Average leaf (solid line) and root (dotted line) Ct-values of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida 4 to 15 months after planting. Values on the y-axis are reversed to illustrate the inverse relationship of Ct-values and CLas titers. *, *** Significant differences between tissues at P < 0.05 and P < 0.001, respectively. No asterisks are shown when P > 0.05. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Average leaf and root starch content of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida. *, **, *** Significant differences between tissues at P < 0.05, P < 0.01, and P < 0.01, respectively. No asterisks are shown when P > 0.05. T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a strong positive correlation of root Ct-values obtained with the TXCChlb and the HLBaspr primer/probe sets.
Percentage of Candidatus Liberibacter asiaticus (CLas)-positive citrus trees with different scions and rootstocks growing in an huanglongbing-endemic environment in southwest Florida based on leaf and root analysis 4–15 months after planting. Trees were defined as CLas-positive when Ct ≤ 32. Percentages are shaded based on their values from blue to red, with darkest blue indicating the lowest percentages and darkest red highest. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a moderate inverse correlation of leaf Ct-values and leaf starch content 15 months after field planting.
Click on author name to view affiliation information
The citrus industry in Florida faces a destructive endemic disease, known as huanglongbing (HLB), associated with Candidatus Liberibacter asiaticus (CLas), a phloem-limited bacterium, and transmitted by the Asian citrus psyllid (ACP). Rootstocks are regarded as critical to keep citrus production commercially viable and help trees cope with the disease. Although most scions are susceptible, some rootstocks are HLB-tolerant and may influence ACP infestation and CLas colonization and therefore the grafted tree tolerance. This study aimed to elucidate the relative influence of rootstock and scion on insect vector infestation and CLas colonization under natural HLB-endemic conditions. Seven commercial rootstock cultivars with different genetic backgrounds were grafted with ‘Valencia’ sweet orange (Citrus sinensis) or were self-grafted (non-‘Valencia’) and planted in an open field where ACP and CLas were abundant. ACP infestation was determined weekly during periods of leaf flushing, and leaves and roots were analyzed every 3 months to determine CLas titers. Trees with ‘Valencia’ scion were more attractive to the psyllids than non-‘Valencia’ scions. This was also associated with a higher number of bacteria and a larger abundance of foliar HLB symptoms. The influence of the rootstock on the psyllid attraction of grafted ‘Valencia’ scion was less evident, and leaf CLas titers were similar regardless of the rootstock. Among the non-‘Valencia’ scions, Carrizo had the lowest and US-942 the highest leaf CLas titers. Root CLas titers also varied among cultivars, and standard sour orange roots harbored more bacteria than some trifoliate orange hybrid rootstocks such as US-942. In some trees, CLas was detected first in the roots 4 months after planting, but root CLas titers remained low throughout the study. In contrast, leaf CLas titers increased over time and were considerably higher than root titers from 7 months until the end of the study, 15 months after planting. Overall, the results of this study demonstrate a greater relative influence of the scion than the rootstock on ACP infestation and CLas colonization during the early stages of infection. This suggests that other cultivar-specific traits, such as the ability to tolerate other stresses and to absorb water and nutrients more efficiently, along with influences on the scion phenology, may play a larger role in the rootstock influence on the grafted tree tolerance during the later stages of HLB progression.
Citrus huanglongbing (HLB or citrus greening) is one of the most destructive diseases of citrus worldwide and was discovered in the Western Hemisphere in 2005 (Gottwald 2010). In Florida, HLB is associated with the phloem-limited bacteria Candidatus Liberibacter asiaticus (CLas), which are transmitted to citrus trees by feeding of the Asian citrus psyllid (ACP) (Halbert and Manjunath 2004). HLB symptoms include chlorosis and blotchy mottling of the leaves, deformed and poorly colored fruit with low internal quality, premature fruit drop, and tree decline (Bové 2006; Gottwald et al. 2007; Wang and Trivedi 2013). Physiological disorders of HLB include phloem collapse and metabolic disturbances, leading to significant starch accumulation in the leaves (Albrecht and Bowman 2008; Brodersen et al. 2014; Etxeberria et al. 2009; Fan et al. 2010), and starch depletion in the roots (Achor et al. 2010; Aritua et al. 2013; Etxeberria et al. 2009; Kim et al. 2009).
Rootstocks can modulate the horticultural characteristics of the grafted tree and impart tolerance or resistance to abiotic and biotic stresses (Bowman and Joubert 2020; Goldschmidt 2014). Under HLB-endemic conditions, some rootstocks have been documented to enhance productivity (Bowman et al. 2016a, 2016b; Caruso et al. 2020; Girardi et al. 2021; Kunwar et al. 2021, 2023) and are generally regarded as imperative for a viable citrus industry. To improve our understanding of the relative influence of rootstock and scion on HLB progression in grafted citrus trees, ACP preference and CLas distribution need to be considered, especially during the early stages of the disease. HLB affects all citrus species and citrus relatives with little known resistance. Sweet oranges (Citrus sinensis) and mandarins (Citrus reticulata) are among the most susceptible species, evidenced by high bacterial titers and disease-induced foliar aberrations and tree decline (Folimonova et al. 2009; Lopes and Frare 2008; McClean and Schwarz 1970). Trifoliate orange (Poncirus trifoliata) and hybrids of trifoliate orange, frequently used as rootstocks, are among the more tolerant species when grown on their own (Albrecht and Bowman 2012; Folimonova et al. 2009; Ramadugu et al. 2016). Possible reasons for their higher tolerance include differences in metabolites, flavonoids, and phytohormones (Albrecht et al. 2020a; Hijaz et al. 2016; Killiny and Hijaz 2016; Killiny et al. 2018).
