Bark morphological and chemical features are differentially correlated with disease resistance and yield in hybrid poplar taxa

In the southeastern United States, the establishment of short‐rotation intensively cultured plantations of hybrid poplar has been hindered by its susceptibility to stem cankers. We evaluated the tradeoffs between biomass yield and disease tolerance in hybrid poplar genotypes belonging to P. deltoides × P. maximowiczii (DM), P. deltoides × P. nigra (DN), P. trichocarpa × P. maximowiczii (TM), and P. deltoides × P. deltoides (DD) taxa. We hypothesized that canker resistant genotypes will have thicker bark but bark thickness and biomass yield will be negatively correlated. After two growing seasons, the DD genotypes developed thicker bark compared to the genotypes of other taxa and bark thickness was not correlated with biomass yield in the DD genotypes (R2 = 0.002). However, in the TM, DM, and DN genotypes, bark thickness was negatively correlated with biomass yield (R2 = 0.33–0.77). Disease incidence studies revealed that the DM genotypes were most susceptible to canker whereas no disease was detected in DD genotypes. Furthermore, bark analysis conducted by Fourier transform infrared spectroscopy coupled with multivariate analysis showed that that DD genotypes to be chemically separate from the three hybrid genotypes and that bark chemistry was correlated with canker disease incidence. Taken together, these results reveal that it is possible to generate hybrid poplar genotypes with thicker bark, disease resistance, and higher biomass yields. This insight should guide further efforts to develop genetically improved hybrid poplar genotypes, both in terms of biomass yield and disease tolerance, for cultivation in the southeastern United States. Hybrid poplar cultivation in southeastern United States is hindered by its susceptibility to stem cankers. We evaluated tradeoffs between yield and canker disease resistance in various hybrid poplar genotypes. After two growing seasons, the DD genotypes showed disease resistance and developed thicker bark that was chemically distinct from the other genotypes. Bark thickness was not correlated with yield in the DD genotypes but was negatively correlated with yield in the other genotypes. These results will guide the development of hybrid poplar genotypes that are both disease resistant and high yielding for cultivation in the southeastern United States.


