Initially, pesticides were always obtained from plants. Gonzalez-Coloma et al. (2010) showed Eugenol to be obtained from laurel (Laurus species), azaridachtina from the neem tree (Azadirachta indica), nicotina from Nicotiana tabacum, and karanjina from Derris indica Karanjina. It also showed insecticidal vegetable oils to be obtained from Brassica napus, Brassica campestris, Capsicum species, Tagetes erecta, Thymus vulgaris and Gaultheria procumbens, insecticidal acids from Varroa destructor and Pyrethrins, chrysanthemats and pyrethrates from Tanacetum cinerariaefolium flowers. Rotenone from the genera Derris, Lonchocarpus and Tephrosia. Natural commercial insecticidal extracts were obtained from the stem of Ryania speciosa, crushed seeds of Schoenocaulon officinale, potato and corn starch syrup (Gonzalez-Coloma et al., 2010). Also, soluble powder of Derris is marketed with different concentrations of rotenone (Undersun Biomedtech, 2022; Yates, 2022).

Yam bean (Pachyrhizus species, Fabaceae), also called jacatupé in southeastern Brazil, jícama in Mexico and Central America and bengkoang in Indonesia is a promising Amazonian plant that could quickly be used as a natural pesticide. It is well adapted and domesticated to the humid tropics due to having been cultivated for the fresh consumption of its white, slightly sweet roots. However, little importance has been given to its seeds, stem and leaves toxic properties. Currently, toxicity is known to be more highly concentrated in seeds and, tests on living beings have shown toxicity to be present on cells in vitro (Estrella-Parra et al., 2014), humans (Sorensen, 1996; Yu et al., 2020), insects (Bejar et al., 2000), mites (Bejar et al., 2000)), fungi (Barrera-Necha et al., 2004), fishes (Sorensen, 1996), and viruses (Phrutivorapongkul et al., 2002).

This bean’s seeds hold 20 iso-flavonoids, mainly rotenoids, with rotenone (Bejar et al., 2000), showing to be the most toxic one among them. This substance has been well known since the beginning of the 20th century on account of its insecticidal action (Roark, 1945a, 1942b, 1943c, 1943d, 1944e, 1945f) and, has already been widely used to control ants (Dolichoderus bidens) and the Mediterranean fruit fly in the USA (Ceratitis capitata) (De Moura and Schlichting, 2007). However, it was extracted from the roots of timbó (Derris species = Lonchocarpus species) and from Tephrosia (Dutta et al., 2019).

Rotenone is a lethal substance due to two reasons: (i) it inhibits electron transport in complex I of the mitochondrial respiratory chain, by blocking the ATP (Catteau et al., 2013)  manufacture,   and  (ii)  produces  reactive  oxygen species, causing oxidative stress (Mohammed et al., 2020).

There are some advantageous aspects in the use of yam bean seeds. First, the ease to be produced under Amazonian conditions, since it needs neither fertilization nor disease and pest control. Second, it has a moderate rotenone content, which ranges from 1.13 to 2.76 mg g^-1 (Dutta et al., 2019). Third, the toxicity of oral rotenone in animals and humans is low. In rats, orally, the LD50 was 60 mg kg^-1 body weight; intravenously, LD50 was 0.2 mg kg^-1, and intra-peritoneally, LD50 was 1.6 mg kg^-1. In humans, orally, LD50 ranges from 300 to 500 mg kg^-1 body weight (Gonzalez-Coloma et al., 2010). Reports of accidental poisoning in humans have been observed when 40 to 80 fresh seeds are eaten (Yu et al., 2020; Fu and Wang, 2012). Only death was observed when 100 seeds were eaten (Narongchai et al., 2005).

The few Pachyrhizus spp. genotypes agronomically and physic-chemically (Silva et al., 2016) characterized (Vasconcelos et al., 2018) in the Amazon are found in the germplasm bank of the National Amazonian Research Institute (INPA), which conserves 64 progenies. Genetic variability was observed in them, which indicates that there could be genetic variability for toxicity. Therefore, it should be studied in this aspect, as well as tested about its effect on some important phytopathogens.

There are two important phytopathogens in the Amazon that cause death at any stage of growth. They attack the neck of the plant, which prevents nutrients from being taken to the aerial part and processed into photo-assimilates. These are the fungus Sclerotium rolfsii in cubiu (Solanum sessiliflorum Dunal) and the bacterium Ralstonia solanacearum in tomatoes. In both cases, control is difficult, and even in resistant varieties, some plants may die.

This work aimed to evaluate toxicity genetic variability in both mature and immature seeds using Pheidole species ants as well as test the toxicity of mature seeds on S. rolfsii and R. solanacearum.

The genetic variability and toxicity experiments on S. rolfsii and R. solanacearum were carried out at the Phytopathology and Vegetable Breeding Laboratories, of the National Amazonian Research Institute (INPA).

Pachyrhizus toxicity genetic variability

Two experiments were carried out using mature and immature seeds flour. This flour was obtained from seeds milled in Wilye Type Micro Knife Mill. The extracts were prepared by flour maceration in filtrated water and then filtered using grade 4 filter paper. The final concentration was adjusted at 1:1000 (1 mg L^-1). The immature seeds, from 22 progenies, were obtained from green pods and, were green-colored as well. In contrast, mature seeds were obtained from dry pods and their color varied from brown to light beige.