The ACP is more attracted to new leaf flush than to mature leaves (Hall and Albrigo 2007; Hall et al. 2016). Therefore, young trees, which flush more frequently during favorable weather conditions, are more vulnerable to CLas inoculation. Results from laboratory and greenhouse studies suggested that ACP feeding habits are determined by the morphology and chemical composition of the leaves. For example, Patt and Sétamou (2010) identified mixtures of volatiles in young shoots of different citrus species that can attract psyllids. Under no-choice conditions in the laboratory, trifoliate orange was found less vulnerable to oviposition and nymph development than sweet orange (Hall et al. 2015). In contrast, less is known regarding the host preference of the insect when they have the freedom to choose among different citrus species in an open field environment.
It is well established that despite systemic infection with CLas, the distribution of the pathogen is not uniform throughout the tree (Folimonova et al. 2009; Kunta et al. 2014; Li et al. 2009; Tatineni et al. 2008; Teixeira et al. 2008), especially during the early asymptomatic stages of disease progression, rendering an accurate diagnosis difficult. CLas can move to the roots, and fibrous root decline was even suggested as one of the primary events following infection (Johnson et al. 2014). However, recent greenhouse studies showed that root CLas titers are considerably lower than leaf titers and that the HLB disease severity is usually determined by the scion rather than the rootstock (Albrecht and Bowman 2019; Bodaghi et al. 2022; Bowman and Albrecht 2020). Nevertheless, some studies have suggested using roots instead of leaves to accurately diagnose the status of infection (Braswell et al. 2020; Louzada et al. 2016).
In this context, the main objective of this study was to determine the relative influence of rootstock and scion on ACP infestation and CLas colonization during the early stages of infection under natural HLB-endemic conditions. We used seven rootstock cultivars that were either grafted with ‘Valencia’ or self-grafted. ACP preference and infestation during flushing time, CLas colonization in leaves and roots, and disease progression were assessed over 15 months of growth in an HLB-endemic environment.
A field trial was established in Mar 2019 at the Southwest Florida Research and Education Center in Collier County, FL (26.46095, –81.43551) and conducted over 15 months. One-year-old grafted trees were planted in two double-rows on raised beds separated by furrows. The distance between the rows was 22 ft (6.7 m), and trees in each row were spaced at 8 ft (2.44 m). Seven commercial rootstock cultivars were used: Carrizo citrange [Citrus sinensis × Poncirus trifoliata (trifoliate orange)], ‘Cleopatra’ mandarin (C. reticulata), standard sour orange (C. aurantium), Swingle citrumelo [‘Duncan’ grapefruit (C. ×paradisi) × trifoliate orange], US-897 (‘Cleopatra’ mandarin × ‘Flying Dragon’ trifoliate orange), US-802 [‘Siamese’ pummelo (C. maxima) × ‘Gotha Road’ trifoliate orange], and US-942 (‘Sunki’ mandarin × ‘Flying Dragon’ trifoliate orange). In the most recent citrus production years in Florida, these rootstocks were among the top 15 most propagated rootstocks (Florida Department of Agriculture and Consumer Services, Division of Plant Industry 2022). ‘Valencia’ is generally not used as a rootstock in Florida because of its high sensitivity to Phytophthora (Castle 2010; Wutscher and Hill 1995), which is prevalent in the poorly drained Spodosol-type soils of southwest Florida (Mylavarapu et al. 2019); it was therefore not included to create reciprocal graft combinations or as a self-grafted plant. All rootstocks were grown from seeds as described in Albrecht et al. (2020b). Plants were grown in the US Department of Agriculture Horticultural Research Laboratory (USHRL) greenhouses in Fort Pierce, FL.
Once the rootstock seedlings reached a suitable size for grafting (4- to 6-mm stem diameter), they were transplanted into 12.7-cm × 24.1-cm plastic tree pots (Stuewe and Sons, Tangent, OR, USA) filled with Pro-Mix BX potting medium (Pro-Mix BX; Premier Horticulture, Inc., Quakertown, PA, USA). After an acclimatization period, rootstock liners were budded with certified disease-free ‘Valencia’ orange (clone 1-14-19) or with themselves (self-grafts) using the inverted T method (Albrecht et al. 2021). Plants remained in the greenhouse under natural light conditions and were fertilized every 2 weeks at a rate of 400 mg N/L; water and insecticides were applied as needed.
