Epidemiology of Cocoa Swollen Shoot Disease (CSSD) on Rehabilitated Cocoa Farms in South-Western Côte DʹIvoire

1. Introduction

Cocoa swollen shoot disease (CSSD) is a vector-transmitted disease that causes drastic yield decreases and rapid decline of an infected cocoa tree. The main symptoms of the disease are red veins banding on young cocoa leaves, fern pattern mosaic on old leaves, and shoot swelling [1]. The disease was first reported in Ghana in 1936 (Steven 1936). Since its reappearance in Côte d’Ivoire in 2003 [2], a cumulative area of 130,986.90 ha of infected trees has been cut off from 2018 to 2022. Meanwhile, at least 60,000 ha of cocoa trees are still affected [3]. These figures do not include the cocoa trees in field that have already died from this disease [4]. The disease is caused by a group of Badnavirus species, namely cocoa swollen shoot virus (CSSV). A complex diversity of the species was reported in Togo, Nigeria, Ghana, and Côte d’Ivoire where virulent strains have been identified [5,6,7,8,9,10,11]. The genome of CSSV is circular of double strains DNA of 7.2 kpb [12] with five open reading frames (ORF) [13]. At least 14 species of mealybugs, belonging to the Pseudococcidae family, are a vector in the natural transmission of the virus [14]. Some tree species have been described as a host of CSSV from which cocoa could be infected [15]. CSSD is controlled by cutting out infected areas, including a buffer zone, then replanting with tolerant planting material, and applying good agricultural practices, including agroforestry systems [16,17,18,19,20]. However, after 100 years of combatting CSSD in West Africa, limited knowledge of the disease epidemiology has led to mixed results [21]. Cocoa farmers in the endemic areas are still reluctant to adopt the removal method to combat CSSD. Instead, they continue applying unsustainable practices such as deforestation. These do not remove CSSD, and lead to landscape degradation, extensive cocoa monoculture, and heightened impacts of climate change. In Côte d’Ivoire, the reemergence of the disease and its quick spread across the cocoa belt [6] justified the intensive phase of cut and replant campaign in 2018 to combat the disease [3]. The limited success on replanted areas could be one of the causes of the bans on improved planting material propagation for a better understanding of the ongoing situation of the disease in field. Thus, the assessment of the epidemiology principles of virus disease in the CSSD management strategy in field will help on optimizing the protection of replanted areas in an integrated management. The present research aims to update the on-field disease propagation pattern of cocoa swollen shoot disease in replanted areas through the evaluation of the statue of rehabilitation technologies for supporting policy decisions around replantation campaigns. In this paper, we reported the monitoring of CSSD in experimental trials over time by assessing the epidemiological pattern of the disease through field observations combined with molecular data on-farm and at landscape level.

2. Materials and Methods

An on-farm cocoa rehabilitation pilot trial of 1200 plots without plants barriers was set up in 2016 including three technologies: grafting on asymptomatic mature trees (GVV), replantation with grafted seedling (RPG), and replantation with hybrid seedlings (RPH). A total of four types of planting material represented by cocoa clones identified by the codes C1, C9, C15, and C16 were used for the grafting technology on old trunk and on seedlings while the hybrids were composed by the improved cocoa planting material already shared to farmers. The performance criteria for selecting those clones were based on the productivity (More than 2 Tones /ha), the butter quality, the resistance to black pod disease and, the compatibility factors between these planting materials. The approach of this study was based on three interventions: (i) survey at landscape level for presence/ absence of the disease in the replanted plots, (ii) on-farm observations of the CSSD characteristics on infected plots, and (iii) molecular in analysis laboratory through asymptomatic cocoa leaves samples collected around the CSSD outbreaks in field.

2.1. Assessing CSSD Epidemiology at Landscape Level in the Rehabilitated Farm

• Determining prevalence and incidence of the swollen shoot disease at landscape level

CSSD prevalence and incidence were monitored through on-site field assessment from 2020 to 2022, which was 4–6 years after the pilot implementation. Two surveys for the presence or absence of the disease in the rehabilitated plots were conducted. The first survey, in 2020, was carried out on a randomly selected representative sample size of 234 plots from the 1200 plots of the pilot. The second survey was performed from 2021 to 2022 on 200 different plots of the same pilot.

