Fruit Yield in Sweet Orange Trees under Huanglongbing (HLB) Conditions Is Influenced by Reproductive Phenological Characteristics of the Scion-Rootstock Combination

Abstract

Since greening (aka HLB), the most economically devastating disease of citrus worldwide was detected in Florida in 2005, citrus acreage and fruit production has reduced by more than 70%. Transmitted by the insect vector Asian citrus psyllid, the disease results in exacerbated preharvest drop, smaller fruit, and a rapid decline of trees leading to a significant reduction of yield. Currently, there is no cure for the disease. A strategy to cope with the disease relies on identifying tolerant or more productive varieties, and understanding factors that make them more productive. Under these circumstances, a combination of rootstock and scion that results in better fruit yield is highly desirable. In this paper we investigated phenological attributes of the main sweet orange varieties planted in Florida, Hamlin and Valencia, grafted on Swingle and US-942 rootstocks, two of the most used rootstocks by Florida citrus growers. Our goal was to better understand the phenology associated with the consistently higher yields of trees grafted on US-942. We assessed fruiting characteristics and abscission dynamics of fruit and leaves. We found that trees on US-942 rootstock, especially from Valencia scion, significantly set more terminals than cluster fruit; terminal fruit is larger and experiences less drop resulting in better yields. In general, rootstock did not have a significant influence on the fruit-bearing characteristics in Hamlin. Similarly, we found that fruit detachment force was not significantly influenced by rootstock. Our results show that in an HLB endemic situation, fruiting characteristics should be considered when selecting rootstocks and rootstock–scion combinations, so genetics resulting in larger, single fruits per fruiting branch should be favored.

1. Introduction

Huanglongbing (HLB) is the most economically devastating disease of citrus worldwide. In Florida, since HLB was first detected in 2005, it has reduced citrus acreage and fruit production by more than 70% [1,2]. HLB in Florida is associated with the phloem-limited bacterium Candidatus Liberibacter asiaticus (CLas) and is transmitted by the insect vector Asian citrus psyllid (ACP) (Diaphorina citri Kuwayama) [3]. The main physiological disturbance caused by the CLas is callose deposition which leads to phloem plugging [4,5,6]. The HLB infection then results in limited nutrient and water uptake due to the severe loss of feeder roots, and leaf blotchy mottle due to the abnormal accumulation of starch [7,8,9]. There is also an exacerbation of leaf drop and twig dieback, which leads to a pronounced reduction in canopy density within a short time [3]. Importantly, fruits from HLB-infected trees are typically smaller, show poorer juice quality [10], and drop more [11,12]. Exacerbated preharvest drop, smaller fruit, and the rapid decline of trees have resulted in a significant reduction in yield, although in recent years it is becoming evident that under endemic HLB conditions, the combination of scion and rootstock also determines final yield [13].

The most common scion varieties in Florida’s citrus processing industry are the sweet orange (Citrus x sinensis (L.) Osbeck) varieties Hamlin and Valencia. Both differ in the duration of the fruit maturation process. Whereas both varieties have their major bloom at the same time under Florida subtropical conditions, around the middle to the end of February, Hamlin oranges complete maturation in about 8–9 months and are harvested by November or December, whereas the fruit from Valencia is fully mature after 13–14 months (March–April of the following year). Consequently, in February, when Valencia trees bloom again, there is fruit completing the latest stages of maturation, meaning that two different crops are in the canopy at the same time.