The experimental design was a randomized split-plot design with rootstock cultivar as the main plot factor and graft type (‘Valencia’ and non-‘Valencia’) as the subplot factor (7 × 2), and 10 single-tree replications. Irrigation was by under-tree microjets. Conventional granular fertilizer (8N–4P–8K; Diamond R, Fort Pierce, FL, USA) was applied at a rate of 0.5 lb (0.23 kg) per tree every 6 months, and slow-release fertilizer (12N–8P–6K, Diamond R) was applied at the same rate three times per year. No insecticides were applied to allow psyllid infestation.
Upon planting, trees were lightly pruned to induce new flush to attract the ACP vector and encourage feeding and CLas colonization. ACP infestation was determined weekly during the flushing periods in Summer 2019 (3 weeks) and Spring 2020 (8 weeks) when the insects were most active. Adult ACP infestation was determined on 10 randomly selected shoots per tree and counting the number of adults present on each shoot and expressed as the number of ACPs per shoot. Immature ACP infestation was determined by inspecting the same 10 shoots selected for adult ACP infestation for the presence of ACP eggs and nymphs and expressed in percent infestation (Monzo et al. 2015; Stansly et al. 2010).
Starting in Jul 2019 (4 months after planting), leaves and roots were collected randomly from each tree every 3 months (except for Mar 2020 when samples were collected 1 month earlier in anticipation of the closing of facilities due to COVID-19). Four to six mature leaves from the most recent flush were collected from different areas in the canopy and stored at –20 °C until analysis. Fibrous roots (<2 mm in diameter) were collected from different areas of the root system, washed, blotted dry, and stored at –20 °C until analysis. Tissues were pulverized with liquid nitrogen using a mortar and pestle. For DNA extraction of leaves and roots, the Plant DNeasy Mini Kit (Qiagen, Valencia, CA, USA) and the Plant DNeasy Pro-Kit (Qiagen), respectively, were used according to the manufacturer’s instructions. Real-time polymerase chain reaction (PCR) assays were performed to detect CLas in the leaves and roots using primers HLBas, HLBr, and probe HLBp developed by Li et al. (2006). Primers COXf, COXr, and probe COXp (Li et al. 2006) were used for internal control and normalization. To validate the accuracy of CLas detection in the roots, the TXCChlb primer/probe set recommended by Park et al. (2018) for root detection in pre-symptomatic trees was used for real-time PCR detection. Both primer/probe sets amplify a 16S rDNA region of CLas. Amplifications were performed over 40 cycles using the iTaq Universal Probes supermix (Bio-Rad, Hercules, CA, USA) and a QuantStudio 3 real-time PCR system (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions. If no amplicon was detected after 40 cycles, a Ct-value of 41 was assigned for calculating average Ct-values.
For starch determination, the same leaf and root tissue collected for CLas detection in Oct 2019, Jan 2020, Mar 2020, and Jun 2020 were used. One-hundred and fifty milligrams of ground tissue were extracted twice in 1 mL of 80% ethanol for 1 h at 70 °C and centrifuged for 5 min at 20,000 gn. Insoluble pellets were used for starch determination. Pellets were dried using an Eppendorf vacufuge concentrator (Thermo Fisher Scientific, Waltham, MA, USA) and resuspended in 900 μL of ultrapure water. To break up the starch, samples were autoclaved for 1 h at 121 °C and 19 psi. Nine hundred μL of sodium acetate buffer (0.1 M, pH 4.65) was added together with 5 μL (14 units) of amyloglucosidase (Sigma-Aldrich, St. Louis, MO, USA), and samples were incubated overnight at room temperature. After centrifugation for 5 min at 20,000 gn, supernatants were diluted 10-fold and used for starch determination. Starch was measured indirectly by enzymatic assay of released glucose (Campbell et al. 1999; Gomez et al. 2007).
Scion and rootstock trunk circumferences were measured at the end of the study at 5 cm above and below the graft union, and the trunk diameters were calculated based on the formula, where d = diameter and c = circumference: .
Foliar HLB disease symptoms were assessed at the end of the study by visual ratings using a scale of 1 to 5 and expressed as HLB disease index. An HLB disease index of 1 represented the best (no HLB leaf symptoms) and 5 represented the worst (76% to 100% of the leaves displayed HLB symptoms). Indices of 2, 3, and 4 represented 1% to 25%, 26% to 50%, and 51% to 75% of the leaves, respectively, displaying HLB symptoms. HLB disease symptoms were defined as irregular blotchy mottling and/or chlorosis of the leaves typical for HLB (Gottwald et al. 2007). Two ratings on opposite sides of each tree were performed and expressed as averages for statistical analysis.
For all variables, except foliar HLB disease ratings, an analysis of variance (ANOVA) was conducted using R version 4.0.3 (R Core Team, Vienna, Austria, 2019). A linear mixed model was used with scion and rootstock cultivar as fixed factors and block as a random factor. Scion effects were analyzed by comparing ‘Valencia’ with non-‘Valencia’ (self-grafted) trees and by comparing self-grafted trees. Rootstock effects were analyzed for ‘Valencia’ trees. Differences were also compared across all 14 graft combinations. Tukey’s honestly significant difference test was used to determine means separation. Differences were defined as statistically significant when the P value was < 0.05. Before analysis, all data sets were tested for normality and homogeneity of variance of the residuals. HLB disease indices were analyzed by nonparametric aligned rank transformation ANOVA using ARTool (Wobbrock et al. 2011). The relationships between root Ct-values measured with the two primer/probe sets, between leaf and root Ct-values, and between Ct-values and starch content were analyzed using Pearson’s correlation coefficient.