For that effect, the farms were investigated individually during the most appropriate period of the year for symptoms appearance (February to April). Only characteristic symptoms of red vein band, fern pattern mosaic, and swollen shoot were considered as indicators of CSSD presence. Incidence data and geospatial information were collected on each farm.

• Evaluation of CSSD impact on cocoa yield

CSSD effect on yield was estimated by collecting production data on 68 selected plots from only GVV technology of the pilot based on the precocity of this technology on pod early appearance. Production data were collected each two weeks in 2021 from quarter 1 to quarter 3 (Q1–Q3) by counting the number of healthy pods harvested.

2.2. On-Farm Characterization of CSSD Typical Infection

CSSD characterization on farm were performed on selected sites where the disease was present through the field survey as described above. For significant statistical data analysis, 3 sites of observations per technology were considered. Equitable number of infected plots per technologies was required and the technology with insufficient infected plots was dropped. The sites were selected from the list of infected farms within each technology based on the level of degradation and the access facility.

Data on CSSD typical infection pattern were collected by considering epidemiology parameters at farm level mainly the number and size of the disease outbreaks, the prevalence, the incidence, the severity of the disease on cocoa varieties or clones and the abundance of mealybugs vectors. The selected plots were re-visited at tree months after the initial recording. Thus, the number of outbreaks per farm was determined by identifying and localizing the disease outbreaks by analyzing CSSD-specific symptoms, die-back, and cocoa tree mortality in group. The prevalence and the incidence were also assessed by counting symptomatic cocoa trees over the time. Only characteristic symptoms of red vein bands, fern pattern mosaic, and swollen shoot were considered to indicate CSSD presence in the farm. The GPS reference of infected cocoa trees was determined per farm and neighboring infections were identified and mapped. Disease severity was estimated per clone and per rehabilitation trial using a scale [18,20]: 1—red vein-banding on flush leaves; 2—chlorotic vein flecking; 3—chlorotic vein clearing/green vein-banding; 4—diffused flecking; 5—swelling and die-back; 6—death of plants. For the vector, mealybug abundance was determined on cocoa stems between ground level and 1.5 m height on pods and young leaves using a magnifying glass. Mealybugs were counted separately on symptomatic and asymptomatic cocoa trees.

2.3. Molecular Characterization of CSSD Epidemiology

Polymerase chain reactions (PCR) methods were used to make the diagnostic of asymptomatic trees in latency phase where infected cocoa trees did not show any symptoms. The latency phase was characterized on three ranges of 6 m, 12 m, and 18 m buffer from the symptomatic trees around the outbreaks based on the currently enforced cutting out procedure. A randomized sample was taken of 10 asymptomatic cocoa trees per range zone for a laboratory PCR analysis. A total number of 3–5 leaves were collected per sampling trees. The DNA extraction of each leaf sample was done the day after sampling at the West and Central African Virus Epidemiology laboratory. The Plant Dneasy kit of Qiagen was used as described by the manufacturer with 50 mg of leaf tissue removed from the bottom part of the leaf blade. The DNA solution obtained was stored at −20 °C for PCR analysis. The Taq DNA polymerase from Promega was used to amplify the viral DNA and the specific primers CSSD1 R/F (5′-CTTCYTCYCCAATTATCCAGACTGC-3′ and 5′-AAYTGGCARAAYGGAGARGC-3′) of 400 bp designed in the ORF3 of the CSSV genome [10].

The PCR reaction mix of 25 µL volume contained 1 µL each of forward and reverse primer, Taq DNA polymerase (0.125 µL), 5 µL of 5X PCR buffer, dNTPs (0.5 µL), Mgcl2 (1 µL) and 2 µL of DNA. The amplification program, run in a Mastercycler Eppendorf type flexlid, consisted of 35 cycles with initial denaturation at 94 °C (2 min), denaturation (20 s), primer annealing 55 °C for 15 s and extension at 72 °C (30 s), and a final extension of 5 min at 72 °C. At the end of the amplification process, 10 µL of amplified product was used for electrophoresis in 0.8% agarose gel and photographed under UV illumination with an imaging system. A clone of CSSV DNA for CSSD1 F/R primers sent by the University of Arizona (US) was used as positive control. Only visible light spot-on agarose gel was considered as positive after the electrophoresis photography.

2.4. Data Analysis

The prevalence rate was calculated using the number of infected farms/trees out of the total number observed. The incidence rate was calculated by considering the number of newly infected farms/trees.