| INTRODUCTION
Hybrid poplar is an important biomass feedstock for wood, fiber, and bioenergy production owing to its fast growth rate and easy clonal propagation (Downing et al., 2011;Sannigrahi et al., 2010). However, in the southeastern and central United States, the establishment of short-rotation intensively cultured plantations of hybrid poplar has been hindered by its susceptibility to stem canker and leaf spot diseases caused by the fungal pathogen Sphaerulina musiva (Feau et al., 2010;Weiland et al., 2003). Canker infection of stems and branches cause lesions eventually leading to stem breakage, limb death, and tree mortality (Weiland & Stanosz, 2007). Therefore, the focus of various hybrid poplar breeding programs has been to identify and develop resistant genotypes (Downing et al., 2011;Qin & LeBoldus, 2014). Conventional breeding approaches have used field-and greenhouse-based evaluation of canker severity to ascertain the susceptibility or resistance of poplar genotypes Qin & LeBoldus, 2014;Weiland et al., 2003). Although effective in identifying resistant clones, it was found that the canker resistant genotypes were generally low yielding (Kim et al., 2021). Tradeoffs between yield and disease tolerance are common features to contend with in crop and biomass feedstock production systems (Ha et al., 2021;Huot et al., 2014). Therefore, detailed understanding of physiological and biochemical basis of disease resistance is needed to ascertain whether it is possible to simultaneously breed hybrid poplar for high disease resistance and biomass yield.
Studies on canker disease infection found that bark features differ among resistant and susceptible genotypes and that bark layers formed before and after infection were thicker in the resistant poplar clone compared to susceptible controls (Biggs, 1985;Blanchette & Biggs, 1992;Weiland & Stanosz, 2007). Furthermore, Qin and LeBoldus (2014) found that the pathogen enters the host tissues through lenticels, pores in woody stems that allow gas exchange between atmosphere and internal tissues, and resistant clones are able to thwart the spread of the pathogen through the formation of thicker bark layers. These detailed histological studies have provided important information about bark tissue formation on canker disease tolerance. Due to the nature of histological experiments, however, they were conducted in one resistant and one susceptible genotype and are not amenable to be incorporated into high-throughput hybrid poplar breeding under field conditions. Genetic and genomic approaches to understand poplarpathogen interactions have revealed that S. musiva fungus is able to successfully interfere with host immune response pathways in susceptible genotypes (Abraham et al., 2019;Lenz et al., 2021;Muchero et al., 2018). In resistant genotypes, the host plant is able to recognize the pathogen and initiate several defense pathways, such as cell wall strengthening, producing wall phenolics, and signaling to induce systemic immunity that potentially limits the spread of the fungus (Lenz et al., 2021;Muchero et al., 2018). It is likely that some of these phenolic compounds increased in the resistant genotypes are incorporated within bark tissues to limit pathogen infection as described by several histological studies (Biggs, 1985;Qin & LeBoldus, 2014;Weiland & Stanosz, 2007). Although these studies imply that bark structures are reinforced in response to pathogen attack, bark tissues are also constitutively produced as a structural barrier that protects the plants from biotic and abiotic factors (Franceschi et al., 2005;Paine et al., 2010;Poorter et al., 2014;Wainhouse et al., 1997). However, detailed analyses of bark morphological and chemical properties have not yet been conducted in hybrid poplar so far and we have no understanding of how these aspects impact disease resistance.
To rapidly evaluate bark features in hybrid poplar genotypes and address the importance of bark in canker resistance, we combined in-field biomass yield performance and disease incidence rating with morphological and chemical analyses of bark tissues in hybrid poplar genotypes. We evaluated hybrid poplar genotypes representing four different taxonomic designations [P. deltoides × P. maximowiczii (DM), P. deltoides × P. nigra (DN), P. trichocarpa × P. maximowiczii (TM), P. deltoides × P. deltoides (DD); different P. deltoides genotypes were used to obtain intraspecific poplar hybrids]. Eastern cottonwoods (P. deltoides) are generally considered to be resistant to canker (Newcombe, 1998;Ostry, 1985) and our previous with yield in the DD genotypes but was negatively correlated with yield in the other genotypes. These results will guide the development of hybrid poplar genotypes that are both disease resistant and high yielding for cultivation in the southeastern United States.

K E Y W O R D S
bark, disease, genotypes, hybrid poplar, taxa, yield study showed that pure eastern cottonwood genotypes had more bark, were disease resistant but were low yielding, whereas the DM genotype with low bark amount was high yielding but disease susceptible (Kim et al., 2021). These studies generally included one/few genotypes per taxon, making it difficult to assess whether the tradeoffs between yield and disease tolerance are associated with individual genotypes tested, or whether they can be more broadly applied to differences in taxa. In this study, we evaluated the tradeoffs between bark formation for defense against canker and biomass yield potential in 39 hybrid poplar genotypes belonging to four different taxa. We hypothesized that there is a strong negative correlation between bark thickness and biomass yield such that DD genotypes with thicker bark are disease resistant but low yielding whereas the elite DM genotypes with thinner bark are disease susceptible but high yielding. We found that there is no correlation with bark thickness and yield in DD genotypes, whereas negative correlation exists between bark thickness and yield in the genotypes belonging to the three other taxa studied. In addition, bark chemical features were distinct in DD genotypes compared to other taxa and were correlated with the absence of disease. The implications of these findings for genetics and inheritance of bark and breeding for both canker resistance and yield in hybrid poplar are discussed.