In Summer 2019, trees were uniformly flushing from the middle of July to the first week of August (3 weeks). In Spring 2020, trees were uniformly flushing from the beginning of February to the third week of March (8 weeks).
There was a significant effect of the scion on the ACP adult infestation. In Summer 2019, adult infestation was nearly twice as high for ‘Valencia’ (6.9 adults per shoot) than for non-‘Valencia’ scions (3.5 adults per shoot) (Table 1). A significant difference was also found in Spring 2020, when 11.0 adults per shoot were found on ‘Valencia’ and 4.4 on non-‘Valencia’ scions. However, the adult infestation did not vary significantly among the non-‘Valencia’ scions, and there was no significant influence of the rootstock on the ‘Valencia’ scion. Across all scion–rootstock combinations, in Summer 2019, the largest number of adults per shoot was found for ‘Valencia’ grafted onto US-942 and the lowest for self-grafted US-942 trees, but there were no differences in Spring 2020 (Supplemental Table 1).
The scion effect was also significant for the immature ACP infestation rate. In Summer 2019, the infestation rate on ‘Valencia’ was significantly higher (30.4%) than on non-‘Valencia’ scions (20.4%) (Table 1). The same trend was found in Spring 2020, when ‘Valencia’ had an infestation rate of 20.8%, whereas non-‘Valencia’ scions had a rate of 14.9%. Among the non-‘Valencia’ scions, significant differences were found in both years. In Summer 2019, ‘Cleopatra’ had a higher immature infestation rate (42.5%) compared with most of the other non-‘Valencia’ scions (11.9% to 17.8%). In Spring 2020, sour orange had the highest oviposition rate (30%), followed by ‘Cleopatra’ (21.1%), whereas the lowest rate was found for US-942 and Carrizo (4.9% and 5.5%, respectively). The rootstock cultivar influenced the immature infestation rate of the grafted ‘Valencia’ scion only in Spring 2020 when US-802 induced a higher infestation rate (30.2%) than US-897 (12.9%). Across all scion–rootstock combinations, in Spring 2020 ‘Valencia’ grafted on US-802 and self-grafted sour orange trees experienced a significantly higher infestation rate (30.2% and 30.2%, respectively) than self-grafted US-942 and Carrizo trees (5.0% and 5.5%, respectively) (Supplemental Table 1).
Trees with ‘Valencia’ scion had significantly lower leaf Ct-values, and therefore more leaf bacteria, than trees with non-‘Valencia’ scions in Oct 2019 (T2), Jan 2020 (T3), Mar 2020 (T4), and Jun 2020 (T5) (Table 2). At the end of the study (T5), Ct-values were 25.2 on average for ‘Valencia’ and 29.6 for non-‘Valencia’ scions. Among the non-‘Valencia’ scions, significant differences were only found at the end of the study when Carrizo had a higher Ct-value (37.7) and therefore less leaf bacteria than the other non-‘Valencia’ scions (26.0–29.7) except US-897. There was no significant rootstock effect on the leaf Ct-values of the ‘Valencia’ scion. Across all scion−rootstock combinations there were no significant differences in leaf Ct-values, except at the end of the study when self-grafted Carrizo trees had a significantly higher Ct-value, than trees with ‘Valencia’ grafted onto Carrizo and onto US-942 (Supplemental Table 2).
The two primer/probe systems used for root detection of CLas yielded equivalent results with a strong correlation of Ct-values (R = 0.93, P < 0.0001) (Supplemental Fig. 1). Significantly higher Ct-values (37.9), and therefore less bacteria, were found in the roots of non-‘Valencia’ trees than in the roots of ‘Valencia’ trees (Ct = 33.4) in Jul 2019 (T1), 4 months after planting (Table 2), but there were no significant differences from Oct 2019 to the end of the study. At the end of the study, average root Ct-values were 33.9 and 34.6 in non-‘Valencia’ and ‘Valencia’ trees, respectively. Among the non-‘Valencia’ trees, significant differences for root Ct-values were observed from Jan 2020 to the end of the study. In Jan 2020 (T3) Swingle roots had a significantly lower Ct-value (Ct = 30.7) and therefore more bacteria than Carrizo roots (Ct = 38.8). In Mar 2020 (T4), Cleopatra, sour orange, and Swingle roots had significantly lower Ct-values (29.0–31.0) than Carrizo roots (Ct = 37.9). At the end of the study in Jun 2020 (T5), Carrizo and sour orange roots had the lowest Ct-values (31.7 and 32.2, respectively) and therefore the most bacteria, and US-802 and US-942 had the highest Ct-values (36.6 and 36.2, respectively) and therefore the least bacteria. Among the ‘Valencia’ trees, root Ct-values varied significantly among rootstocks in Mar 2020 and in Jun 2020. In Mar 2020 (T4) sour orange roots had a significantly lower Ct-value (29.9) and therefore more bacteria than US-942 roots (Ct = 36.3). At the end of the study in Jun 2020 (T5), sour orange roots had a significantly lower Ct-value (30.1) and therefore more bacteria than Swingle, US-802, US-897, and US-942 roots (Ct = 35.1–36.9). Across all scion−rootstock combinations root Ct-values were significantly higher, and therefore CLas titers lower, in self-grafted Carrizo trees than in self-grafted Swingle trees in Oct 2019 and Jan 2020 (Supplemental Table 2). In Mar 2020, self-grafted Carrizo trees had a significantly higher Ct-value, and therefore less bacteria, than ‘Valencia’ trees grafted onto Carrizo and self-grafted Cleopatra trees.