The rootstock is an important component of commercially grown citrus trees and may determine the success or failure of a citrus operation [14]. Rootstock selection is based on tolerance to pests, diseases, soil conditions, and cold, as well as on the desired effect on scion vigor, fruit size, fruit quality, and yield [14]. Hybrids of trifoliate orange (Poncirus trifoliata L. Raf.), such as Swingle citrumelo [‘Duncan’ grapefruit (Citrus paradisi) x Poncirus trifoliata] and US-942 [‘Sunki’ (Citrus reticulata) x ‘Flying Dragon’ trifoliate orange (Poncirus trifoliata)] are among the most used rootstocks for production of sweet oranges in Florida. In the 2020–2021 growing season the number of budded rootstocks was 1,285,560 for US-942, and 468,558 for Swingle, making them the number 1 and number 4 most propagated rootstocks, respectively, during the past season according to the Bureau of Citrus Budwood Registration) [15]. Among all the commercially available rootstocks, US-942 [16] has a clear advantage for commercial use in an HLB-endemic scenario, because yields are superior [17]. HLB also appears to have less impact on US-942 as compared to other rootstocks, including Swingle, one of the classical rootstock choices for growers in Florida. The causes for higher yields coupled to a particular rootstock are not well understood as they have not been studied in detail. Many factors may play a role in the differences in yield when comparing different rootstocks, but in the situation of endemic HLB, the disease has been indicated to be a major factor in yield loss [17]. Still, differences in yield are visible when comparing rootstocks, and US-942 has been shown to be consistently superior in terms of yield. It has been suggested that US-942 is somehow tolerant to HLB because it seems to be less affected in terms of cropping and tree growth while maintaining high titers of bacteria and showing leaf symptoms [17]. However, the reasons for the higher yield have not been investigated and are unknown. In this paper, we investigated the phenological attributes of Hamlin and Valencia scions grafted on Swingle and US-942 rootstocks, with the aim of better understanding the phenology associated with consistently higher yields of trees grafted on US-942. We showed that trees on this rootstock, especially from Valencia scion, significantly set more terminal than cluster fruit, and experienced less fruit drop resulting in better yields.

2. Material and Methods

2.1. Plant Material and Fruiting Evaluation

Mature trees (7–8 years old) from sweet orange (Citrus x sinensis (L.) Osbeck) varieties Hamlin and Valencia grafted on US-942 and Swingle citrumelo rootstocks were selected at the University of Florida, Institute of Food and Agricultural Sciences (IFAS) at Southwest Florida Research and Education Center demonstration (SWFREC) grove, Immokalee, Collier County, Florida (26.42° N and 81.43° W). The trees were uniform in size and HLB symptomatic. Because HLB has been endemic in Florida since 2013, the HLB incidence was 100%. The trees were spaced 3.0 to 3.6 × 6.7 m within and between rows and were not pruned during the duration of the experiment. The soil in Immokalee is fine sand (sandy, siliceous, hyperthermic Arenic Alaquods). Standard orchard management practices were followed throughout the study for nutrition, irrigation, disease, and pest management. Dry fertilizer (12N–0P–13K; Howard Fertilizer, Lake Placid, FL, USA) was applied in the fall, spring, and summer of every year at a rate of 36 lb N per acre (40 kg/ha) per application. All trees were irrigated with a micro-sprinkler placed approximately 15 cm from the tree trunk to keep the soil near field capacity. Three replications of four trees per each rootstock–scion combination were selected on a randomized complete block design and fruit were counted using a 0.5 m² quadrat. The quadrat was placed on the west and eastern sides of the canopy and the number of terminal and clustered fruits were counted after the June drop for two consecutive seasons.

2.2. Fruit Detachment Force (FDF) Measurement

FDF was measured as described before [18] with some variations. Briefly, symptomatic terminal fruits containing a 5 cm long section of the stem were harvested from each experimental plot in October and December for Hamlin and in January, February, and March for Valencia. A total of four replicates of ten terminal fruits, from all parts of the tree, each with the peduncle or the fruit-holding stalk, were collected from each experimental plot. Collected fruits were then transported to the laboratory, where the detachment force of the fruits from the peduncle at the pedicel was measured. The force was measured in Kilogram-force (KgF) using a digital pull force gauge (Force One; Wagner Instruments, Greenwich, CT, USA); fruit stems were secured by a clamp connected to the force gauge, and fruits were then pulled.