Each experiment followed the completely randomized design with three repetitions. One Petri plate (diameter=10 cm) per experimental unit. Each plate was filled with 15 ml of extract then 20 ants were put in it.

Ants (Pheidole species) were picked up using plastic cups with sausage slices (Estrella Alimentos ®) outside buildings in INPA/Campus 3. After 0.5 to 1 h cups were taken up to the laboratory. Toxicity was assessed by counting the dead ants every 10 min for 60 min. Lethality percentage was calculated as [1- SN/TN] × 100, where SN= survivor number and TN= Total number of ants.

Toxicity on S. rolfsii

Six potato-dextrose-agar (PDA) culture media were prepared using 200 g potato, 10 g dextrose (Biotec Reagentes Analíticos) and 20 g agar (Agargel) to 1 L of distilled water. This was mixed with aqueous extracts from mature seeds of genotypes P8, P40 and P49. The final concentrations were for P8 1:500 (2g L^-1), P8 1:1000, P40 1:1000, P49 1: 1000, PDA (control), and Cabrio® Top (4g L^-1) (control).

This was followed by a completely randomized design with treatments in a 6×5 factorial scheme (six culture media) and five observation times (two, three, four, five and seven days) with three replications. The experimental unit was a Petri dish (9 cm) holding approximately 15 ml of solution. S. rolfsii sclerotia was placed in the center of the plate with a forceps. The mycelium diameter was measured on the fifth and seventh days.

Toxicity on R. solanacearum

Yeast-peptone-glucose (YPG) liquid culture was prepared using 5 g yeast extract (Kasvi), 5 g peptone (BD Difco), and 5 g glucose (Biotec Reagentes Analíticos) to 1 L of distilled water, pH=7 (Kpêmoua et al., 1996). This was then mixed with aqueous extracts of the genotype P40. The mixing was carried out so the final concentrations would be 1: 200, 1: 1000 and 1: 2000. Three R. solanacearum isolates, were used: V1 (Biovar 3), V13 (Biovar 2), and V15 (Biovar 1).

Thus, the experiment followed a completely randomized design, with treatments in a 5×3 factorial design (three extract concentrations + two controls) × (three R. solanacearum isolates) with 24 repetitions. The experimental unit was a 96-well Elisa microtiter plate. The two controls mentioned earlier were YPG+Tetracycline® (240 mg L^-1) and YPG by itself. 50 µl of each solution were placed in each well with the aid of a micropipette and the bacterium R. solanacearum was added. Absorbance was assessed through the iMark Microplate Reader spectrophotometer, every 24 h for three days.

Statistical analyses

All data were submitted to analysis of variance to determine the significant effect of the extracts. The averages of the treatments were compared using the Duncan test (P<0.05). The SAS 9.0 software, PROC GLM procedure was used (SAS Institute, Cary, NC). To observe the genetic diversity of yam bean toxicity, standardized averages of immature seed toxicity in Pheidole were used to construct a dendogram. This standardization was done for each time (10 to 60 min). For this purpose, genotypes were clustered based on Euclidean distances (bootstrap=1,000 samples) using unweighted pair group method with arithmetic mean (UPGMA) method. The Darwin 6.0 software was used (Perrier and Jacquemoud-Collet, 2006).

Many rotenone and rotenoid toxicity tests on insects have been carried out since the early 20th century (Roark, 1945a, 1942b, 1943c, 1943d, 1944e, 1945f). These toxins were extracted mainly from the Derris genera. However, the rotenoids from Pachyrhizus sub-species were not tested sufficiently on plant’s pests and diseases. In this work, the toxicity genetic variability of genotypes Pachyrhizus was assessed on Pheidole spp. ants, as well as, the effect of Pachyrhizus extracts over the fungus S. rolfsii and the bacterium R. solanacearum.

Yam bean toxicity genetic variability

Usually, toxicity tests have been carried out on several species such as the arthropod Artemia salina, zooplankton Daphnia magna, the worm Eisenia andrei and the collembola Folsomia candida (Danabas et al., 2020; Bandeira et al., 2020). There are still other species being used to determine chronic or acute toxicity in different environments. Yet, these methods need that these species be bred. In this work, Pheidole spp. ants were chosen for two reasons. First, because they are abundant in the Amazon region and are easy to capture. Second, they are abundant around the world and their control is difficult (Ali and Ali, 2020).

The results showed the mature seeds to present 100% lethality on Pheidole spp. after 10 min of treatment (Table 1). However, immature seeds had approximately the same lethality after 60 min. This indicates mature seeds to be six times more toxic (Tables 2 and 3).

The analysis of variance showed a significant effect for genotypes, which indicates there to be toxicity genetic variability in immature seeds at 10 to 60 min (Table 1). This result was observed in a dendogram (Figure 1). With 50% of dissimilarity, only two clusters were observed, but with 20% of dissimilarity, six clusters were observed. P9, P58, P7, P10, P61 and P28 showed no genetic variability for toxicity. However, water, P5 and P50 are a distant cluster from others. All these results indicate that the most toxic genotypes could be selected following 10 min.

