Sublethal exposure to an emulsion based on Pogostemon cablin (Lamiaceae) essential oil impairs digestibility, feeding, fecundity, and mobility of the coffee berry borer, Hypothenemus hampei

Abraão Almeida Santos (  abraaoufs@gmail.com ) University of Florida https://orcid.org/0000-0001-5284-3294 Cliver F. Farder-Gomes Universidade Federal de Vicosa Arthur V. Ribeiro University of Minnesota Twin Cities Campus: University of Minnesota Twin Cities Thiago L. Costa Universidade Federal de Vicosa Josélia Carvalho Oliveira França Universidade Federal de Lavras Leandro Bacci Universidade Federal de Sergipe Antônio Jacinto Demuner Universidade Federal de Vicosa José Eduardo Serrão Universidade Federal de Vicosa Marcelo Coutinho Picanço Universidade Federal de Vicosa


Introduction
Insecticide misuse in agriculture has been associated with food, water, and environmental contamination, increasing the search for safer substances in pest control, including environmentally friendly molecules that can be used without negatively affecting human and environmental health (Bolzonella et  The coffee berry borer (CBB), Hypothenemus hampei Ferrari (Coleoptera: Curculionidae: Scolytinae), is the leading pest of coffee crops worldwide, causing economic losses of up to US$500 million annually (Vega et al. 2002). Females bore the fruits and lay eggs in the endosperm, where newly hatched larvae and adults feed. CBB causes economic damage by reducing coffee grains' quality and market value (Infante 2018). Also, immatures and adults can feed on a single fruit simultaneously because of the fast development of CBB (estimated cohort generation time of 45 days) aligned with the high longevity of adults (Baker et al. 1992).
For many years, conventional insecticides, such as endosulfan, were at the forefront of CBB management (Vega et al. 2015). Despite its e cacy, endosulfan has been banned from 70 countries, including Brazil, one of the primary coffee producers worldwide, because it is a persistent organic pollutant that causes human and environmental hazards (Vega et al. 2015; Menezes et al. 2017). The limited options for the management of CBB led to the search for alternative molecules, but mainly conventional synthetic insecticides have been investigated (e.g., Plata-Rueda et al. 2019).
The insecticide effect of EOs against CBB has been documented (e.g., Reyes et al. 2019). However, the development of emulsions containing these oils and knowledge on their sublethal effects is lacking. Therefore, we sought in this study to test the lethal and sublethal effects of Pogostemon cablin (Benth.) Blanco (Lamiaceae) and its emulsion against CBB. The EO of P. cablin has been previously reported showing insecticidal activity (Albuquerque et al. 2013;Rocha et al. 2018), but its effect on CBB is still unknown. Thus, we addressed the question: Can a emulsion of P. cablin EO achieve lethal and sublethal toxicity similar to that of the pure oil on CBB?

Insects and essential oil
We obtained adult females of H. hampei from infested coffee berries (Coffea arabica L. cv. Catuaí) collected in an organic eld (-20.739994º, -42.847968º), Viçosa, state of Minas Gerais, Brazil. The females were kept in the dark at 27±0.5°C, 65±10% relative humidity to produce one generation (F1). The F1 females were removed from the fruits 24 h before the bioassays. The EO of P. cablin (98% purity) was purchased from Empório Laszlo Corp. (Belo Horizonte, State of Minas Gerais, Brazil).
Essential oil and emulsion: chemical and physical analyses Fifteen days after preparation, the EO chemical composition and EO-based emulsion physical aspects were determined. Quantitative analysis of EO components was done in a Shimadzu GCMS-QP5050A apparatus tted with a capillary DB-5 column (30 m × 0.25 mm, with 0.25 µm lm thickness) coupled with a ame ionization detector, using helium as the carrier gas ( ow rate of 1.8 mL min−1) and injector temperature of 220°C. The initial column temperature was 40°C. Isotherm was kept for 2 min, and then the temperature was increased to 240°C at a rate of 3°C min −1 . This system was maintained for 15 min before injecting the sample (1.0 µL) in a split ratio of 1:20 and a column pressure of 100 kPa. Component concentrations were calculated as the percentage of the corresponding peak area to the total area of all peaks. The qualitative analysis was done in a Shimadzu GC17A apparatus tted with a capillary DB-5 column (30 m × 0.25 mm, with 0.25 µm lm thickness) using electron impact ionization (70 eV). Injector temperature, isotherm, column pressure, and heating rate were similar to those mentioned above. Thereafter, components were identi ed by comparing their retention times to a series of alkanes (C9 -C27) and their mass spectra to the library database of Wiley and Nist (05, 08, and 11).
The physical parameters of the emulsion related to conductivity, microparticle size, pH, and zeta potential were characterized. Zeta potential is related to the repulsive forces existing between the particles, where an increase in these forces provides more stability and avoids the agglomeration of EO particles within the emulsion. Thereby, values higher than + 25 mV or lower than − 25 mV indicate high degrees of strength (Shi et al. 2017;Hashem et al. 2018). For pH values, those lower than 4 or higher than 7.5 indicate that the dispersion would be expected to be stable (Oliveira et al. 2017). Furthermore, conductivity indicates percolation transition in water/oil rates (Liu et al. 2011).
The size of emulsion samples was obtained using an Olympus BX-53 light microscope, coupled with an Olympus DP 73 digital camera (Olympus Optical Corp., Tokyo, Japan). The computer program ImageJ was used to measure the diameter of the particles for 250 observations (National Institutes of Health, NIH, Maryland, USA). Zeta potential and conductivity were analyzed using a ZetaSizer Nano ZS90 (Malvern Instruments Inc., Worcestershire, UK) and pH using a bench pHmeter (Digimed DM 20) in triplicate. Previously, samples were diluted with 1:100 (v:v) ultra-pure water, and evaluations were performed at 25°C in triplicate (Oliveira et al. 2017).