2.3. In Vitro Treatments

To assess abscission in fruit and leaf explants of different rootstock/scion combinations, we used an in vitro system as described elsewhere [19]. Terminal symptomatic fruits with their peduncle and stem (5 cm) were detached from each variety-rootstock combination. For each rootstock–scion combination, 30 fruit explants were used. Treatments were water (control) and 1-aminoacyl-cyclopropane (ACC, 1 mM). Each treatment consisted of 15 fruits divided into three biological replications. An open-ended 5 ml disposable plastic Pasteur pipette was attached to each explant and the union was sealed with parafilm to avoid solution leakage. The open-end of the attached pipette was filled with the ACC solution or water and re-filled when consumed. The number of abscised fruits was evaluated every two days for 14 days. This study was performed during cell enlargement (June) and repeated at peel color break stages for both Hamlin (October) and Valencia (January).

In the case of leaves, we detached mature symptomatic leaves containing the petiole. Three replications of 48 leaves were used for each rootstock–scion combination and treatment. Leaf explants were prepared by removing more than 95% of the leaf blade to avoid auxin interference with the abscission process, leaving only the petiole and a small 1 cm long portion of the leaf blade and the central vein above the leaf abscission zone. Petioles were then immediately placed in 96 well plates containing 200 µL of ACC solution or water. Abscised explants were counted daily for 7 days.

2.4. Pre-Harvest Fruit Drop (PHFD) and Yield Assessment

The PHFD was monitored after the color break until harvesting for both varieties grafted on both rootstocks. Each experimental unit consisted of four trees, replicated three times, as above. For Hamlin, the number of dropped fruits was counted in September, October/November, and December. On Valencia, PHFD was counted in December, January, and March. On each date, the fruits dropped underneath the trees in the experimental plots were manually counted and cleared. Fruit weight and the number of fruits per experimental unit were recorded in situ at harvest to determine yield.

2.5. Statistical Analysis

Analysis of variance was performed on all the data collected using the generalized linear mixed model estimation (GLIMMIX) procedure in the Statistical Analysis System (SAS 9.4) software (SAS Institute, Cary, NC, USA). The GLIMMIX procedure was used to check the normality and homogeneity of the data. Poisson, beta, and normal distributions were used for count, percent, and weighed data in the model, respectively. The means were separated using least-square means using Type III tests. Any of the following p-values were used where appropriate for both field evaluation and laboratory assessments, p < 0.05, p < 0.01, p < 0.001, p < 0.0001 and indicated by *, **, *** and **** or letter(s).

3. Results

3.1. Fruit Type Distribution

The number of clustered and terminal fruits was assessed for two consecutive seasons in Valencia and Hamlin varieties grafted on two different rootstocks, Swingle and US-942. In general, both Hamlin and Valencia had significantly (p < 0.05) more terminal than clustered fruit irrespective of the rootstock. However, in the case of Valencia, the rootstock had a significant (p < 0.05) impact on the fruit type distribution. Valencia grafted on US-942 had 12% more terminal fruits than grafted on Swingle. Conversely, Valencia on Swingle had 18% more clustered fruits than on US-942. These trends were also present in Hamlin, although differences, in this case, were not significant at the 0.05 level (Figure 1).

3.2. Fruit Detachment Force

In general, FDF was not influenced by rootstock. In Hamlin, FDF was measured in October (at the end of the color break stage) and in December (at the end of maturation). In this variety, FDF was different with the year. In 2020, FDF from Swingle tended to be lower, although not significant at the 0.05 level. In any case, FDF remained steady during these two months (Figure 2A). In Valencia, we measured FDF in January and March, at the end of the color break and the end of the maturation stages, respectively. In this cultivar, FDF decreased significantly (p < 0.05) during maturation, irrespective of the rootstock. Significant differences (p < 0.05) in FDF were observed in March 2020 when considering rootstock, where US-942 showed lower FDF (Figure 2B).