The yields were estimated using the formula Y (kg. ha⁻¹) = N * D * 0.04 (N= Number of healthy pods per tree;

D = Numbers of trees per hectare; 0.04 = bean weight (kg) per pod according [22]. The resulting production count was compared between asymptomatic and symptomatic plots at landscape level.

The severity data at farm level were analyzed by cumulating the score per infected cocoa clones. CSSD vectors were characterized by mealybug observation on each selected farm. Thus, Statistical comparative analysis was performed on the resulting data. Statistical analysis was performed using SPSS. Diagrams and box plots were designed with R. The maps were generated using QGIS software version 3.16.8-Hannover.

3. Results

3.1. CSSD Epidemiology at Landscape Level

3.1.1. Prevalence and CSSD Mapping in the Pilot Network

Characteristics of the CSSD infection rate on the pilot plots network during the three surveys revealed that the disease appeared in the rehabilitated farms with a prevalence of 3.42% in 2020, i.e., four years after implementation in field. Also, the number of infected plots in the observed sample more than doubled from year 5 (2021) to year 6 (2022)—from 10 to 23 infected plots. The location of asymptomatic and CSSD diseased pilot farms is shown in Figure 1. Infected farms are unequally distributed on the areas covered by the rehabilitation in the Nawa region. Infected farms are mostly observed in the areas of Méagui, Gnammangui, and Buyo. No infection was observed in the areas of Dabouyo and Okrouyo near Gueyo.

3.1.2. Spread Dynamics of CSSD Infection Within the Rehabilitation Pilot

Infection data of the selected samples throughout the three surveys for 2020, 2021, and 2022 show a CSSD incidence rate of 1.05% between year 4 (2020) and year 5 (2021) after planting. The proportion of this rate becomes exponential between year 5 (2021) and year 6 (2022) where an incidence rate of 6.95% has been observed (Figure 2). These results indicated the CSSD spreads rapidly over time in the rehabilitation plots despite use of improved planting material. Also, the repartition of infected plots per technology shown a high rate of infection with GVV plots (70%) while RPG presented the lower rate (10%), Table 1.

3.1.3. Impact of CSSD Infection on Cocoa Yield in the Pilot

From the total number of 68 GVV plots monitored for cocoa production in quarters 1 and 3 in 2021, 52 were asymptomatic and 16 were symptomatic. The overall recorded yield was between 1000 and 1900 Kg/ha. No significant difference was found for cocoa yield between symptomatic and asymptomatic plots when considering the entire cocoa plot irrespective of the cocoa variety and technology applied (Table 2).

3.2. CSSD Epidemiology At-Farm Level

3.2.1. CSSD Mapping on Rehabilitated Farm

The required number of tree infected plots for this study was obtained with GVV and RPH rehabilitation technology. The mapping of CSSD-infected cocoa trees inside the six selected farms for GVV and RPH technologies indicated unequal distribution of the infection on each plot. In the GVV trials, CSSD presents massive infection on cocoa trees in the centre of the plots and sporadic points along the borders. When using seedlings of improved planting material (RPH), CSSD infections were identified mostly at the plot borders on the rehabilitation plots. This is evidence that reinfection of replanted young farms after the cutting out process depends on the disease pressure in the neighboring plots. There were fewer CSSD reinfections in RPH trials when the surrounding CSSD outbreaks were limited in size (case of Koda and Logboya 1). However, the disease spread rapidly, and the number of infected trees grew with increasing size of the neighboring outbreaks (case of Logboya 2).

In both cases (GVV and RPH), the rehabilitations were at risk of infection in the presence of bordering CSSD outbreaks. These had previously occurred in Petit-Tiémé and Krohon B for GVV, and at Logboya 2 for RPH (Figure 3).

3.2.2. CSSD Prevalence and Incidence On-Farm

Prevalence and incidence showed a high rate on GVV farms in comparison to RPH farms, confirming the mass infection on the farms with grafting technology (GVV) (Table 3). Results indicated that CSSD propagated more rapidly on GVV farms where the initial number of infected trees was found to be increased. CSSD infection spread slowly on farms that were replanted entirely.

3.2.3. Mealybug Abundance on Infected Farms

At on-farm level, the results indicate a slightly increased number of mealybugs observed on the trunks of asymptomatic cocoa trees compared to infected cocoa trees (Figure 4).

