Chemical composition and repellent activity of essential oils of Tithonia diversifolia (Asteraceae) leaves against the bites of Anopheles coluzzii

Tithonia diversifolia is widely used in African traditional medicine for the treatment of a large number of ailments and disorders, including malaria. In the present study, we evaluated the repellent activity of essential oils (EO) of this plant against Anopheles coluzzii, a major vector of malaria in Africa. Fresh leaves of T. diversifolia were used to extract EO, which were used to perform repellency assays in the laboratory and in the field using commercially available N,N-Diethyl-meta-toluamide (DEET) and Cymbopogon (C.) citratus EO as positive controls and vaseline as negative control. The repellency rates and durations of protection of the human volunteers involved were used as measures of repellent activity. Chemical composition of the T. diversifolia EO was established further by gas chromatography coupled with mass spectrometry. The moisture content and oil yield were 81% and 0.02% respectively. A total of 29 compounds in the T. diversifolia EO was identified, with d-limonene (20.1%), α-Copaene (10.3%) and o-Cymene (10.0%) as the most represented. In field studies, the mean time of protection against mosquito bites was significantly lower in T. diversifolia EO-treated volunteers compared to treatments with C. citratus EO (71 min versus 125 min, p = 0.04), but significantly higher when compared with the non-treated volunteers (71 min vs 0.5 min, p = 0.03). The same pattern was found in laboratory repellency assays against A. coluzzii. In contrast, repulsion rates were statistically similar between T. diversifolia EO and positive controls. In conclusion, the study suggests promising repellent potential of leaves of T. diversifolia EO against A. coluzzii.

www.nature.com/scientificreports/ appeared. Once the distillation process is completed, the distillate was left to rest for 1 h at room temperature and after this short stay in the separating funnel, two phases were formed. For each batch, 1 mL of EO on top of the water was collected with the micropipette and then introduced into a labelled bottle. The T. diversifolia EO are collected in 10 mL hermetically sealed glass vials wrapped with aluminum foil after dehydration of the oil-floral water mixture using anhydrous sodium sulfate (Na 2 SO 4 ) then stored at + 4 °C until further analysis. The EO yield was calculated using the following formula: C EO = [(V EO /M d ) × 100] ± [(ΔV EO /M d ) × 100], where C EO is the EO yield (mL/g), V EO is the volume of EO collected (mL), ΔV EO is the reading error, and M d the mass of dried vegetal material (g).
Determination of chemical composition of essential oils from T. diversifolia. The composition of volatile elements was determined using gas chromatography coupled with mass spectrometry (GC-MS). Separation of the EO components was performed on Hewlett-Packard silica fused capillary columns (50 cm length, 0.32 mm internal diameter, and 0.3 µm film thickness Carbowax). Helium was used as carrier gas at a flow rate of 1 mL/min. The GC column oven temperature varied from 60 to 240 °C at a rate of 7 °C/min for 20 min. Mass spectra were taken in scan mode in the range of 50-500 m/z mass to charge, operated at 70 eV, and the ion source temperature was maintained at 250 °C. The total run time was 23.82 min. The identification of compounds was done by comparing their retention index (RI) and mass spectra with those from the Wiley/National Bureau of Standards, and the National Institute of Standards and Technology libraries stored in the GC-MS database. The concentration of volatile compounds (Ci) was calculated using the following formula: %Ci = [(Ai × Fi)/Vol] where, Ai is the peak air product, Fi is the proportionality factor and Vol is the volume injected.
Laboratory rearing of A. coluzzii. Eggs of A. coluzzii (Ngousso strain) were obtained from the "Organisation de Coordination pour la lutte contre les Endémies en Afrique Centrale, Yaounde", Cameroon and reared in plastic bowls (300 eggs per bowl) for 24 h at the Insectarium of the Faculty of Medicine and Pharmaceutical Sciences to obtain larvae. The different larval stages obtained were fed with fish food every morning. Once the larval stages evolved into pupae, they were separated and kept in net-covered mosquito cages until emergence of adults. A cotton piece soaked in a 10% glucose solution was placed inside the cages to feed the emerged adults.