| Study site description and planting
The study site is located on the Mississippi Agricultural and Forestry Experiment Station Pontotoc Ridge-Flatwoods Branch Experiment Station (34°08′19.29″ N 88°59′41.51″ W) in Pontotoc, MS. In general, this site is a marginal upland site with severely eroded well drained soil with silty loam (until a depth of 8 cm) and silty clay loam (from 9 to 120 cm). Over 100 different poplar genotypes belonging to four different taxa were planted in blocks in April 2019. Unrooted cuttings of the DD genotypes were produced by Mississippi State University from the collection of eastern cottonwood that was started in 2012. Unrooted cuttings of the other hybrid poplar genotypes were produced by GreenWood Resources in Oregon. The unrooted cuttings were harvested in January 2019 and held in storage at −2°C until planted. The unrooted cuttings (23-46 cm long) were manually planted at 2.7 m × 1.8 m of inter-and intra-row spacing in a randomized complete blocked design with two cuttings/ per genotype and each block replicated eight times (for a total of 16 cuttings/genotype). Of these eight blocks, four were designated to be coppiced at the end of 2 years of planting and four blocks were designated to be harvested after 4 years of planting. Weed control at the time of planting was provided by spraying pre-emergence herbicides (pendimethalin [Pendulum 3.3 EC] at 4.7 L/ha and Oxyfluorfen -23% [Goal 2XL] at 4.7 L/ha) and postplanting by directed spray of Clethodim -12.6% ([Select] at 1.1 L/ha) between rows in the first year. Cottonwood leaf beetles and Japanese beetles were controlled using Imidacloprid ([Admire Pro] at 0.1 L/ha). Imidacloprid was applied from the ground during the early first year growth followed by aerial applications, when needed, due to the height attained by the trees.

| Field sampling, disease incidence, and specimen processing
In the four blocks designated for coppicing, the height and diameter at breast height (DBH) of 2-year-old trees were measured. After the assessment of the volumetric growth of trees grown, a subset of 39 genotypes showing the lowest, moderate, and highest biomass yield per taxa were selected for further characterization (Table 1). Although cankers can be caused by different fungal pathogens, hybrid poplars grown in the southeastern United States are particularly susceptible to cankers caused by S. musiva infection, which is endemic to this region (Feau et al., 2010;. If/when there was mechanical damage on the trees, mainly due to movement of weeding equipment, it was below 1 m of the trees and was recorded separately. For identification of stem cankers caused by S. musciva, we followed the protocol as described previously  with some modifications. Briefly, cankers were identified on the stem as a sunken portion of the bark, that can be identified by the weeping appearance on the stem. Cankers on the branches were identified by ruptured bark along a branch forming an expansion. The main stem and branches were examined to determine the presence of canker (and rated as 1) or T A B L E 1 Hybrid poplar genotypes and taxa grown as part of the study. absence of canker (and rated as 0). Thereafter, the main stems were harvested for each genotype in four blocks (n = 4 replicates) and stem disks (5.1 cm long) were collected at breast height (1.37 m). The diameter and weight of each stem disk were measured at the time of harvest. All stem disks were dried at 40°C for >3 days to reduce moisture content. In year 3, stem canker incidence was again recorded in both the coppiced and non-coppiced plots.

| Bark thickness measurements and correlation with yield
To quantitatively characterize the bark features of the different taxa (as listed in Table S1), bark thickness was measured in the stem disks in all the 39 genotypes (n = 4 replicates) as follows. Stem disks were imaged by placing the (transverse) cut surface on a flatbed scanner (Epson Perfection 4490) and the diameter of stem disks and thickness of bark at four locations were measured using ImageJ software (Abramoff et al., 2004) as shown in Figure 1. Bark thickness was then normalized based on the stem disk diameter by calculating percent bark (bark thickness/ stem disk diameter × 100).

| FTIR analyses of bark and wood samples
The chemical signature of the bark and wood samples were obtained by Fourier transform infrared (FTIR) spectroscopy. Stem disks were debarked using a 2.5 cm width hand planer and the separated bark and wood samples were each milled to a <40 mesh particle size with a Wiley Mill (Thomas Scientific). The chemical fingerprints of the bark and wood samples were collected using a Perkin Elmer Spectrum Two FT-IR spectrometer. About 250 μg of bark or wood samples were placed on an attenuated total reflectance (ATR) accessory and FTIR spectra over the range of 4000-600 cm −1 were collected with a spectral resolution of 4 cm −1 and eight scans per sample. Three spectra were collected for each replicate sample per genotype. Principal component analysis (PCA), a multivariate analysis, was performed on the FTIR spectral data to visualize any differences or similarities between the samples. Prior to the analysis, the spectra were mean normalized, the variables reduced by a factor 4, and a multiplicative scatter correction (MSC) was applied in The Unscrambler® X software version 9 (CAMO software).