Bioassays
All experiments were performed in a completely randomized design using newly emerged adult females and treatments involving P. cablin EO, the EO-based emulsion, and acetone (control).
Four replicates were used per tested dose, with each replicate consisting of 10 females caged in a Petri dish (7.5 cm diameter x 2 cm height) containing 1.5 g of coffee berry for feeding. The topical application consisted of 0.5 µL of the treatment solution applied on the thorax of each female with a Hamilton microsyringe (10 µL). A previous bioassay indicated that neither acetone nor the mixture used in the emulsion without the EO affected the survival of the insects (ANOVA: F = 0.16; df = 2,9; P = 0.85, Fig. S1, supplementary information). After the topical application, the insects were kept in the dark at 27±0.5°C, 65±10% relative humidity, and the number of dead insects was counted for 72 h.

Time-mortality bioassay
The calculated LD 90 for the EO and the EO-based emulsion and control were topically applied as aforementioned with ve replicates per treatment and 10 females per replicate (n = 50 females per treatment; 150 in total). The number of live insects was registered every 10 min up to 4 h, every 2 h up to 12 h, and every 12 h up to 72 h.

Fecundity
The calculated LD 25 for the EO, EO-based emulsion and acetone control were topically applied as aforementioned. Twenty replicates, each of which was one insect, were used for each treatment (n = 20 females per treatment; 60 in total). Each female was supplied with a coffee berry for the oviposition site three hours after exposure to the chemicals. The coffee berries were washed with 1.0% sodium hypochlorite, distilled water, and air-dried for two days before being exposed to the treated females. The potential bias from fruit size was reduced using berries with similar sizes [diameter (ANOVA; F = 0.85; df = 2,57; P = 0.43) and height (ANOVA; F = 0.46; df = 2, 57; P = 0.63)]. Fruits were kept separately in individual vials (5.0 cm diameter x 2.5 cm height) at 27±0.5°C and 65±10% relative humidity in the dark. The coffee berries were mechanically opened after 30 days, and the number of H. hampei larvae counted.

Feeding
We investigate if feeding by H. hampei adult females treated with the LD 25 of either P. cablin EO or the EO-based emulsion reduces the mass of coffee berries. Ten replicates, each one with one berry and ve females, were placed in a Petri dish (7.5 cm diameter x 2 cm height; n = 50 females per treatment; 150 in total). The coffee berries were weighed on an analytical balance (Genaka, AG200) before and after 72 h of exposure to the females. Another set of 10 replicates containing only coffee berries was used to correct the mass in the treatments and control to reduce the bias due to water loss. The Petri dishes were kept in a desiccator at 27±0.5°C and 65±10% relative humidity in the dark to reduce the interference of relative humidity on grain weight. The temperature (25.80±0.6°C) and relative humidity (54.20±5.80%) inside the desiccator were monitored using a digital thermohygrometer (Kasvi K295070H). The coffee berries used had similar initial weight (ANOVA: F = 0.14; df = 3,36; P = 0.94). Results were calculated as percent berry weight loss (BWL) using the formula:

Mobility
The H. hampei females were exposed to the LD 25 of P. cablin EO, EO-based emulsion, and acetone as the control to evaluate insect mobility. Fifteen replicates were used per treatment, in which each replicate consisted of one female (n = 15 females per treatment; 45 in total) in a Petri dish (7.5 cm diameter x 2 cm height). The insects were kept at 27±0.5°C and 65±10% relative humidity in darkness for 72 h. Each female was kept at room temperature for one hour for acclimatization, transferred to a Petri dish, and recorded for 10 min using an automated video tracking system equipped with a charge-coupled device camera (ViewPoint Life Sciences, Montreal, Canada) for the calculation of the displacement (cm) and walking velocity (cm/s).

Histopathological analysis
Adults of H. hampei from the control (n = 5), and those topically exposed for 72 h to the LD 25  Nussloch GmbH, Heildelberger, Germany). Sections with 3 µm thickness were stained with hematoxylin and eosin and analyzed using an Olympus BX-53 light microscope, coupled with an Olympus DP 73 digital camera (Olympus Optical Corp., Tokyo, Japan).

Statistical analyses
All statistical analyses were performed using the R software (R Core Team 2021) and packages described below for analyses and the R software [with the package ggplot2 (Wickham 2016)] and Corel Painter (Essential 7, Ottawa, ON, Canada) for graphics design.
The lethal doses of the P. cablin EO and the EO-based emulsion to H. hampei were determined with doseresponse curves using PROBIT analysis, available in the ecotox package (Hlina 2020). Models with a probability of acceptance of the null hypothesis (P > 0.05) were accepted by the chi-square goodness-oft test (χ 2 ) and estimated 95% con dence intervals (CI 95%).
The time mortality was estimated using Kaplan-Meier estimators in the package survival (Therneau 2021). Subsequently, the curves and LT 50 s (including CI 95%) from treatments were compared using the Log-rank test with Bonferroni correction to avoid false-positive results.
The fecundity and feeding data were tted in generalized linear models (GLM) using the package lme4 (Bates et al. 2015), setting treatment as the xed effect and error distribution accordingly. Data from the bioassays for the number of larvae were initially investigated under a Poisson error distribution but subsequently re tted using Quasi-Poisson (link function = log) as overdispersion was detected. Similarly, we initially used binomial error distribution but updated it to Quasi-Binomial (link function = logit) due to overdispersion for berry mass loss data. Models validity and quality were checked using the package performance (Lüdecke et al. 2021). When data showed the signi cance of treatments, means were compared by multiple pairwise comparisons using least-square means at α = 0.05 (package emmeans) (Lenth 2020).
Mobility data were submitted to analysis of variance (ANOVA). Normality of residuals and homogeneity of variance were checked by Shapiro-Wilk and Bartlett's test, respectively, and data transformation was unnecessary. When data showed the signi cance of treatments, means were compared by Tukey's test (α = 5%).
The midgut epithelium of H. hampei had a single-layered epithelium with digestive cells showing homogeneous cytoplasm and a spherical nucleus with the predominance of decondensed chromatin (Fig. 7A). The apical surface had a well-developed brush border (Fig. 7B). After 72h of exposure to the EO, the midgut epithelium showed histopathological features, including irregular epithelium architecture, disorganization of brush border, cytoplasm vacuolization, and release of cell fragments into the gut lumen ( Fig. 7C-D). In the adults treated with EO-based emulsion, those histopathological features were more severe with irregular epithelium, high disorganization of brush border and cytoplasm vacuolization, a nucleus with condensed chromatin, and release of cell fragments into the gut lumen (Fig. 7E-F).
In the brain and ovaries of H. hampei exposed insects, there was no damage compared with control. The brain had the neuropil region of the mushroom bodies and antennal lobe with a homogeneous aspect, and Kenyon cells had spherical nuclei with decondensed chromatin (Fig. 8A, C, and E). The insect ovarioles presented a well-developed tropharium region, oocyte with a prominent germinal vesicle (oocyte nucleus), surrounded by a layer of follicular cells, and a peritoneal sheath coating the entire ovariole ( Fig. 8B, D, and F).

