Ecological risk assessment of Piper aduncum essential oil in non-target organisms

One possible alternative to chemotherapeutic agents in the treatment and prevention of diseases in fish farms is the use of Piper aduncum essential oil. However, ecotoxicological data are required to ensure its proper use and to prevent adverse effects on non-target organisms. These data are relevant since this essential oil is described as having insecticidal, molluscicidal and cytotoxic activitiy that may be associated with its chemical composition. Thus, the aim of this study was to evaluate the ecotoxicity of P. aduncum essential oil to five test organisms using the species sensitivity distribution (SSD) statistical approach. The chemical composition of the essential oil was characterized by means of gas chromatography coupled to mass spectrometry (GC-MS) and gas chromatography-flame ionization detection (GC-FID) for identification and quantitation purposes, respectively. The main component (75.5%) of the essential oil was dillapiole. The hazardous concentration for 5% of biological species (HC5) was calculated to determine the 95% protection level, resulting in a value of 0.47 mg L-1 (with a confidence interval of 0.028 1.19 mg L-1. ). A concentration range related to the level of protection for aquatic communities (the predicted no-effect concentration, PNEC) was determined through the application of safety factors to the HC5 value. The ecotoxicity parameters showed that P. aduncum essential oil can be used safely in water bodies at a concentration equal to or below 0.09 mg L-1.


INTRODUCTION
Essential oils have been considered as a possible method of disease prevention and treatment in aquaculture that causes less adverse environmental effects than veterinary drugs (Elumalai et al. 2020). Essential oils have shown interesting properties for aquaculture when administered correctly, such as sedative (Aydin and Barbas 2020), anesthetic (Hoseini et al. 2018), antimicrobial (Sutili et al. 2014), immunomodulation (Al-Sagheer et al. 2018), and stress reducing activity (Souza et al. 2017). However, it is important to consider that plantderived active substances can be stressor factors or even toxic if the conditions (way of administration, concentration used, species-specific actions, chemotype and chemical composition) are not suitable (Souza et al. 2019).
Piper aduncum L. (Piperaceae) is a plant native to the Amazon region that stands out for its therapeutic action against fungi, stomach ache, rheumatism, fever and general infection (Mgbeahuruike et al. 2017), protozoan parasites (Dal Picolo et al. 2014), and for its repellent properties (Mamood et al. 2017). The chemical characterization of the metabolites present in P. aduncum reveals that the major compound is dillapiole (Gaínza et al. 2016).
The essential oil of P. aduncum has been used to treat fish diseases (Queiroz 2012;Corral et al. 2018), yet no information exists on its behavior against aquatic non-target organisms. Non-target organisms may eventually come into contact with the essential oil when it is used for therapeutic purposes in aquaculture production or when the aquaculture effluents reach the water bodies in the surrounding areas (Bártíková et al. 2016). Therefore, the ecotoxicological assessment is required to determine the parameters for the safe use of the essential oil in fish farming (Souza et al. 2019;Bashir et al. 2020).
Th ecotoxicological evaluation assesses the potential toxicity of the substance to aquatic life, how much of the substance is expected to get into the environment, and what are the potential effects of its use on the environment (FDA 2018). A useful parameter in this context is the hazardous concentration (HC5), which is the estimated concentration that protects 95% of the species in a community and can be predicted from a small number of toxicological data (OECD 1995). For this purpose, the species sensitivity distribution (SSD) is used, which can be estimated with data from at least five species belonging to at least four taxonomic groups (Batley et al. 2018), and considers the variation in the sensitivity to the tested substance among the species. SSD is a relevant tool in the definition of quality standards with respect to the maximum values of potentially toxic substances allowed in the environment in order to protect species diversity (Aldenberg and Slob 1993;OECD 1995). The determination of the "no observed effect concentration" (NOEC) is a necessary step to establish a tolerable concentration in the environment (OECD 1995).
In this study we aimed to assess the ecotoxicity of the essential oil of P. aduncum. The chemical composition of the essential oil sample was analyzed, and its HC5 was determined against five non-target organisms: a microalga (Pseudokirchneriella subcapitata Korshikov), lettuce (Lactuca sativa L.) seeds, a nematode (Panagrolaimus sp.) and two microcrustaceans (Daphnia magna Straus and Artemia salina L.). In addition, safety factors were applied to the HC5 value in order to establish levels of protection for aquatic communities.