| Py-GC/MS analysis
Detailed bark compositional analyses were performed on three DD (ST-70, DD 27-5, and 110412) and three DM (9707, 24265, and 24345) genotypes by pyrolysis gas chromatography mass spectrometry (Py-GC/MS). Approximately 200 μg of bark sample were pyrolyzed at 450°C for 12 s using a Frontier EGA/Py-3030 D pyrolyzer. Then, a Perkin Elmer Clarus 680 gas chromatograph coupled with a Clarus SQ 8C mass spectrometer was used for the chromatographic separation of the pyrolysis vapors and the identification of the evolved components, respectively. The GC column was 60 m length × 0.25 mm inner diameter × 0.25 μm film thickness (Agilent Technologies). Helium was used as the carrier gas at a 1 mL/min flow rate and an 80:1 split ratio was maintained. The GC furnace temperature was gradually increased from 50°C with a rate of 5°C/min to 280°C and then held at 280°C for 5 min. A total of 50 chromatographic peaks with S/N ≥ 2000 were extracted from the pyrograms using the TurboMass GC/MS software, then tentatively identified by comparing each of the mass spectra to the NIST library of fragmentation patterns from the Turbomass software. The peaks from the four replicates were averaged to obtain a representative chromatogram per genotype and differential peaks in the DD and DM genotypes were identified manually.

| Statistical analyses
To compare the differences among genotypes and taxa in bark features, DBH and tree height, analysis of variance (ANOVA) was conducted in SAS 9.4 (SAS Institute) at a significance level of p < 0.05 and means, when different, were separated using the Tukey test. The presence/ absence of disease was compared between taxa using ANOVA at a significance level of p < 0.05 in SAS 9.4. The strength of linear correlation between bark thickness versus stem disk diameter, DBH, and height of trees F I G U R E 1 Quantitative measurements of bark thickness. Schematic representation of bark thickness and stem diameter measurements (a) and representative image of stem cross section used for making the bark measurements (b) using ImageJ software.

| RESULTS
3.1 | Bark morphology, thickness, correlation with yield, and disease incidence in four hybrid poplar taxa After harvest, examination of stems revealed qualitative differences in bark morphology among the four hybrid poplar taxa. The DM, DN, and TM genotypes had smooth bark with characteristic white speckles, whereas the bark of DD genotypes was rough textured with ridges and furrows ( Figure 2). The color of the bark in the different taxa was also variable, with TM genotypes having lighter greyish bark, the DM and DN genotypes having charcoal grey bark with lighter speckles and DD genotypes was brown colored. We found that bark thickness was statistically higher in DD genotypes with average of 2.28 mm compared to the average ranging between 1.17 and 1.26 mm for the other taxa (Table S2) and that bark thickness was correlated with stem disk diameter in DD (R 2 = 0.54) and DM (R 2 = 0.66) genotypes ( Figure S1). To normalize bark amount within each stem disk, we calculated percent bark and compared the patterns within genotypes belonging to different taxa. The results showed that percent bark was also significantly higher in DD genotypes at an average of 4.57% compared to the average ranging between 2.76% and 3.43% in the other taxa ( Figure 3; Table S2).
Evaluation of DBH and height revealed that overall DD genotypes had statistically higher DBH, averaging 4.32 cm compared to the average of 3.5 cm in DN and 2.83 cm in TM genotypes, whereas the DD, DM, and DN genotypes had similar tree heights, ranging between 4.9 and 5.1 m, and higher height than the TM genotypes averaging 4.0 m (Figure 4; Table S2). Correlation analyses were conducted to compare percent bark with DBH and tree height data. Since the results were similar among these comparisons, only the percent bark and DBH correlation data are discussed in detail. Overall, across all 39 genotypes, there was no correlation between percent bark and DBH ( Figure S2). However, when the percent bark and DBH were compared separately among taxa, interesting patterns emerged ( Figure 5). There was a strong negative correlation between percent bark and DBH in TM genotypes (R 2 = 0.78), a moderate negative correlation between percent bark and DBH in DM (R 2 = 0.56) and DN (R 2 = 0.34) genotypes and no correlation between percent bark and DBH in DD (R 2 = 0.002) genotypes. Similar results were obtained when correlation between percent bark and stem disk diameter ( Figure S2b) and percent bark versus tree height ( Figure S2c) were compared.
Analysis of stem canker incidence revealed that there were no diseased trees among the DD genotypes (0/48 trees), whereas 18% of the trees (11/45) were diseased among the DN genotypes, 36% of the TM trees (8/22) were diseased, and about 42% of the DM trees (15/36) were diseased (Figure 6a). In year 3, canker incidence was evaluated on the stems that regenerated from the trees harvested in year 2 (coppiced trees) and the results showed that there was no disease detected in the 48 trees belonging to the DD F I G U R E 2 Morphology of bark in the different hybrid poplar taxa. After 2 years of growth, the main stems of trees were harvested and the differences in bark morphology in DD (a), DM (b), TM (c), and DN (d) hybrid poplar genotypes was recorded.
genotypes, whereas there was 11% (5/45), 39% (14/23), and 47% (18/34) diseased stems in DN, TM, and DM genotypes (Figure 6b). Stem canker incidence was also evaluated in 3-year-old non-harvested trees (non-coppiced) that were clones of the individual trees as shown in Figure 6a and the results showed that there were no diseased trees among the 45 trees of DD genotypes, whereas 13% (6/45), 50% (11/22), and 54% (16/35) of trees were diseased among the DN, TM, and DM genotypes (Figure 6c). These results show that the DD genotypes, at this point of time, are completely resistant to canker infection and that the DN genotypes are moderately resistant to canker infection. The resistance of eastern cottonwoods (pure P. deltoides genotypes) and the susceptibility of the DM genotypes to stem canker has been previously reported (Feau et al., 2010;, however, to our knowledge, our study is the first report that correlates bark thickness with disease with taxonomic identity of genotypes. Overall, the results presented in Figures 2-6 indicate that there are quantifiable differences in bark morphology and thickness measurements between the four hybrid poplar taxa tested and that the DD genotypes showed unique bark characteristics and were superior in terms of biomass yield and disease tolerance compared to the other taxa.