Repellency tests. The repellent potential of the T. diversifolia EO was evaluated by human landing assays both in the field against wild type mosquitoes and in the laboratory using established adult female A. coluzzii. All volunteers freely received antimalarial prevention before experiments, as per the national guidelines for malaria management 27 .  www.nature.com/scientificreports/ Repellency activity against wild-type mosquitoes. Nightly captures of adult female mosquitoes were conducted in April 2020 and among human volunteers in the Mabanda neighbor. Mabanda is a popular neighborhood located in the third division of the city of Douala (Littoral region, Cameroon) and where insalubrity is very present and the population lives in high promiscuity. It is also a neighborhood with a large number of mosquito breeding sites due to human activities. 2 mL of EO was dripped onto each leg from the bottom of the knee to the end of the ankle then was applied by hand onto the legs of each volunteer, from the knee to the end of the ankle to cover a skin area of about 200 cm 2 (0.01 mL/cm 2 of EO). The nine catchers were divided into three groups of three participants. The three catchers from each group were placed one-meter away from each participant in the group while the distance between groups was variable, depending on the capture sites (next to a stream between two plank houses, next to the field and next to the hard houses inland). In each group working simultaneously in one spot, the volunteer received either C. citratus EO 100% known to have repellent properties (positive control), or T. diversifolia EO 100% (test group), or no substance (negative control) to control for any capture biases. C. citratus (citronella) is greatly used as repellent by populations. In contrast, DEET repellents are few used by populations due to absence of knowledge, cost and beliefs on its toxicity. In this context, we preferred using citronella-based repellents as positive control for field experiments. The capturers were instructed not to rub, touch, or wet their substancetreated legs. The field experiments were conducted between 7.00 p.m. and 11.00 p.m. (4-h exposure) with the field temperature varying from 29 to 30 °C during each capture night. The volunteers had no contact with oils, perfumed soaps, lotions or perfumes on the day of the assay. Mosquitoes species are counted and are identified using morphological traits and dichotomous keys [28][29][30] .
Mosquito rearing and repellency activity against laboratory A. coluzzii strains. Larvae mosquitoes were fed daily with 50 mg fish powder for 6-7 days. A piece of cotton wool soaked with glucose (30 mL, 10%) was deposited onto the net-covered for the nutrition of adults that emerged Cage. The repellency activity was assessed using the Armin cage test as described by Schreck et al., and WHO guidelines on efficacy testing of mosquito repellent for human skin 31,32 . Briefly, 3-5-day-old and adult female A. coluzzii mosquitoes (n = 100) without blood supply for one to two days were kept in a net-covered cage (35 cm × 35 cm × 35 cm) 24 . The DEET 30% (positive control), or vaseline (negative control), or one of the different concentrations of the T. diversifolia EO (10%, 30%, 50% and 100% in vaseline to reduce its volatility) were used to perform the assay as described previously 24 . Only the forearm of each volunteer was exposed and the remaining area was protected with rubber gloves. The experiments were conducted in two days between 8.00 p.m. and 03.00 a.m., and were performed during a 1-h exposure period in triplicate (three different human volunteers' hands were used per test and each volunteer received the three treatments at different times) with 1 g of each formulation puting the EO treated arm and control arms into the cages at the same time for a full hour. At the day 1, the positive control arm (DEET 30%) and the one treated with 30% EO were introduced simultaneously into the mosquito cage for each volunteer and at the day 2, the DEET was replaced by the vaseline (negative control). The exposure was stopped when the first bite of the volunteer was noted. Before the experiment, the volunteer's forearm was treated with petroleum jelly as control and exposed for 30 s to check for repulsion. If at least two mosquitoes landed on the arm, the repellency assay is continued as this means that petroleum jelly has no repellent effect on mosquitoes, so it would not affect the repellent activity of the EO of T. diversifolia to be tested. Likewise, the volunteers had no contact with other substances (i.e., body oils and lotions, perfumed soaps, fragrances) on the day of the assay.
Repellency assay outcomes. Field and laboratory-based repellent activities were assessed by determining the repulsion rate and the protection time for each substance. The repulsion rate (Re) was determined using the following formula: Re = [(N C -N EO )/N C ] × 100, where N C is the number of mosquitoes captured by volunteers treated with negative control, and N EO is the number of mosquitoes captured by volunteers treated with the test substance. This rate was used to assess reduction in the attractiveness of mosquitoes to humans treated with the substance tested. The protection time is the time interval (min) between the application of the substance and the first mosquito bite.