Chemical EO composition
The ratio and type of substances present in EO of a species can vary as a function of several factors, such as growth conditions, altitude, soil type, agricultural methods and practices, developmental stage, part of the plant extracted, harvesting time, climate and fertilization (Moghaddam and Mehdizadeh 2017). Therefore, the chemical composition of the EO used in this study was determined, as a reference for the ecotoxicological analysis.
Chromatographic method: The chromatographic conditions of the standard solutions and of the EO sample were the same. We used a Varian 3900 gas chromatograph coupled to a Varian Saturn 2100T mass spectrometer (GC-MS) and a Shimadzu 2010 (AOC 5000 injector) gas chromatograph with a flame ionization detector (GC-FID). VOL. 51(1) 2021: 71 -78 ACTA AMAZONICA For chromatographic separation, we used an OV-5ms bonded column (30 m x 0.25 mm, 0.25 μm) (Ohio Valley Specialty Co., Marietta, OH, USA). In the GC-MS system, helium gas was used as the carrier gas, whereas in the GC-FID system, nitrogen was used, both with a flow rate of 1 ml min -1 . The optimized parameters were: injector temperature = 220 °C, initial temperature = 50 °C, temperature gradient = 3 °C min -1 , final temperature = 310 °C), range of mass/charge ratio analyzed = 41 -450 m/z, and injected volume = 1 μl, splitless mode, electron impact at 70 eV.

Identification and relative area of EO compounds
The EO compounds were identified by GC-MS by comparison with the mass spectra of the NIST 2.0 library (NIST 2005), through retention index (RI) calculations according to Van den Dool and Kratz (1963) for chromatographic assays with linear temperature programming in relation to the hydrocarbon chain (C7-C40) (Equation 1) and by referring to data in the literature.
where x = chemical compound; n = carbon number of the hydrocarbon eluting before compound x; t x = retention time of compound x; t n e t n+1 = retention times of the hydrocarbons eluting before and after compound x, respectively.
The identified compounds were quantified by GC-FID under the same analytical conditions used in the GC-MS by means of the percentage obtained from the area of each peak relative to the total area of the peaks of the compounds found.

Assay with Pseudokirchneriella subcapitata
In this assay, the effect of the EO on the growth of the microalga Pseudokirchneriella subcapitata was evaluated. The methodology followed Becaro et al. (2015). The algae were cultured in suspension in NPK medium under controlled light (1300 lux) and temperature (20 ± 2 °C) . The suspension was distributed in a 96-well polystyrene culture plate, with a final density in each well of 1.4 x 10 6 cells ml -1 . The algae suspensions were exposed to definitive EO test concentrations of 0.1, 1, 10 and 100 mg L -1 . Twelve replicates (wells) were used for each test concentration. Tween® 80 was used as a solubility adjuvant in the test solutions. Besides, in the same 96-well culture plate, six wells were filled with a control (250 μl of algae suspension and 50 μl of NPK medium + Tween® 80). Algae growth was estimated through cell suspension absorbance (750 nm) measured with a microplate reader (Tecan Sunrise) to calculate the growth rate. Readings were taken at 0, 24, 48, 72, 96, 120 and 168 h.