Discussion
The increasing concerns with human and environmental safety expedite progress towards adopting alternative substances in opposition to conventional synthetic pesticides to manage insect pests.
However, research on formulations conferring higher stability for applications and the potential sublethal effects of new molecules is still needed. This study demonstrated that a emulsion containing 18% of P. cablin EO has similar acute toxicity and sublethal effects as its EO against the CBB, H. hampei. These ndings contribute to the development and improvement of eco-friendly molecules and the advancement of the knowledge about plant-based insecticides and their lethal and sublethal effects, including histopathological aspects, thus helping to ll this current gap.
EO-based emulsions have been developed to overcome EO's rapid degradation and particle size, resulting in similar or higher insecticidal activity than its pure form. The estimated lethal doses obtained here are equivalent between P. cablin EO and its emulsion, indicating the potential of the emulsion for the control of H. hampei. Thus, we hypothesize that the main reasons for such success are: i) a reduction in the EO microparticle size associated with suitable conductivity, pH, and zeta potential (i.e., greater particle repulsion), which can increase the e ciency and uniformity of a substance deposition on a surface; and ii) slower release of active ingredient resulting in prolonged insecticidal activity, which has been found for other EOs (González et al. 2015;Oliveira et al. 2017;Rocha et al. 2018). In the P. cablin EO evaluated here, patchoulol is the most abundant component and claimed to be responsible for this oil's insecticidal activity (Albuquerque et al. 2013;van Beek and Joulain 2018). So, it is reasonable to assume that this main compound was kept under a suitable stable system with nominal particle size in the emulsion, contributing to acute toxicity similar to that of the pure essential oil, as found in previous studies for leafcutting ants (Rocha et al. 2018).
Noteworthy, the insecticidal effect of an EO is not uniquely related to the major compounds because synergistic, additive, or antagonistic interactions among components have been reported (Tak and Isman 2015;Tak et al. 2016;Melo et al. 2020). Those interactions among EO components may cause rapid mortality, which occurred with the P. cablin EO as indicated by the short lethal time. This fast mortality may be attributed to some mechanisms, including the interference in calcium channels and the inhibition of octopamine and acetylcholinesterase as reported for EOs of other Lamiaceae species (Rattan 2010;Park et al. 2016). Interestingly, the emulsion has a lethal time similar to that of the EO despite its lower EO content (18% of EO), which may be due to the (co) surfactants' proprieties that increase the penetration of the insect cuticle by EOs (Kogan and Garti 2006;Hashem et al. 2018).
Sublethal effects of the P. cablin EO and the EO-based emulsion to H. hampei include the impairment of reproduction and feeding with the low number of larvae produced and a decrease in the mass of coffee berries consumed by treated females. The increase in EO bioavailability through formulations is known to cause extensive damage to insect organs, impairing processes such as feeding (Hashem et  Autophagy is primarily a survival mechanism that preserves the turnover of cell components (Noguchi et al. 2020). However, this process also kills cells after insecticide exposure, such as found for Anticarsia gemmatalis (Hübner) (Lepidoptera: Noctuidae) exposed to neem oil (Farder-Gomes et al. 2021c) and squamocin (Fiaz et al. 2018). Besides autophagy, the nuclear chromatin condensation indicates that EO and EO-based emulsion exposure may induce cell death by apoptosis in the midgut of H. hampei. Chromatin condensation is a common aspect of apoptotic cell death in organs undergoing degeneration, oxygenated sesquiterpenes represented 44% of the EO composition and may be responsible for the higher displacement observed in the treated insects. Similar results have been reported in Atta sexdens (Linnaeus) (Hymenoptera: Formicidae) exposed to both P. cablin EO and its nanoformulation (Rocha et al. 2018).

Figure 4
Number of larvae from female adults of coffee berry borer (Hypothenemus hampei) 30 days after topical exposure to Pogostemon cablin essential oil (EO) and an EO-based emulsion (18% of P. cablin EO) at their respective lethal dose to kill 25% of the population. Box plots indicate the range of data dispersion ( rst and third quartiles and extreme values), median (solid line), and data from each replicate (points). Different lower case letters indicate signi cant differences among treatments (GLM Quasi-Poisson distribution; F = 4.73, df = 2, 57; P = 0.01).

Figure 5
Berry mass loss (%) of coffee berries 72 hours after exposure to female adults of coffee berry borer (Hypothenemus hampei) topically treated with Pogostemon cablin essential oil (EO) and an EO-based emulsion (18% of P. cablin EO) at their respective lethal dose to kill 25% of the population. Box plots indicate the range of data dispersion ( rst and third quartiles and extreme values), median (solid line), and data from each replicate (points). Different lower case letters indicate signi cant differences among treatments (GLM Quasi-Binomial distribution: F = 8.74, df = 2, 27; P = 0.001).