| Bark chemistry and correlation with disease incidence in four hybrid poplar taxa
Given that previous analyses of bark differences revealed histological changes among resistance and susceptible F I G U R E 3 Variation of bark % among and between the genotypes of four hybrid poplar taxa. Identity of individual genotypes and order within each taxa is as shown in Table S1. Bark % values are means (n = 4) per genotype and error bars represent standard error values. Grey line indicates the lowest mean value of bark % in the DD genotypes. Letters indicate significant difference (p < 0.05) between taxa.

F I G U R E 4 Average values of DBH (a) and tree height (b)
among and between the genotypes of four hybrid poplar taxa. Values are means (n = 4) per genotype and error bars represent standard error values. Identity of individual genotypes and order within each taxa is as shown in Table S1. Grey line indicates the lowest mean value in the DD genotypes. Letters indicate significant differences in DBH and height (p < 0.05) between the four hybrid poplar taxa.

F I G U R E 5
Correlation between bark % and diameter at breast height (DBH; cm) in the different hybrid poplar taxa. After 2 years of growth, the DBH of the main stems of trees were measured, the tree was harvested and the bark % of the main stem was calculated (see materials and methods for details).
hybrid poplar genotypes (Biggs, 1992;Qin & LeBoldus, 2014;Weiland & Stanosz, 2007), we wanted to evaluate whether, in addition to changes in bark morphological features, hybrid poplar genotypes differed in bark chemistry. We elected to analyze the chemical changes in the bark tissues using FTIR spectroscopy since it is a rapid and easy technique that enables non-destructive chemical screening of plant tissues (Alonso-Simón et al., 2011;Carpita et al., 2001;Straková et al., 2020). The spectra obtained from the bark were analyzed using PCA and the variance in spectra was visualized in the score plot ( Figure 7a) where principal component (PC) 1 represented 54% of the variation and PC2 represented 18%. The results revealed that DD genotypes chemically separated from the genotypes of the three other hybrid poplar taxa by both PC1 and PC2 (Figure 7b,c; Table 2). PC1 was largely defined by the absorbance peaks at 1031 cm −1 , assigned to C-O stretching, that are usually present in carbohydrate fractions (Bhagia et al., 2022;Horikawa et al., 2019) and the bands at 783 cm −1 , which could be tentatively assigned to PO 2 stretching, present in phospholipid fractions in the samples (Nzai & Proctor, 1998;Yoshida & Sakai, 1973) as well as 1319 and 1631 cm −1 , assigned to ring vibration and COO− antisymmetric stretching of ester groups (Kim & Daniel, 2018;Szymanska-Chargot & Zdunek, 2013). PC2 was defined by several peaks, including the positive intense bands at 1043 cm −1 , assigned to C-O or C-C stretching from carbohydrate fractions and 1127 cm −1 , assigned to C-O-C stretching and aromatic C-H deformation (Faix, 1991;Rodrigues et al., 1998), and the negative band at 959 cm −1 attributed to C-O bending present in the samples. While it is not possible to assign these FTIR bands to specific constituents of the bark material, these results suggest that FTIR spectroscopy can be effectively used to chemically differentiate bark in hybrid poplar genotypes and that the bark of the DD genotypes is compositionally distinct from the other taxa.