Average lethality of immature seeds at 10 min of exposure ranged from 5.5 to 100.0% (Table 2). This indicates there would be genetic variability for rotenone or other toxic substances. P9, P58, P7, P10, P61 and P28 showed to be the most toxic genotypes all with 100% lethality. Conversely, P50 (5.5%) and P5 (8.5%) were the least toxic ones. Thus, P50 and P5 could be edible as long as they are cooked (Bidwell, 2020), dried or roasted (Catteau et al., 2013), or soaked by changing the water as it is usually in lupine (Lupinus spp.) preparation.

The analysis of variance also showed significance for the 'extracts versus control' contrast (Table 1). It suggests immature seeds to have a toxic action. This toxicity was observed on humans who had  eaten  40  to  100  cooked Pachyrhizus immature seeds (Yu et al., 2020; Narongchai et al., 2005; Silva et al., 2016). They had diarrhea, dyspnea, unconsciousness or death (Silva et al., 2016).

Mature seeds presented no significant difference on their genotype-induced lethality following 10 min (Table 3). Yet, a difference was  detected  between  genotypes  and control. Lethality averages for genotypes and control were 100 and 27%, respectively. This indicates all these genotypes to bear high toxicity.

Pachyrhizus seeds bear isoflavonoids and coumarins (Jiménez-Martínez et al., 2003). Rotenone and pachyrrhizin are each group’s most toxic ones. Thus, the genetic variability can be for these substances, one of them or their interaction.

S. rolfsii

The analysis of variance (Table 4) and Duncan test detected a significant effect of the treatments on the S. rolfsii growth. The mycelial diameter after five days ranged  between  20  and  60 mm  in  culture  media  with Pachyrhizus seed extracts (1:1000= 0.1%). Whereas mycelial diameter was 75 and 0 mm in PDA and PDA+Cabrio® Top controls. Thus, Pachyrhizus extracts reduced by 73% the mycelial growth.

P40 extract, at 1: 1000 concentration inhibited mycelial growth by 73%, following five days of cultivation (Table 5). P8 extracts at concentrations 1: 500 and 1: 1000 reduced mycelial growth by 32.6 and 48.9%. This shows the less concentrated extract to bear greater toxicity.

On the other hand, Cabrio® Top was shown to have inhibited the mycelium growth. Thus, its use could be recommended to control this fungus. Even though there is no recommendation for doing so. Cabrio® Top is an agricultural pesticide indicated against various fungal diseases. Its composition holds two pesticides: (i) pyraclostrobin which inhibits the mitochondrial respiratory chain, and (ii) metiram that reacts non-specifically with fungus sulfhydryl enzymes (Registration 01303, from the Ministry of Agriculture, Livestock and Supply, Brazil).

Rotenone and Cabrio® Top bear some similarities in action mechanism (Catteau et al., 2013). Both prevent the mitochondrial respiratory chain and the production of ATP in the fungus. Thus, Pachyrhizus seeds could be used to control this soil fungus. They could be applied in the pit as extracts or powder prior to placing the seedlings.

R. solanacearum

The best way to assess bacterial multiplication is by absorbing light. To standardize the absorbance  readings, decision was made to estimate the difference between the 48 and 24 h ones (Table 6). The results showed the absorbance differences for the extracts to have ranged from 0.29 to 0.51. In controls they were 0.06 (Tetracycline) and 0.68 (without extracts and no Tetracycline). This indicates these extracts to have reduced the multiplication of R. solanacearum by 17 to 39%. The extract holding the concentration of 1: 200 showed to be the one which exerted the most toxic effect for R. solanacearum.

Likewise, biovar 3 showed the largest difference in absorbance (0.52), suggesting it to be the most virulent. In contrast, biovar 1 showed the smallest difference in absorbance, which indicates it to be less virulent.

It is very difficult to cultivate tomato in the Amazon region, mainly due to the R. solanacearum. Yet, there was the singular case of the success of tomato production in Central Amazon. It was carried out at the Agro-industrial Adventist Institute in the 1970s and showed to yield up to 53 t ha^-1 (Prance, 1989). According to Dr. Noda, methyl bromide was used, in this institute to sterilize  the  soil  (Oliveira, 2015). But this substance was prohibited because it was harmful to human health and the environment.

Therefore, Pachyrhizus seeds can be used as bactericide and insecticide in tomato cultivation mainly to control R. solanacearum and the mole cricket (Gryllotalpa brachyptera), which seems to be its vector.

There is genetic variability for toxicity in Pachyrhizus spp. It opens the possibility of screening all germplasm of this genus. The most toxic genotypes can be selected and included in a genetic improvement program. The aim of which should be to maximize seed and rotenoids yield.

Pachyrhizus seeds control R. solanacearum and S. rolfsii. Thus, Pachyrhizus-based products can be used for preventing soil diseases. Nevertheless, some novel methods, for treating the soil through the use of these biopesticides, will have to be developed.

The authors have not declared any conflict of interests.