In the ‘Valencia’ trees the average root Ct-values were significantly lower, and therefore CLas titers higher, than the leaf Ct-values in Jul 2019, 4 months after planting (Fig. 1). The percentage of trees identified as CLas-positive, at this time, was 57% based on root detection compared with 21% detected in leaves (Supplemental Fig. 2). However, from Oct 2019 until the end of the study in Jun 2020 (T2–T5) leaf Ct-values were significantly lower, and therefore CLas titers higher, in the leaves than in the roots. Leaf CLas titers increased over time, whereas root titers decreased, resulting in 91% CLas-positive plants and 23%, respectively, at the end of the study. A different trend was observed for the non-‘Valencia’ trees, where root Ct-values were significantly lower, and therefore CLas titers higher, than leaf Ct-values in Oct 2019 (T2) and Mar 2020 (T4), but higher at the end of the study in Jun 2020 (T5). At that time 73% of the non-‘Valencia’ trees were identified as CLas-positive based on leaf analysis and 29% based on root analysis.
Citation: HortScience 58, 4; 10.21273/HORTSCI17039-22
A weak but statistically significant correlation between leaf and root Ct-values (R = 0.38, P < 0.0001) was found in Jul 2019, whereas there was no significant correlation in Oct 2019, Jan 2020, and Mar 2020 (data not shown). At the end of the study in Jun 2020, the correlation coefficient was R = –0.27 (P = 0.0018) (data not shown).
Significantly more starch was found in ‘Valencia’ (4.3–12.2 µg·mg−1) than in non-‘Valencia’ leaves (3.1–11.0 µg·mg−1) in Oct 2019, Mar 2020, and Jun 2020 (Table 3). Significant differences were also found among the non-‘Valencia’ trees. In Oct 2019, Cleopatra and sour orange leaves contained more starch (6.1 µg·mg−1 and 5.4 µg·mg−1, respectively) than Carrizo and US-897 leaves (1.0 µg·mg−1 and 1.7 µg·mg−1, respectively). Similarly, in Mar 2020 Cleopatra and sour orange leaves had significantly more starch (7.6 µg·mg−1 and 6.2 µg·mg−1, respectively) than the other non-‘Valencia’ leaves (3.6–4.3 µg·mg−1). In Jan 2020, sour orange and US-802 had more leaf starch (13.0 µg·mg−1 and 14.1 µg·mg−1, respectively) than Carrizo (6.9 µg·mg−1). A similar trend was observed in Jun 2020, when sour orange, US-802, and Swingle had more leaf starch (13.3–14.4 µg·mg−1) than Carrizo (5.2 µg·mg−1). No significant rootstock effect on the leaf starch content of the ‘Valencia’ scion was found at any timepoint. There was a moderate inverse correlation (R = –0.66, P < 0.0001) between leaf Ct-values and leaf starch at the end of the study in Jun 2020 (Supplemental Fig. 2).
Across all scion–rootstock combinations significant differences for leaf starch were found in Mar 2020 and Jun 2020. In Mar 2020 all ‘Valencia’ trees as well as self-grafted Cleopatra and sour orange trees had significantly more leaf starch (6.2–7.6 µg·mg−1) than self-grafted Carrizo, US-802, and US-942 trees (3.6–4.0 µg·mg−1) (Supplemental Table 3). A similar trend was observed for Jun 2020 when most of the ‘Valencia’ trees and self-grafted sour orange trees had significantly more leaf starch than self-grafted Carrizo trees.
In Oct 2019, Jan 2020, and Mar 2020, the root starch content was significantly influenced by the scion. Less starch was measured in roots of ‘Valencia’ trees (0.7–1.6 µg·mg−1) than non-‘Valencia’ trees (1.1–2.5 µg·mg−1) (Table 3). Among the non-‘Valencia’ trees, significant differences in root starch were measured in Jan 2020 and Mar 2020. In Jan 2020, sour orange roots had the most starch (4.7 µg·mg−1) and US-802 the least (0.6 µg·mg−1). In Mar 2020, Cleopatra and sour orange roots had the most starch (2.1 µg·mg−1 and 1.9 µg·mg−1, respectively) and US-802 and US-897 had the least (0.2 µg·mg−1). There was a significant rootstock effect among the Valencia trees in Oct 2019 and Jan 2020. Both times Cleopatra roots had more starch (2.8 and 3.4 µg·mg−1, respectively) than Swingle, US-802, and US-942 roots (0.7–1.1 and 0.4–0.8 µg·mg−1, respectively). No significant rootstock effect was found for Mar 2020 and Jun 2020. There was a weak negative correlation (R = –0.24, P < 0.0045) between root Ct-values and root starch content at the end of the study (data not shown).