To further evaluate the differential bark chemical features, we selected three DM and three DD genotypes, which showed the largest differences in the PCA score plot (Figure 7a Kline et al., 2020;Sykes et al., 2008). These phenolic compounds found in the three DD genotypes are known to be incorporated within the lignins (Sykes et al., 2008) and are likely involved in restricting disease spread in resistant genotypes (Qin & LeBoldus, 2014;Weiland & Stanosz, 2007). Further studies are needed to evaluate whether these compounds are present at consistently high levels in all of the DD genotypes leading to the distinctly different chemistry of the DD genotypes and whether they are causally related to their stem canker resistance.

F I G U R E 6
Incidence of stem canker disease was assessed in different hybrid poplar taxa. The incidence of canker on the stems was evaluated in 2-year-old (a), 3-year-old coppiced (b), and 3-year-old non-coppiced (c) trees. The coppiced trees were the same individual trees as described in (a) whereas the non-coppiced trees were clones of individual coppice trees. Letters indicate significant differences in disease incidence (p < 0.05) between the four hybrid poplar taxa.

| Wood composition differences among DD and DM genotypes
Several studies have been conducted to analyze chemical differences within genotypes of Populus species and the consensus in the literature is that biomass chemical characteristics are not strongly correlated with genotype ( Verlinden et al., 2013). However, our results indicate that the pedigree of the genotype is an important determinant in bark composition (Figure 7). Given that most studies of Populus species were conducted with debarked tissues (Happs et al., 2021;Verlinden et al., 2013), we were unable to ascertain whether the observed genotype-specific bark compositional differences (Figure 7) were due to the tissues analyzed (debarked vs. barked tissue) or due F I G U R E 8 Pyrograms obtained from the pyrolysis-GC/MS conducted with bark tissues of three DM and DD genotypes; The differential peaks with the m/z values and the phenolic compounds associated with the peaks are shown. Each peak from the four replicates was averaged to obtain a representative pyrogram per genotype. Identity of the compounds was designated by comparing m/z values from published studies (Kline et al., 2020;Sykes et al., 2008).

F I G U R E 9
Principal component analysis score plot of FTIR spectra collected on wood samples of the DD and DM hybrid poplar genotypes. Three spectra were collected for each replicate sample per genotype.
to differences between different hybrid poplar characteristics. To address these alternative possibilities, we conducted the FTIR analyses of wood samples from the DD and DM genotypes. We focused on this subset of genotypes since DM genotypes are considered the early-age high-yielding and disease sensitive genotypes whereas the DD genotypes are more disease tolerant and, in our study, also high yielding (Figures 5 and 6; Feau et al., 2010;Kim et al., 2021;Stanton et al., 2010). The PCA score plot revealed no clear separation between the FTIR spectra obtained from the wood samples of the DM and DD genotypes (Figure 9). These results indicate that there are no chemical differences between wood samples of the DM and DD genotypes as opposed to genotypespecific spectral differences observed in bark tissues (Figure 7a) of different hybrid poplar taxa.