Assay with lettuce (Lactuca sativa)
This assay evaluated the effect of the EO on the growth of lettuce (Lactuca sativa) seedlings. The assay followed the methodology of Castro et al. (2018). Lettuce seeds were obtained commercially and the trials carried out in the dark for 168 h at (20 ± 1 °C) in 12-well polystyrene plates, according to Bautista et al. (2013). Seeds were exposed to four EO concentrations (0.1, 1, 10 and 100 mg L -1 ) and two controls (distilled water and distilled water + DMSO). A total of 216 seeds were used. Seeds were placed individually in each well containing a Whatman no. 2 paper disk. Three plates were used per test condition, with a total of 18 plates. Dimethylsulfoxide (DMSO) was used as solubility adjuvant, and 0.4 mL test solution was added to each disk. Growth was estimated by measuring root elongation of the emerged seedling. The seeds were observed and photographed every 24 h using an Optika camera 4083B3 coupled to a stereomicroscope. Root size was measured using the software Optika View, Ver 7.1.1.5.

Assay with Daphnia magna
This assay evaluated the acute toxicity of the EO on the microcrustacean Daphnia magna according to OECD guidelines (OECD 2004). The D. magna came from populations kept at the Laboratory of Ecotoxicology and Biosafety at Embrapa Meio Ambiente, bred in aquaria (40 x 20 x 15 cm) containing water reconstituted with nutrients, with pH 7.5, total hardness of 53.58 mg L -1 in calcium carbonate (CaCO 3 ) and conductivity of 111.4 μs cm -1 . The temperature and light intensity were maintained at 20 ± 1 °C and 1000 lux, respectively. The organisms were fed once daily with the alga P. subcapitata (Jonsson and Maia 1999). Neonates less than 24 hours of age were separated from the cultures and used as test organisms.
The EO test concentrations were defined after preliminary testing as 1, 2, 4, 8 and 16 mg L -1 . As controls we used reconstituted water and reconstituted water + Tween® 80 (final concentration: 25 mg L -1 ). For each treatment level and control, one 12-well polystyrene plate was used (seven plates in total), with two D. magna neonates per well in a final volume of 5 mL test or control solution.
The daphnids were considered to be immobilized if they were not able to swim within fifteen seconds after gently moving the test plate (even if they still could move their antennae) (OECD 2004). Immobilization was recorded visually every 24 h during 48 h (two counts) using a colony counter equipment. Test values were compared with control values in order to determine the estimated concentration that immobilized 50% of D. magna in 48h.

Assay with Artemia salina
This assay evaluated the acute toxicity of the EO on the nauplii of the brine shrimp, Artemia salina. The test organisms were obtained from the eclosion of viable Artemia cysts (Maramar®) at the rate of 3 g cysts per liter, as recommended by the manufacturer. The hatching proceeded with the cysts remaining 24 h in a constantly aerated saline solution (3%) VOL. 51(1) 2021: 71 -78 ACTA AMAZONICA prepared with distilled water and sea salt to produce nauplii (Castro et al. 2018). Nauplii were exposed to EO test concentrations under controlled light (1000 lux) and temperature (20 ± 2 °C) conditions (USEPA 2002). The EO test concentrations were defined after preliminary tests as 10, 18, 32.4, 58.3 and 105 mg L -1 . As controls we used saline solution and saline solution + Tween® 80 (final concentration: 25 mg L -1 ). For each treatment level and control, one 12-well polystyrene plate was used (seven plates in total), with two A. salina nauplii per well in a final volume of 5 mL test or control solution. A nauplius was considered immobilized if its appendages and antennae did not move for 10 seconds of observation (Evans et al. 2010). Immobilization was recorded visually every 24 h during 48 h (two counts) using a stereomicroscope (Optika, Italy). Test values were compared with control values in order to determine the estimated concentration that immobilized 50% of A. salina in 48h.

Assay with Panagrolaimus sp.
This assay evaluated the acute toxicity of the EO on nonparasitic nematodes of the genus Panagrolaimus. The assay was carried out according to the methodology by Castro et al. 2018. Synchronized nematode cultures were maintained in oat medium (Lara et al. 2007) at the Laboratory of Ecotoxicology and Biosafety of Embrapa Meio Ambiente, and the assay was performed for 96 h at 24 ± 1 °C in the dark without renewal of the test solutions. The EO test concentrations were defined as 0.1, 1, 10 and 100 mg L -1 . As controls we used K medium (KCl and NaCl; Boyd et al. 2012) and K medium + Tween® 80 (final concentration: 25 mg L -1 ). For each treatment level and control, one 24-well polystyrene plate was used (six plates in total), with ten nematodes per well in a final volume of 1 mL test or control solution. Immobility was assessed for four days by counting the number of exposed nematodes that appeared immobile under a microscope and did not respond to a stimulus using a small metal wire (Castro et al. 2018). Test values were compared with control values in order to determine the estimated concentration that immobilized 50% of Panagrolaimus sp. in 96 h.