Across all scion–rootstock combinations significant differences were found only in Jan 2020 when self-grafted sour orange trees had significantly more root starch (4.7 µg·mg−1) than many of the other graft combinations (0.5–1.6 µg·mg−1) (Supplemental Table 3). In ‘Valencia’ trees, root starch content was significantly lower than leaf starch content; the same was observed for non-‘Valencia’ trees except in Oct 2020 (Fig. 2).
Citation: HortScience 58, 4; 10.21273/HORTSCI17039-22
After 15 months of field growth, ‘Valencia’ scions had significantly smaller trunk diameters (2.2 cm) than non-‘Valencia’ scions (2.7 cm) (Table 4). Among the non-‘Valencia’ scions US-802 had a significantly larger trunk diameter than Cleopatra. There was no significant rootstock effect on the trunk diameter of the grafted ‘Valencia’ scion. Rootstock trunk diameters did not vary significantly between ‘Valencia’ and non-‘Valencia’ trees and among non-‘Valencia’ or ‘Valencia’ trees.
HLB disease indices of ‘Valencia’ and non-‘Valencia’ trees were 2.9 and 2.5 on average, respectively (Table 4). Although differences were significant, indices indicated that trees were not yet severely affected in the early state of tree growth and development. HLB disease indices varied significantly among the non-‘Valencia’ trees. Self-grafted Cleopatra and sour orange trees had a higher index (2.9) than self-grafted US-942 trees (1.9). There was no significant rootstock effect on the disease index of the grafted ‘Valencia’ scion.
Seven citrus rootstocks were compared as grafted trees with ‘Valencia’ orange in the scion position and as self-grafted trees in a region with high ACP and HLB incidence to assess the relative influence of rootstock and scion on vector and disease colonization during the early disease stage.
Citrus trees have several new growth flushes each year, with the major flushes occurring under the climatic conditions of spring and summer (Dalal et al. 2013; Hall and Albrigo 2007). It has been demonstrated that the presence of new flush increases the presence of ACP and the probability of CLas infection (Hall et al. 2016) and that once established, the psyllids prefer to feed and reproduce on young leaves and shoots rather than on mature ones (Cifuentes-Arenas et al. 2018; Hall et al. 2011, 2016; Patt and Sétamou 2010; Wenninger et al. 2009). It was also shown that under controlled conditions using 8-month-old potted plants, the scion variety can affect feeding and biology of the ACP (Alves et al. 2018). However, less information is available for psyllid preference under natural field conditions.
We found that adult ACPs were more attracted to and preferred laying eggs on trees with ‘Valencia’ scion than on non-‘Valencia’ scions in both spring and summer flushing periods. Among the non-‘Valencia’ scions (self-grafted rootstocks), there were no measurable differences in adult infestation, but the immature infestation rate was generally higher on ‘Cleopatra’ and sour orange scions compared with the trifoliate orange (Poncirus trifoliata) hybrids, suggesting that Citrus species are more favorable hosts than Poncirus species for oviposition and development. This is similar to the findings of Westbrook et al. (2011), who identified less infestation by ACP on P. trifoliata seedlings under field conditions compared with several Citrus species. The different host preference of the ACP has also been reported in other studies, albeit mostly under controlled environmental conditions (Alves et al. 2021; Felisberto et al. 2019; Hall et al. 2015; Richardson and Hall 2014). Reasons suggested for the different host preferences of the ACPs include host-specific volatiles as well as differences in leaf color and morphology (Hall et al. 2011; Paris et al. 2015, 2017; Patt and Sétamou 2010; Wenninger et al. 2009). The leaves of the trifoliate hybrid cultivars used in our study are usually darker green and have a different morphology than ‘Valencia’, Cleopatra, and sour orange leaves, which could have contributed to them being less attractive to the ACPs. Another possible reason for the preference of the ACPs for ‘Valencia’ over non-‘Valencia’ scions in our study may be differences in flushing dynamics among the scions (de Carvalho et al. 2020), which was not evaluated here. The higher preference of ACPs for the ‘Valencia’ scion resulted in higher levels of foliar CLas titers at the end of the study, reinforcing the association of ACP infestation and CLas distribution. Similar to our results, Boava et al. (2015) found a considerably higher incidence of CLas infection in trees with ‘Pera’ sweet orange scion, regardless of the rootstock, than in neighboring trees with different citrandarin (mandarin × trifoliate orange) scions, although ACP colonization was not documented in that study. However, there was no clear association between the immature infestation rate and CLas titers within the non-Valencia scion cultivars, except for Carrizo, which was the least preferred cultivar.