| DISCUSSION
In both natural and agricultural systems, plants use their limited resources for either promoting growth or bolstering defense responses, depending on internal or external factors (Ha et al., 2021;Herms & Mattson, 1992;Romero & Bolker, 2008). The tradeoffs between allocation for growth or defense are important for the survival and reproduction of plant species and have profound ecological, agricultural, and economic implications (Huot et al., 2014;Paine et al., 2010). Understanding how plants balance these tradeoffs will enable us to identify traits that enhance plant fitness and develop effective breeding and engineering strategies to improve crop productivity. Our previous research revealed that there are tradeoffs between biomass yield and stem canker disease tolerance in hybrid poplar genotypes, and that these characteristics are correlated with bark thickness (Kim et al., 2021). To further understand how bark structure and chemistry impacts canker disease resistance and to test whether it is biologically feasible to overcome the tradeoffs and identify high-yielding genotypes with disease tolerance, we screened hybrid polar genotypes for growth, biomass yield, and canker incidence. We found that the DD genotypes had visually distinct and thicker bark compared to the DN, DM, and TM genotypes (Figures 2 and 3). In addition, the DD genotypes had no correlation between bark thickness and biomass yield, whereas the DN, DM, and TM genotypes had moderate to strong negative correlation between bark thickness and yield ( Figure 5). Furthermore, DD genotypes showed no canker development ( Figure 6). Collectively, these data indicate that there is wide variability among the various taxa, in terms of yield and bark thickness and that there is potential to make significant gains in selection, improve diversity in hybrid poplar breeding, and identify genotypes that are both high yielding and disease resistant in hybrid poplar. Several studies have indicated that resistant genotypes of hybrid poplar develop thicker bark constitutively, and reinforce it in response to pathogen infections (Biggs, 1992;Kaufert, 1937;Lenz et al., 2021;Qin & LeBoldus, 2014;Weiland & Stanosz, 2007). The constitutive and induced production of bark is considered as an expensive trait that requires allocating carbon resources to defense at the expense of wood production (Paine et al., 2010;Romero & Bolker, 2008). However, when the allocation to defense enables the plant to withstand pathogen attacks, the plant is able to survive in unfavorable conditions and eventually grow/reproduce further (Feau et al., 2010;Huot et al., 2014). Since the progenitors of the DD genotypes were native to wet and humid ecosystems in the southeast region of the United States where canker disease is prevalent (Feau et al., 2010;Nelson & Tauer, 1987;Stanton et al., 2010), traits that facilitated balancing the tradeoffs between defense and growth were potentially favored. These traits were phenotypically expressed in the intraspecific DD hybrids that were developed from crossing the pure P. deltoides (D) genotypes. We were able to evaluate DD genotypes that were similar to the genotypes from the interspecific hybrids (DM, DN, and TM) in expressing their hybrid vigor, and then we were able to effectively compare their bark features, yield, and disease resistance. Since there was no correlation between bark features and yield in the disease resistant DD genotypes, we conclude that it is possible to balance tradeoffs between defense and yield and to obtain genotypes with thicker bark, disease resistance, and higher biomass yields in hybrid poplar. Further studies are needed with other DD genotypes in different locations to assess whether these characteristics of thicker bark, higher yields, and disease resistance in DD genotypes are heritable across various environments. In the long term, these studies will enable us to develop genetically improved hybrid poplar genotypes, both in terms of biomass yield and disease tolerance, for cultivation in the southeastern United States.
In addition to increased bark thickness, the DD genotypes were also found to have chemically distinct bark compared to the DM, DN, and TM genotypes. The FTIR and Py-GC/MS data indicated increased accumulation of phenolic compounds in the DD genotypes in the six genotypes analyzed. However, there were no chemical differences in the wood tissues obtained from the DD and DM genotypes, strongly indicating that there is not overall change in composition in the wood tissues of the DD genotypes and the compositional changes observed were specific to bark tissues. These results are not surprising since wood tissues generally consist of mature xylem with tracheid and vessel elements, that are functionally dead at maturity (Biggs, 1992). Whereas, bark tissues consist of both dead outer layers, composed of lignified and suberized cells, and living inner layers consisting of functional vascular cambium, phloem, phelloderm, and cork cambium (Biggs, 1992). Therefore, it is reasonable to observe more changes in chemical composition of bark tissues than wood tissues among the