3.2.4. CSSD Infection on Cocoa Clones in Field

For the rehabilitation with grafting technology, the infection rate by clones indicates that CSSD highly infected all four types of clones (C1, C9, C15, and C16), even if C9 showed slightly lower infection rates (Figure 5a). All the clones presented the same level of susceptibility according to the severity score of CSSD-infected cocoa clones (Figure 5b).

3.2.5. CSSD Infection on Hybrids Cocoa Varieties

CSSD outbreaks at initial stage were observed for the rehabilitation farms with cocoa hybrids five years after planting of the seedling. The number of infected cocoa seedlings in the plots increased with the number of CSSD out-breaks (Figure 6).

3.3. Molecular Epidemiology of CSSD Around the Disease Spots on Farm

The selected plots were categorized by visibly infected trees per outbreak, according to the official protocol of cutting out. All the tree GVV plots were in the category of 11–100 infected trees. On these farms, cuts of 12 m in diameter around infected trees is officially recommended to eliminate the virus. The three other RPH farms containing 1–10 infected trees per outbreak are categorized for cutting out in a 6 m diameter. The molecular PCR results (Figure 7) on symptomless trees adjacent to visibly infected ones showed CSSD infection with variable infection rate according to the distance around the outbreaks.

For the outbreaks containing 1–10 infected trees, the collected samples in the 6 m buffer zone showed an 87% CSSD infection rate on asymptomatic trees. However, in the 12 m buffer around the same outbreaks,80% of infection rate was observed with the PCR analysis on asymptomatic trees. Even in the 18 m buffer, a 20% infection rate was detected (Figure 8). Under current control measures, these latency infections in the 12 and 18 m buffers zones around the CSSD spots would not be considered and thus not removed during the treatment of those outbreaks.

In the outbreaks containing 11–100 infected trees, the buffer zone from 6–12 m showed an 83% infection rate detected by PCR. This confirms the approach of cutting out of all trees in this radius when a respective infection outbreak occurs. However, the buffer of 18 m around the outbreaks still showed a 30% latency infection rate that is neglected according to the official CSSD treatment control measure. The results indicate a high percentage of missed infection in the small outbreaks (80%) in comparison to the large outbreaks (30%).

4. Discussion

This study was conducted to understand the cocoa swollen shoot reinfection and how it spreads over time through rehabilitation plots with three innovative propagating technologies. The data presented here was collected on a subsample of the pilot, as we are not able to visit all the farms of the pilot that included 1200 plots. However, the random sample method using statistical approach made the results representative of what is happening in field. The pattern of the CSSD infection in the pilot is discussed at landscape and farm level.

4.1. CSSD Propagation at Landscape

The CSSD spread was mapped at a landscape level. Some areas such as Gueyo were still virus- free, while other localities including Méagui and Buyo showed infection/reinfection of newly planted farms. These results confirmed previous observations in the Soubré area where less CSSD infection was observed in areas with diversified landscape farms compared to extensive cocoa monoculture [23]. The spread of CSSD is greatly facilitated at this level of landscape diversity at Méagui, Soubré, and Buyo. However, the lower rate of infection observed in the Gueyo area could be explained by the alternance of palm oil plantations with cocoa farms [19].

4.2. CSSD Reinfection and Sustainability in Rehabilitated Cocoa Farms

This study was conducted in the Nawa region on cocoa rehabilitation pilots. The pilots were established in 2016 for validating the on-farm agronomic performance of innovation technologies at small scale. Results of the on-field survey indicate early CSSD infection/reinfection in this pilot, four years after the rehabilitation. The CSSD prevalence recorded in the pilot at landscape level showed an exponential propagation incidence up to 7%, which doubled between year 5 in 2021 and year 6 in 2022. The lack of barriers around these trial plots, likely associated to the latency infection phase, facilitated the rapid infection observed. For GVV trials, the initial selection of virus-free mature farms for rehabilitation by grafting was based on visual observations for presence/absence of CSSD characteristic symptoms. The mapping of the disease spots inside those trials showed that most infected cocoa trees have been in the Centre of the plots. This result indicates that some of the selected plots for grafting were infected from the beginning of the pilot but did not show any symptoms at the time of selection. The latency phase of CSSD constitutes one of the main challenges for infield inspections. In addition, results indicate that the disease spread from outbreaks in neighboring plots and increased pressure from the virus on the trial plots. Thus, for rehabilitation success, the use of an early detection tool in-field is a key to scale up any productivity-increasing technology based on improved planting material [24].