Statistical analysis. Data were keyed in an Excel spreadsheet (Microsoft Office, USA) and then exported to the statistical package for social sciences v16 (SPSS, IBM, Inc., IL, Chicago, USA) and GraphPad v5.03 (Graph-Pad Prism, Inc., San Diego, California, USA) software for statistical analysis. Non parametric Mann-Whitney and Kruskal-Wallis tests were used to compare mean values between groups while Fisher's exact and Pearson's independence chi-square were used to compare proportions. Level of statistical significance was set at p < 0.05.
Ethics approval and consent to participate. The study was conducted in accordance with ethical guidelines related to research on humans in Cameroon. The study received ethical clearance from the Institutional Committee of Ethics for Research for Human Health of the University of Douala (no. 1976/CEI-UDo/06/2019/T). Before enrollment, subjects were informed on the purpose and process of the investigation (background, goals, methodology, study constraints, data confidentiality, and rights to opt out from the study), and signed informed consent was obtained from all those who agreed to participate in the study in accordance with the Helsinki Declaration. Participation was voluntary, anonymous and without compensation.
Collection of wild plant material was carried out in accordance with national guidelines and Cameroonian legislation (provisions of the law on forest and environmental management in Cameroon No 94/01 of 20 January 1994 relating to the forest, fauna and fishing regime in Cameroon and explicitly recognizing the rights of local populations use on various forest products) and an identification certificate was obtained at the National Herbarium of Cameroon.

Results
Phytochemical screening of the T. diversifolia essential oil. The moisture content of the leaves of T. diversifolia was 81% and the yellow essential oil yield was 0.02% after hydrodistillation of 50 kg of dry leaves of T. diversifolia. The GC-MS analysis revealed the presence of 29 compounds in the T. diversifolia EO (yellow pale-colored), with d-limonene (20.06%), α-Copaene (10.29%) and o-Cymene (10.0%) as the most represented ( Table 1). The 29 compounds identified belong to four groups namely monoterpenes, sesquiterpenes, phenylbutane derivatives and triterpenes, which accounted for 54.13%, 35.22%, 8.87% and 1.78% of the compounds, respectively.
Field repellent activity. The mean time of protection against mosquito bites was significantly higher in volunteers treated with the C. citratus EO as compared to those treated with the T. diversifolia EO (125 min versus 71 min, p = 0.04) (Fig. 2).
The analysis of the entomofauna captured revealed a similar distribution of Culex, Anopheles and Mansonia mosquitoes between the two groups treated with the EO (Fig. 4). Regarding Anopheles mosquitoes collected in the groups, these were predominantly identified as A. gambiae s.l.
It should be noted that minor burning sensations were reported only in volunteers treated with the C. citratus EO. No adverse signs/symptoms were observed in those treated with the T. diversifolia EO.
In vitro repellent activity. Finding from the in vitro repellency assay showed that the time of protection from biting was on average higher in the group of capturers treated with DEET as compared to the other groups treated with varying concentration of the T. diversifolia EO, and the difference was statistically significant www.nature.com/scientificreports/ (p = 0.0016) (Fig. 5a). In contrast, no difference was found in terms of the number of mosquitoes captured and repulsion rate between DEET-treated group and the T. diversifolia-treated groups (p > 0.05) (Fig. 5b,c). No allergic events were reported during the laboratory experiments.  www.nature.com/scientificreports/

Discussion
The emergence of mosquito populations resistant to synthetic insecticides and commercially available repellents hinders their use at the population level for control of mosquito-borne diseases such malaria. Unlike their synthetic counterparts, there is no evidence on emergence of resistance to natural substances. In this context, there is a need for new molecules with good repellent potentials and safety. We therefore evaluated the repellent potential of EO from leave parts of T. diversifolia, a plant traditionally used in Cameroon for treating chickenpox, and in other countries (e.g., Mexico and Nigeria) for treating malaria and other diseases [32][33][34][35] .
The yield of T. diversifolia EO (0.02%) differs from that found for the plant collected in another city in Cameroon in a previous study 36 . This finding is consistent with other previous studies showing a difference in yields depending on plant's harvest location, harvest period, leaf condition, drying or extraction time or technique with a yield that varies between 0.01% and 0.1% for essential oils extracted from leaves 24,[37][38][39] . Another study revealed that in addition to the storage time, the transport conditions also impact the EO yield of this plant 40 .