Determination of the EC 50
We determined the average and 95% confidence interval of the effective concentration causing immobility to 50% of the test organisms (EC 50 ) in the acute toxicity assays of the microcrustaceans and nematodes during the respective test periods.
Pseudokirchneriella subcapitata cell growth-rate and L. sativa root growth-rate were calculated at the end of the exposure period by calculating the angular coefficients of the linear regression curves (from algae suspension absorbance and root size) as a function of time (Basu and Pal 2011;Becaro et al. 2015). The EC 50 and its 95% confidence intervals was determined for the parameter concerning the phytotoxic effect on the growth rate. The analyses were made using the "Probit Analysis" and "Simple Regression" modules of the Statgraphics Centurion XVII program, version 1.17.04 (StatPoint Technologies 2014).

Estimation of the EO hazardous concentration (HC5)
NOEC was estimated for the five test organisms by considering EC 50/10 (Elmeggard and Akkerhuis 2000). The NOEC values estimated for each test species were used to construct the SSD curve. Based on the SSD, a hypothetical environmental concentration at which only 5% of the species in the aquatic environment would be affected (HC5) was determined by the log-logistic distribution of the NOEC values (OECD 1995;Jonsson et al. 2015) using simple logistic regression in the "Logistic" module of the Statgraphics Centurion XVII program (Statpoint Technologies 2014; Zolezzi et al. 2005). The fit of the distribution of the NOEC values was assessed with a Kolmogorov-Smirnov test. The HC5 was calculated with a confidence level of 50%, which corresponds to the "most likely" estimate of the 5th percentile (Traas and Van Leeuwen 2007;Liu et al. 2016). An estimate of the exposure level at which organisms in the ecosystem would not suffer any damage was calculated by applying safety factors of "1" and "5" to the HC5 value to determine the predicted no-effect concentration (PNEC), which is an ecological safety threshold that represents the critical concentration at which no effects are expected in a given ecosystem (Liu et al. 2016;Castro et al. 2018). The respective safety factor values represent the minimum and maximum limits of a scale from 1 to 5 in a risk assessment (Traas and Van Leeuwen 2007;Mertens 2018).
We also used data for the fish Arapaima gigas Schinz from Queiroz (2012) to calculate the NOEC for this species and added it to the log-logistic distribution. Queiroz (2012) reported that the immersion of A. gigas in a bath of a solution of aqueous extract of P. aduncum at 80 ml L -1 over 24 h did not affect the viability of the fish. As Queiroz (2012) used the aqueous extract, not the EO, we applied an arbitrary safety factor of 1000 to the concentration of 80 mL L -1 and thus, we estimated an EC50-24h > 80 mg L -1 and, consequently, a NOEC = 8 mg L -1 .
NOEC and HC5 values calculated for each organism are shown in Table 2 and the curve fitted to the NOEC values is shown in Figure 2. Derived from the fitted curve, a value of p = 0.8073 was obtained with the Kolmogorov-Smirnov test, which means that the data correspond to a logistic function. The application of safety factors (5 and 1) to the HC5 value (0.47 mg L -1 ; Table 2) allowed us to obtain a PNEC concentration range of 0.09 to 0.47 mg L -1 .