3.3. In Vitro Analysis of Fruit Abscission

Natural fruit explant abscission was monitored for 10 days for all rootstock–scion combinations at two different times during fruit development, cell enlargement (June for both Hamlin and Valencia) and color break (October for Hamlin and January for Valencia) using an in vitro system (Figure 3A). The response to ACC was also assessed. Fruit abscission was influenced by rootstock. In non-mature, growing fruits of both Hamlin and Valencia on Swingle in June, abscission was absent in controls, but the fruits readily responded to ACC. In breaker fruit, abscission occurred, although was delayed in both non-mature Hamlin and Valencia on US-942. In contrast, the fruit from both varieties, irrespective of rootstock, readily abscised at the breaker stage both in controls and ACC treatment (Figure 3B,C).

3.4. In Vitro Analysis of Leaf Abscission

A leaf explant assay was performed in vitro (Figure 4A). Leaf explant abscission was monitored for 7 days. In general, all four rootstock–scion combinations performed similarly. Explants readily responded to ACC treatment, advancing abscission significantly (p < 0.001) during the whole duration of the experiment regardless of the combination of rootstock and scion (Figure 4B,C).

3.5. Fruit Drop and Fruit Size Assessment

Fruit drop was assessed after the color break until the end of the maturation process for two consecutive seasons. Hamlin and Valencia had opposite behavior. In Hamlin, in both years fruit drop was significantly (p < 0.05) higher in trees grafted on US-942 than in trees grafted on Swingle. In Valencia, fruit drop was significantly higher (p < 0.05) in 2020 in trees grafted on Swingle compared to US-942, and the same trend was maintained in 2021, although no significant differences at the 0.05 level were found in that season (Figure 5). We assessed the size of the fruit that dropped. In all cases, the larger and heavier the fruit, the fewer fruit drop, irrespective of the variety and the rootstock (Figure 5).

3.6. Yield

The yield per tree was assessed at the end of the maturation period, coincident with the peak of commercial harvest for both varieties, December for Hamlin, and the end of March for Valencia. The yield was significantly (p < 0.05) higher in trees grafted on US-942 than on Swingle in both varieties. Only in the 2020 season, there were no differences for Hamlin, but in the 2021 season, the yield was 30% higher in trees on US-942. In Valencia, the yield was consistently higher on US-942 with 37% more fruit in 2020, and 30% more fruit in 2021 (Figure 6).

4. Discussion

In this work, we have investigated rootstock phenotypical attributes that may influence Hamlin and Valencia productivity under endemic HLB conditions. For this, we determined fruit characteristics and monitored FDF and PHFD for two consecutive seasons in both varieties of sweet orange grafted on US-942 and Swingle rootstocks. These rootstock–scion combinations are among the most used currently in Florida. Furthermore, under HLB conditions, both varieties grafted on US-942 are among the top rootstock–scion combination choices [15], and growers consistently report better yields on this rootstock.

Rootstock did not have a significant influence on the fruit-bearing characteristics in Hamlin. However, in Valencia, the US-942 rootstock induced significantly more terminal than cluster fruits as compared to the Swingle rootstock. Similarly, we found that FDF was not significantly influenced by rootstock in general. Only in Valencia fruit, when measured in March, there was a significant decline in FDF, but only in one year, 2020, and this was not maintained in the following season. This indicates that FDF does not consistently depend on rootstock–scion combination and may depend on other factors such as climatic conditions, which vary greatly from year to year. Interestingly, the decrease in FDF was observed in March in mature fruit, just after the new fruit set and before the June fruitlet drop in Valencia. Trees from Valencia bloomed in mid-February. This suggests that competence for photoassimilating resources between newly set fruit and mature fruits in Valencia could be the cause for the increased fruit drop. The source-sink regulation of this type of competence for the carbohydrate pool available to the fruitlets acts as a regulatory element linking the carbohydrate status and the severity of fruitlet abscission, leading to June drop [20]. However, in this case, the situation may be more complex, as the competence for photoassimilation would be established between developing fruitlets and mature fruit. In any case, and interestingly, this decrease in FDF was not observed in Hamlin, as the fruit of this variety completes maturation by December, well before the tree blooms.