The rootstock did not influence ACP preference for the ‘Valencia’ scion, and differences in the immature infestation rate were only induced by US-802 compared with US-897 in Spring 2020. Similarly, under controlled environmental conditions, Alves et al. (2018) did not find any influence of the rootstock on the feeding behavior of the ACP. A recent study by de Carvalho et al. (2021) found that the rootstock cultivar can influence the scion flush intensity. In that study, most graft combinations had a similar flushing frequency, but the intensity was higher on some of the more vigorous rootstocks, such as ‘Florida’ rough lemon (C. × limonia var. jambhiri), than on less vigorous ones, such as ‘Flying Dragon’ trifoliate orange. Similarly, Rodrigues et al. (2020) found a lower abundance of flush shoots of the ‘Valencia’ scion on three semidwarfing rootstocks and ‘Flying Dragon’ compared with more vigorous rootstocks. This might also explain the observed difference between US-802 and US-897 in our study: the former is known for its high vigor, whereas the latter, a ‘Flying Dragon’ hybrid, is known for is low vigor (Bowman and Joubert 2020). However, Rodrigues et al. (2020) suggested that other mechanisms in addition to tree vigor are involved in the host–vector relationship.
US-897, US-942, US-802, and Carrizo rootstocks have been classified as HLB tolerant and Cleopatra as HLB susceptible when grown as nongrafted trees (Albrecht and Bowman 2012). The results from this study suggest although these rootstocks might influence ACP preferences in the grafted ‘Valencia’ scion, they may not substantially reduce CLas colonization. The lack of a rootstock influence on CLas titers of the grafted scion was also found in a recent field study in Florida (Kunwar et al. 2023).
Aside from vector exclusion, inoculum removal is one of the best strategies to reduce the spread of HLB in areas where the disease is not yet endemic (Bassanezi et al. 2020). Because HLB can spread from pre-symptomatic trees to other trees in the field (Lee et al. 2015), accurate diagnosis is important to decrease spread of the disease. Current strategies to identify HLB-affected trees typically rely on visual disease symptoms and real-time PCR analysis of leaves to detect CLas. However, PCR detection of CLas can be inaccurate because of the uneven distribution of the bacteria in the tree (Folimonova et al. 2009; Kunta et al. 2014; Li et al. 2009; Tatineni et al. 2008; Teixeira et al. 2008). Moreover, HLB symptoms may not be apparent until many months after infection, even under favorable infection conditions and are often confused with nutritional deficiencies (Folimonova et al. 2009; Gasparoto et al. 2012; Hung et al. 2001; Lopes et al. 2009). Effort has therefore been directed toward finding more effective methods of CLas detection such as the PCR analysis of roots instead of leaves (Braswell et al. 2020; Louzada et al. 2016; Park et al. 2018). The motivation for using roots for CLas detection may have been prompted by observations that CLas infection causes root decline before symptoms manifest in the canopy (Graham et al. 2013; Johnson et al. 2014).
The better suitability of using roots for CLas detection was not obvious in our field study where trees grew under high HLB pressure. Although CLas was detected at a higher percentage in the roots than in the leaves 4 months after planting, root Ct-values were consistently high, whereas leaf Ct-values decreased, and therefore CLas titers increased, during the 15 months of the study. Consequently, leaf tissue analysis identified 86% of the trees as infected by the end of the study, whereas root tissue analysis identified only 66%, and there was no correlation of root and leaf Ct-values. These results suggest that CLas detection is less reliable when using roots as opposed to leaves in an HLB-endemic production environment such as Florida. However, in regions where HLB and/or its insect vector are not yet endemic, fibrous root analysis may allow CLas detection before leaf symptoms appear in some instances.
CLas is a phloem-limited bacterium, and therefore the movement of carbohydrates is likely to influence CLas distribution in the tree, especially during the early stage of tree growth and development. After transplanting trees to the field, photosynthates need to be mobilized from the leaves to support root system establishment and growth (Watson 2005), allowing CLas bacteria to be translocated along the same path. This may explain the earlier detection of CLas in the roots of some of the trees in our study.
One common characteristic of HLB is that starch accumulates in the leaves, whereas it is depleted in the roots (Achor et al. 2010; Aritua et al. 2013; Etxeberria et al. 2009). We found a moderate correlation of starch and CLas titers in the leaves. In addition, more starch was measured in ‘Valencia’ leaves, which were more HLB symptomatic, than in non-‘Valencia’ leaves. A similar relationship between starch accumulation, CLas titer, and HLB disease symptoms was also found by Boava et al. (2017). In that greenhouse study a higher accumulation of starch was found in HLB-symptomatic ‘Pera’ sweet orange and ‘Sunki’ mandarin trees compared with several asymptomatic varieties, which were hybrids of ‘Sunki’ mandarin and trifoliate orange. In contrast to the leaves, there was no clear relationship of CLas titers and starch concentrations in the roots, nor was there a consistent influence of the rootstock. Root starch concentrations were less than half that of the leaves, but both showed a similar seasonal trend and no apparent inverse relationship. However, root Ct-values were low, and the experiment was ended before trees were severely affected by HLB.