DISCUSSION
Variability in the chemical composition of an essential oil may influence its intrinsic ecotoxicity (Almeida et al. 2017). While we obtained a 75.5% dillapiole composition in our P. aduncum, Silva et al. (2013) reported a 85.6% dillapiole composition in P. aduncum EO from the same locality (Manaus), and Maia et al. (1998) reported a proportion of dillapiole from 31.5% to 97.3% in P. aduncum samples from eight different locations in the Amazon region. This variability in EO composition data indicates that the safety parameters calculated in here are likely specific to the composition of our EO samples, and probably need to be estimated for a wider range of the EO composition spectrum in order to better establish its overall safety parameters.   The higher sensitivity of the microcrustaceans D. magna and A. salina to the EO of P. aduncum found in here is supported by another study. Quignard et al. (2003) evaluated the lethality for Artemia franciscana Kellogg of 74 species of Amazonian plant extracts at a concentration of 500 mg L -1 for 24 h, and P. aduncum was among the most active plants, with lethality above 90%. Microcrustaceans also have high sensitivity to a variety of chemical compounds (USEPA 2002). The short duration (40 -50 days) of the life cycle of Daphnia magna (USEPA 2002) may be related to higher sensitivity, as the usual experimental exposure periods to stressor agents of one or two days represent a higher exposure relative to their life span. The sensitivity level also depends on the development phase, tested concentrations, exposure time and species used (Rand and Petrocelli 1985).
Other studies have also observed the deleterious effect of Piper sp. compounds against arthropods. The EO of P. aduncum had lethal effects on larvae of the coleopteran Tenebrio molitor at concentrations above 2.5% v.v -1 (Fazolin et al. 2007). Piper aduncum extract at a concentration of 0.1 mg L -1 showed the highest insecticidal activity among three Piper species, promoting 92% mortality on larvae of the lepidopteran Ostrinia nubitalis (Bernard et al. 1995). These studies suggest the insecticidal action of dillapiole, as it is the main compound in the chemical composition of P. aduncum oil.
No adverse effect of the EO was observed on L. sativa seeds, in contrast to the study by Alves et al. (2004), who identified the allelopathy of different EOs on L. sativa germination and root length, and demonstrated the inhibitory potential of volatile extracts of cinnamon (Cinnamomum zeylanicum), pepper-rosmarin (Lippia sidoides), lemon grass (Cymbopogum citratus) and clove basil (Ocimum gratissimum) for L. sativa seeds. There is no obvious explanation for the lack of phytotoxicity of P. aduncum EO on plant seeds and algae in our study, even at the highest concentration of 100 mg L -1 . It may be related to the inhibitory cholinesterase effect that was described for compounds of several Piper species (Silva et al. 2017), which can have appreciable effects on arthropods, in contrast to plants and algae.
No toxicity data and NOEC estimates of P. aduncum EO exist for fish. However, Queiroz (2012) reported that the immersion bath in a solution of aqueous extract of P. aduncum (80 mL L -1 for 24 h) did not affect the viability of Arapaima gigas. Due to the fact that Queiroz (2012) used aqueous extract, not the EO, we applied an arbitrary safety factor to the concentration of 80 mL L -1 of aquous extract for the estimation of the NOEC. The attribution of non-toxicity of P. aduncum EO to fish could also be reinforced by the absence of adverse effects in A. gigas treated at a high oral dose (80 mL kg -1 bw) for 15 days (Corral et al. 2018).
The PNEC concentration range obtained (0.09 -0.47 mg L -1 ) justifies the application of a factor in the range of 1 to 5 to the HC5 on a case-by-case basis for practical purposes in establishing a maximum permissible concentration for a water body. The factor depends on environmental evidence supported by data on the presence of nonnative species, the number of species involved and any gaps between laboratory and field data (Traas and Van Leeuwen 2007;Liu et al. 2016).

CONCLUSIONS
The essential oil of Piper aduncum differentially affected the growth and mortality of target organisms of different trophic levels, with microcrustaceans showing higher sensitivity than plant seeds, algae and nematodes. The safest concentration estimated indicated for use of this oil in aquaculture production without compromising the co-existing biota was 0.09 mg L -1 . This concentration may be considered a preliminary result, pending on further ecotoxicological evaluations of other target organisms and a wider range of chemical composition of P. aduncum essential oil to improve the risk estimate.