Hydrogenated monoterpenes and sesquiterpenes were the main compounds found in the T. diversifolia EO. This finding is also consistent with previous reports on the chemical composition of the plant EO from other regions of Cameroon and beyond 24,33,35,38,41 , though other previous reports from Cameroon outlined a predominance of sesquiterpenes in flower EO 36 . Our detailed analysis of the chemical composition identified 3 of the 29 compounds at high levels viz. (d)-α-limonene (monoterpenes), α-Copaene (sesquiterpenes) and o-Cymene (sesquiterpenes). Lamaty and colleagues found that (Z)-β-ocimene (40.2%) was the main compound from EO of T. diversifolia collected in the town of Yaoundé, Centre region of Cameroon 36 . This discrepancy highlights the role of environment in shaping biological development of plants, and thus its impact on their chemical composition and biological activities 24,37,41 . Our study is the first to evaluate the repellent activity of T. diversifolia against A. coluzzii, a major malaria vector in Africa. We demonstrated the repellent potential of this plant against laboratory strains of A. coluzzii and against natural mosquito bites in field studies. Our findings support previous studies on the same plant that reported repellent activities of its EO fractions against A. gambiae, Aedes aegypti and Culex quinquefasciatus 23 . This repellent activity in natural condition was higher than that observed against the laboratory strain, probably due to the time between the harvest, the time of extraction (4 months after harvest) of essential oil, the natural condition test (15 days after extraction) and the laboratory test (1 month after the natural conditions test) that would have an impact on the volatile properties of certain terpenic compounds 38 . The high content of C. citratus EO in α-pinene (34.4% vs 7.72% for T. diversifolia EO) could explain it strong repellent activity (92% vs 89.5%) compared to the T. diversifolia EO rich in (d)-α-limonene, and studies have shown that α-pinene is a powerful repellent [42][43][44] . The difference in the duration between the harvest period and the extraction could have induced the volatility of compounds (difference in concentration) and the conformation change of certain compounds of the T. diversifolia EO tested in our study, as has been suggested by Walker et al. 41 The latter could also explain the low repellent activity of T. diversifolia EO compared to C. citratus EO. Besides, it would also be interesting to conduct further studies to identify the chemical compounds responsible for repellent activity of T. diversifolia along with elucidating mechanism of action. Given high content of (d)-α-limonene, this compound is probably greatly involved in the repellent activity of T. diversifolia EO, possibly in association with other major compounds found (α-Copaene and o-Cymene).
The results showed a shorter protection time in T. diversifolia EO-treated volunteers compared to the individuals treated with the positive control for both the laboratory (DEET) and field (C. citratus EO) assays. In natural conditions, the mean protection time was 71 min for T. diversifolia EO and 125 min for the positive control. These findings suggest that the T. diversifolia EO has similar repellency, but that the volatile nature of its compounds greatly impacted the protection time of T. diversifolia EO. These preliminary results on protection time against wild mosquitoes suggest that T. diversifolia EO should be used each hour in to maintain its repellent activity, www.nature.com/scientificreports/ which is not achievable in practice. In this context, it would be helpful to develop release-control formulations in order to reduce its the volatility. In the laboratory, the concentrations of the T. diversifolia EO at 50% provided a better time protection (11.9 min) against A. coluzzii than the 100% extract (9.6 min), but this time was significantly less than that of DEET at 30%. Oyewole et al. in Nigeria 24 had found protection times of 120, 160 and 210 min, respectively for the 10%, 50% and 100% formulations, higher than those found in our study. These results corroborate with previous study suggesting the influence of the solvent used in the formulations; unlike hexane which was used by Oyewole et al., the petroleum jelly used in our study would slow the volatility of the EO 24 . Variability in the sensitivity of an EO against Anopheles species has also been reported, A. gambiae was more sensitive to repellents than A. coluzzii 18 . One solution to improve protection time of T. diversifolia EO could be to develop controlled-release formulations in order to increase the duration of repellence activity. The volatile nature of the compounds of the EO as well as a combination of other factors such as chemical composition, natural resistance of the mosquito vector species and experimental conditions can explain discrepancies observed between the groups treated and positive control.

Conclusion
The present preliminary study showed promising repellent activity of the leave T. diversifolia EO against A. coluzzii and, suggests that it could be used as an effective vector control tool at the individual level to complement conventional control methods at the community level. However, the limited time of protection compared to controls outlines the need for extensive research on development of release-control formulations and inocuity before its potential introduction as a commercial repellent. www.nature.com/scientificreports/