At the end of the experiment, ‘Valencia’ scion trunks were smaller in diameter than non-‘Valencia’ trunks. Although the higher sensitivity of the ‘Valencia’ scion to HLB may have contributed to these differences, cultivar-specific differences and grafting effects associated with the different taxonomic affinities of the grafting partners (Goldschmidt 2014) likely contributed as well. The differences in trunk diameters among the non-‘Valencia’ scions measured at this early stage of disease were also likely attributable to the genetic potential of the cultivar rather than to HLB.
Trials conducted in Florida under HLB-endemic conditions indicated that some rootstocks can improve the grafted tree performance (Bowman et al. 2016a, 2016b, 2021; Kunwar et al. 2021, 2023; Singerman et al. 2021). The same was observed in studies conducted in Brazil (Boava et al. 2015; Girardi et al. 2021), where HLB is present, albeit less severe than in Florida (Bassanezi et al. 2020). The results from this study confirm earlier observations that CLas titers are lower in the roots of infected sweet orange trees with certain HLB-tolerant rootstocks as compared with more sensitive rootstocks (Bowman and Albrecht 2020). This may be one mechanism for the better performance of some graft combinations with ‘Valencia’ or other HLB susceptible scions. However, our observations indicate that rootstocks neither considerably affect ACP infestation nor CLas titers in the grafted scion itself. We did not determine the relative influence that other factors, besides reduced CLas titer in the roots, may have in the observed better performance of sweet orange trees with some rootstocks in the longer term.
Compared with sweet oranges and other commercial citrus scion cultivars, some cultivars commonly used as rootstocks are less prone to ACP infestation and CLas colonization when grown as seedlings or as a grafted scion. In contrast, graft combinations with a susceptible scion, such as ‘Valencia’ orange, are vulnerable to HLB even in graft combination with tolerant rootstocks. Except for the initial stages of infection, CLas are detected more consistently and at higher levels in the leaves than in the fibrous roots, regardless of the rootstock. The results from this field study reinforce the greater relative influence of the scion than the rootstock in HLB progression, previously demonstrated under greenhouse conditions. This suggests that other rootstock-specific traits, such as a broader stress tolerance and a better ability to absorb water and nutrients, along with influences on the scion phenology, may play the primary role in the grafted tree tolerance reported for some rootstocks.
Average leaf (solid line) and root (dotted line) Ct-values of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida 4 to 15 months after planting. Values on the y-axis are reversed to illustrate the inverse relationship of Ct-values and CLas titers. *, *** Significant differences between tissues at P < 0.05 and P < 0.001, respectively. No asterisks are shown when P > 0.05. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Average leaf and root starch content of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida. *, **, *** Significant differences between tissues at P < 0.05, P < 0.01, and P < 0.01, respectively. No asterisks are shown when P > 0.05. T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a strong positive correlation of root Ct-values obtained with the TXCChlb and the HLBaspr primer/probe sets.
Percentage of Candidatus Liberibacter asiaticus (CLas)-positive citrus trees with different scions and rootstocks growing in an huanglongbing-endemic environment in southwest Florida based on leaf and root analysis 4–15 months after planting. Trees were defined as CLas-positive when Ct ≤ 32. Percentages are shaded based on their values from blue to red, with darkest blue indicating the lowest percentages and darkest red highest. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a moderate inverse correlation of leaf Ct-values and leaf starch content 15 months after field planting.
Contributor Notes
The manuscript is associated with a presentation given at the Florida State Horticultural Society Annual Meeting, held in Daytona Beach, FL, USA, 26–28 Oct 2021.
Funding for this study was provided by the Citrus Research and Development Foundation (CRDF), USDA National Institute of Food and Agriculture Hatch project 1011775, and USDA-NIFA project 2017-70016-26328. We thank the plant physiology team at Southwest Florida Research and Education Center for their hard work and enduring enthusiasm.
U.A. is the corresponding author. E-mail: ualbrecht@ufl.edu.
Average leaf (solid line) and root (dotted line) Ct-values of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida 4 to 15 months after planting. Values on the y-axis are reversed to illustrate the inverse relationship of Ct-values and CLas titers. *, *** Significant differences between tissues at P < 0.05 and P < 0.001, respectively. No asterisks are shown when P > 0.05. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Average leaf and root starch content of young ‘Valencia’ and non-‘Valencia’ trees growing in an huanglongbing-endemic environment in southwest Florida. *, **, *** Significant differences between tissues at P < 0.05, P < 0.01, and P < 0.01, respectively. No asterisks are shown when P > 0.05. T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a strong positive correlation of root Ct-values obtained with the TXCChlb and the HLBaspr primer/probe sets.
Percentage of Candidatus Liberibacter asiaticus (CLas)-positive citrus trees with different scions and rootstocks growing in an huanglongbing-endemic environment in southwest Florida based on leaf and root analysis 4–15 months after planting. Trees were defined as CLas-positive when Ct ≤ 32. Percentages are shaded based on their values from blue to red, with darkest blue indicating the lowest percentages and darkest red highest. T1 = Jul 2019, T2 = Oct 2019, T3 = Jan 2020, T4 = Mar 2020, T5 = Jun 2020.
Scatter plot showing a moderate inverse correlation of leaf Ct-values and leaf starch content 15 months after field planting.
