from the Cardiocaryon section, but from Asia. This walnut species occupies the most northern area and is a very frost resistant type [4,6]. Both mentioned Juglans species are quite rare in Hungary, they can be found mainly individually in botanical gardens or dendrological collections. It is

Discussion
The hint of a former hybridization event was detected and studied in an old black walnut stand in Hungary. Twenty-two putative hybrid trees with intermediate morphology were selected for deeper investigation. The selected hybrids represent half of the walnut stand, and 18 trees show even superior diameter growth and height compared with the pure black walnut individuals. Out of these, four outstanding trees with DBH above 45 cm were designated for tree breeding purposes as selected plus trees. A detailed forestry assessment is in progress, with a complex survey of tree height, crown structure and trunk quality. Observations of phenological traits (leafing period, flowering features, fruitification, etc.) have been started as well, to provide a full description of the selected plus trees.
As the spontaneous occurrence of hybrids in such a high number is not a common phenomenon, in this preliminary study, a genetic analysis was conducted to confirm the hybrid state of the observed trees. Based on the genetic analysis, it is clearly visible that the two parental species' groups are completely separated, while putative hybrids sampled from the study site form a distinct group. The applied two ordination methods (PCoA, UPGMA clustering) divide the analyzed hybrids into two subgroups. Both approaches set the study plot's hybrid subgroup (with one additional hybrid from Kámon, Hungary) closer to the black walnut group, while the subgroup of the other hybrids (including the known hybrid cultivars) is located between the parental species (in the case of PCoA), or even closer to the J. regia group (based on the UPGMA).
On the other hand, the Bayesian analysis by STRUCTURE shows a different genetic pattern along the analyzed data set, where the hybrids of the study site form a completely distinct group as if the samples of a third species had been evaluated. On the contrary, the genotypes of the reference hybrids follow a different pattern. From the latter result, the question would arise whether the putative hybrids are real hybrids at all, or if they are rather the progenies of a third Juglans species. As the origin of the propagation material is unknown, this scenario should also be considered. The use of full-or half-sib progenies for forestation was not uncommon in Hungary, especially for exotic tree species-mainly conifers, but also in some rare indigenous broadleaved species-where the propagation material was not available. In these cases, seed collection from solitary trees in parks or botanical gardens used to be a general practice in forestry. In Hungary, two other walnut species can be found besides J. regia and J. nigra. One is the North American butternut (J. cinerea L.), which belongs to the Cardiocaryon section. Another possible species would be the Manchurian walnut (Juglans mandshurica Maxim.) also from the Cardiocaryon section, but from Asia. This walnut species occupies the most northern area and is a very frost resistant type [4,6]. Both mentioned Juglans species are quite rare in Hungary, they can be found mainly individually in botanical gardens or dendrological collections. It is not unlikely, however, that J. mandshurica individuals were used in forestry test experiments between the 1950s and 1980s in Hungary, originated from the territory of the former Soviet Union. Nevertheless, no descriptions or notes are available proving this hypothesis. Morphological characteristics of the selected trees are also not congruent with the ones of butternut and Manchurian walnut. The typical large leaves with pubescent hairiness and the brownish and hairy twigs are missing, as well as the very emblematic 'moustache' above the leaf scar in the case of butternut. Being a hybrid of one of these mentioned species is again not very likely. Butternut forms hybrids only within its own section, and not with the black walnut or common walnut. Manchurian walnut hybridization may occur with common walnut, but the morphological features are still lacking. For these reasons, it can be concluded that the origin of the analyzed hybrids is rather the one that was proposed previously. However, this first STRUCTURE result can be verified in a future analysis with a broader sample set from various resources of different Juglans species, or with different approaches, such as nuclear barcoding markers or even biochemical profiling [30][31][32].
Nevertheless, the applied fingerprinting method was suitable to prove the uniqueness of untypical trees in the black walnut stand. One certain result of the study is the unrelatedness of the observed hybrid trees and the black walnuts on the study site. The sampled hybrids and black walnut trees formed two distinct groups with different allele compositions, and even private alleles occurred in both groups in a moderately high number. The preliminary hypothesis, that putative hybrids are originated from one (or a few) Juglans nigra mother tree(s) and were mingled with other J. nigra half-sibs, can certainly be excluded. The applied fingerprinting method proved to be a reliable tool for identifying half-or full-sibs based on the shared allele content. On the other hand, field observations with morphology descriptions can provide a reliable and trustable result as well. The random sampling of a few selected individuals could be a timesaving and less expensive approach for the future.
The main goal of the present study was to highlight the existence of spontaneous hybrids in older black walnut stands in the Central European forests, even if the origin of the seed lot is not evident. These hybrids may represent a considerably high economic value due to the increased growth and timber production, combined with better hardiness and resistance, as it has already been recognized for the NG38 × RA, NG23 × RA and MJ209 × RA hybrid walnut clones in the Mediterranean region. However, these wellknown hybrid varieties have not yet been tested in the eastern and rather continental part of Europe, and the chance of a lower tolerance towards continental environmental conditions can be predicted. The fact that the locally adapted genetic resources are given a high priority in various national common walnut breeding programmes would also point towards this concept [33][34][35]. From this aspect, the selection of local hybrids, originated from local resources, can be a key question, as the J. regia trees acting as pollinator in the hybridization process represent a quite different race (Carpathian race) compared to the members of the Atlantic-Mediterranean French race [36]. Nevertheless, this hypothesis should still be proved, and for this purpose, the establishment of common garden trials is needed besides the genetic analysis.
We can specify at least three aims for hybrid walnut selection: timber production, multi-purpose usage in agroforestry, or as future breeding material. The most straightforward option is the use of high-quality timber, similarly to the other hybrid cultivars. The most expensive, but also the most profitable option would be the intensive clonal plantation for timber production purposes applying vegetatively propagated hybrid genotypes. A more affordable solution for forestry application could be the use of hybrids with generative origin, or mixing the expensive vegetative plating material with seedlings of walnut or other noble hardwood species. From the aspect of timber production, the low fertility or even sterility of the hybrids would be preferred, as trees could allocate all their resources for vegetative development. An additional benefit of sterility would be the blocked invasiveness, promoting the acceptance and use of such non-native, clonal material in vulnerable riparian forests.
Another potential option is agroforestry and the multi-purpose usage, provided hybrids are fertile and produce edible nuts. Even if the European consumption is focusing mainly on the common walnut, hybrid nuts could fill in a niche and help the diversification of the nut market. The recent invasion of the Walnut husk fly (Rhagoletis completa Cresson) [37] is again an emerging and serious problem that would characterize future breeding concepts with the application of less sensitive genotypes of other Juglans species through directional crosses.
Finally, considering that the locally selected hybrids as potential genetic resources for future breeding is a less defined field of usage, manifold aspects can be included within it. We have already referred to the walnut husk fly damage and a possible solution using new genotypes with less sensitivity. Another important field of improvement is rootstock breeding. In fact, this is one of the main pillars of plantation management and is not highlighted enough in our region. Furthermore, from the ongoing climate change point of view, the test of various genetic resources with higher resistance and hardiness, and the selection of locally adapted genotypes, would have more and more importance not only in fruit breeding, but also in forestry [38,39]. All this already existing knowledge should be built into a future breeding plan focusing on the changes we are facing locally or worldwide.