To obtain a more in-depth understanding of the possible effect of rootstock on abscission capacity and drop, we studied abscission in fruits and leaves in controlled conditions by using an in vitro system, so external factors could be excluded. The role of ethylene in enhancing fruit abscission in citrus has been proven in explants [21]. It has also been demonstrated that ethylene is the in vivo hormonal activator of leaf abscission [22]. We treated explants of fruit and leaves with ACC, the biochemical precursor of ethylene that accelerates abscission. Even though ACC accelerated explant abscission, we did not find any significant difference in the abscission rate when comparing rootstocks or in response to ACC, ruling out the influence of the rootstock in abscission when considering only HLB. Rates were very similar in both leaves and fruit explants, and in the case of fruit explants, the abscission rate was determined by peel maturity status as previously reported [19]. This evidence, hence, reinforces the notion stated in the above paragraph that rootstock does not influence abscission and in consequence, fruit drop. Our data, on the contrary, suggest that the combination of rootstock and scion imparts differential characteristics in abscission and fruit drop.

Next, we studied the characteristics of the fruit as influenced by rootstock. It has been suggested that the severity of HLB is associated with negative effects on citrus fruit growth, size, and weight of fruit at maturity, and the rate of preharvest fruit drop. In fact, out of a number of physical and compositional attributes such as seed number, leaf number, size and chlorophyll, and soluble sugars, only fruit weight and size have been linked to the likeliness of a fruit to drop. Irrespective of overall tree HLB symptoms it has been observed that fruit with smaller diameter or weight is more likely to drop than larger or heavier fruit in both Hamlin and Valencia [12,23]. The same seems to be true for healthy trees: in a study performed before HLB was endemic in Florida, the fruit was more prone to drop in the central hours of the day, when evapotranspiration was higher and fruit size and weight were lower [24]. In the present study, we found that terminal fruit, larger in size and more common in trees on US-942 than on Swingle, were less prone to drop, resulting in higher yields. Clustered fruits, proportionally more abundant on Swingle than on US-942, especially in Valencia, were smaller in size, probably due to competition for resources. Previous work has shown that carbon shortage reduces hormonal stimulators of growth, such as GAs, and increases stress-sensitive signals, such as ABA and ACC levels. It has been then suggested that this mechanism would allow, through abscission, the regulation of fruit load in accordance with the severity of the sugar deficiency [25].

In both seasons studied, the yield was consistently higher in Valencia on US-942 (p < 0.05), but in Hamlin, this varied between seasons. This reinforces the notion that rather than the rootstock, it is the combination of rootstock–scion that imparts the differential characteristics to the tree. The higher yield of US-942 is well documented, whereas the yield on Swingle is considered intermediate [17] and the reasons for this higher yield have not been determined so far, although tolerance to HLB has been observed. A link between fruit size and HLB-associated fruit drop has been previously indicated, and the involvement of water relations in the fruit as they vary during the growing season has been suggested [26]. This work did not focus on physiology associated with the responses of the different genotypes studied to varying environmental conditions; instead, we focused on how rootstock–scion combination can dictate fruiting characteristics and how this may help to explain differences in yield under HLB conditions. In any case, we cannot rule out other factors such as climactic and/or hormonal/nutrient interactions, that can vary between seasons. Future work will focus on these variables and will cover different climactic areas for a longer period.

5. Conclusions

Our data suggest that part of the higher yield consistently observed in trees grafted on US-942 can be explained because this rootstock produces more terminal than clustered fruits. Terminal fruits are larger and less subjected to competence for photo-assimilates, especially in Valencia. These results show that in an HLB endemic situation, fruiting characteristics should be considered when selecting rootstocks and rootstock–scion combinations, in order to produce larger, single fruits per fruiting branch.