For RPH technology with hybrid planting material, the selected infected farms were cut individually without treating the neighboring infected farms before establishment of the pilot. Observations indicated the presence of the disease outbreaks around the trials and the appearance of infections on seedlings along the border. This clearly indicates the disease migration from infections outside of the newly planted cocoa farms. The lack of barriers around these trials left the pilot more susceptible to disease reinfection, which thus greatly facilitated the spread of CSSD [19]. However, the sporadic occurrence of disease observed in the Centre of the trials could be explained by poor application of the CSSD management risks protocol/guide in the rehabilitated farms by farmers. The early reinfection observed on those rehabilitated plots with improved planting material also indicates that partial or sporadic cutting out in areas of CSSD mass infection was not the appropriate measure to properly eliminate the virus pressure and all sources of infection in the surrounding environment. Systematic cleaning of the virus in the environment before replanting in area of CSSD mass infection associated with farmers education on risks management would be helpful on mitigating the CSSD rapid spreading and ensure the sustainability of replanted cocoa farms.

4.3. CSSD Infection and Planting Material Tolerance on Farm

The improved planting materials (clones and hybrids) in these trials were infected four years after their establishment under farmer field conditions. They were thus all susceptible to CSSD. This situation calls for building a research program for development of CSSD-resistant cocoa varieties. A harmonized approach could combine field trials with molecular screening while considering major productivity challenges in West Africa [25]. Thus, the use of tolerant planting material without integrating any other control measure cannot protect adequately the replanted farm against CSSD reinfection. Consequently, more investigation is needed for the breeding program in West Africa regarding the selection of resistant cocoa varieties for CSSD control. To date, the best way on combatting the disease will be integrated management that combines cross-protection tools, agroforestry, barriers, tolerant planting material and capacity building for farmers.

4.4. Missed Infections in the Current Application of the CSSD Control Measures

The PCR early detection results on the randomly selected asymptomatic trees around the disease outbreaks on farm show that the recommended buffer largely underestimates the actual infection on the three-buffer zone delimited around the CSSD outbreak. For category 1 (between 1–10 infected trees) with an officially recommended 6 m buffer, a substantial part of asymptomatic trees remains in the field. Results show 80% of infection on asymptomatic trees in the 6–12 m buffer and 20% in the 12–18 m buffer. In category 2 (between 11–100 infected trees), with a recommended 12 m buffer, a non-neglectable part of asymptomatic trees remains outside of the buffer zone. This remains true even if the largest part of asymptomatic trees is cut out with the recommended method. The recommended buffer for category 2 reflects the reality of more CSSD infection. Even so, it underestimates the need to remove all the infected trees around the outbreak. This finding can help refine the recommendations for the categories 1 and 2 to ensure better success of the cut-out method to combat CSSD. This result is in conformity with [26] and [27], which concluded that about 30–40% of asymptomatic trees around CSSD outbreak are in latency phase. The latency infection causes the failure of the CSSD control in field (when this control is based solely on visual inspection for delimiting cut-out areas around the disease outbreak).

Furthermore, results indicate that the rate of trees in latency phase is high at the initial stage of CSSD infection when 1–10 symptomatic trees were observed. Thus, training farmers on CSSD risk management should be a focus of detection strategies in the asymptomatic phase and the spread of the virus by the vector. In fact, before the mass infection, the disease presents early characteristic signs and symptoms that farmers should recognize for an early intervention. According to [28] and [29], control measures are more efficient when the disease is detected at an early stage in field. Thus, strengthening the notion that an in-field detection tool to detect the asymptomatic trees can play an important role by enabling frequent surveillance of new planting farms [24].

5. Conclusions

This study addressed epidemiology characteristics of CSSD in newly planted/established/ rehabilitated farms in regions like Nawa that had existing CSSD infection pressure. Before rehabilitation, farmers should consider cutting all the trees on infected farms rather than the current practice of sporadic cuts of only infected areas. Based on results, the following suggestions for improving the CSSD control measure in field in Côte d’Ivoire can be drawn: (i) establish a permanent/sustainable education system for farmers on CSSD detection and risk management using an update guide of CSSD control for training extensionists, (ii) use virus-free cocoa planting materiel and appropriate barriers in replantation and, (iii) investigate new replantation frequently for early detection of the disease using early detection tools.