resistant and susceptible hybrid poplar genotypes. Several previous studies have shown that bark tissues formed in non-diseased and diseased trees accumulate phenolic compounds and are necessary for disease/pest resistance (Franceschi et al., 2005;Qin & LeBoldus, 2014;Tegelaar et al., 1995;Weiland & Stanosz, 2007). Metabolite analyses conducted with hybrid poplar genotypes inoculated with S. musiva also revealed increased accumulation of phenolics and lignin in resistant lines (Lenz et al., 2021;Zhang et al., 2018). Detailed chemical analyses of bark composition and its association with canker disease resistance are needed to ascertain whether the mechanism of increased accumulation of phenolics in bark tissues is causally associated with increased canker resistance across all the DD genotypes. A better understanding of the mechanisms involved in disease tolerance when combined with genetic and genomic tools available in hybrid poplar will enable us to rapidly identify genes associated with balancing the tradeoffs between biomass yield and disease resistance and to develop genetically improved hybrid poplar with both higher yield and disease resistance simultaneously.
In our analysis of bark morphological and compositional features in interspecific and intraspecific hybrid poplar genotypes, we observed that the inheritance of bark features is dependent on taxa identity. In particular, we observed that only the DD genotypes had brown bark that was thicker and chemically different to the other studied taxa. Whereas, the interspecific DM and DN hybrids, had dark gray and smooth bark, even though these genotypes inherited 50% of their genetic material from their D (P. deltoides) parent. This suggests that inheritance of bark and disease resistance in the interspecific hybrid poplar genotypes is likely recessive. Remarkably, Newcombe and Ostry  had previously suggested that the inheritance of disease resistance from the P. deltoides parent is probably recessive. However, they were unable to conclusively determine the inheritance structure because of the long duration of growing the hybrid poplar (>5 years of trials) and variable weather causing damage to the trees that was unrelated to canker disease . Given that most of the hybrid poplar breeding programs utilize cuttings from the F 1 generation (first progeny after crossing two pure lines) for further propagation and screening (Stanton et al., 2010), it is unlikely to obtain any disease resistant hybrid poplar genotypes if inheritance of resistance is recessive. Therefore, further research is critically needed to examine whether the phenotypic expression of disease resistance inherited from the D genome is recessive and if there are multiple alleles that may control resistance or susceptibility in the different genotypes. These trials need to include F 2 progeny or P.deltoides backcross progeny of the interspecific crosses since phenotypic expression of recessive alleles is possible in the F 2 or backcross generations . Furthermore, if it is conclusively determined that bark features are causally related to disease tolerance phenotypes, bark morphology could be used as a screening tool to identify resistant hybrid poplar genotypes. This will be advantageous compared to screening directly for disease resistance, since canker incidence is climate dependent and variable across locations and time (Feau et al., 2010;Stanton et al., 2010). Overall, our results demonstrate the utility of high-throughput methods to evaluate bark features in hybrid poplar genotypes and the complexity of interspecific and intraspecific breeding of hybrid poplar for identifying canker disease resistance and enhanced productivity in the southeastern United States.

| CONCLUSIONS
Our analysis of bark morphological and chemical features in taxonomically distinct hybrid poplar genotypes is the first study, to our knowledge, that demonstrated the differential impact of these bark features on biomass yield and canker disease resistance. We showed that the bark tissues of the DD genotypes have distinct morphological and chemical properties compared to the bark tissues of DM, DN, and TM hybrid poplar genotypes. We found that bark thickness was not correlated with yield in the DD genotypes but negatively correlated with yield in the TM, DM, and DN genotypes. Furthermore, we observed that DD genotypes were resistant to canker compared to genotypes of other taxa and this resistance was correlated with bark composition. Taken together, these results reveal that it is possible to obtain genotypes with thicker bark, disease resistance, and higher biomass yields. Future research is needed to quantify the differential bark composition within hybrid poplar genotypes and ascertain whether specific compounds in the DD genotypes are causally related to canker disease resistance. A better understanding of the mechanisms involved in disease resistance and their mode of inheritance in hybrid poplar will guide further efforts to develop genetically improved hybrid poplar genotypes, both in terms of biomass yield and disease tolerance, for cultivation in the southeastern United States.