Infection parameters of Heterorhabditis amazonensis (Nematoda: Heterorhabditidae) in different stages of Hibiscus pink mealybug

Abstract The pink hibiscus mealybug Maconellicoccus hirsutus (Hemiptera: Pseudococcidae) is an invasive pest of an enormous variety of crops and has become a concern in many parts of the world. Early attempts to control M. hirsutus with chemical insecticides and cultural methods have failed due to the cryptic habit of the insect. We assessed the entomopathogenic nematode Heterorhabditis amazonensis as a biological agent against different insect stages. Comparing different concentrations of the nematode, insect females were very susceptible, with more than 90% of the insects killed. In second and third nymphal stages mortality rates varied from 20 to 60% depending on the nematode concentration. The first nymphal stage as much less susceptible to nematodes due to their small size. The number of nematodes capable of invading the insect host did not vary between the different concentrations. However, the LC50 for females (35.2 IJ/insect), second and third nymphal stages (83.9 IJ/insect) demonstrated that H. amazonensis should be considered as a potential biocontrol agent of the pink hibiscus mealybug.

Initial measures to control the pest were done using chemical insecticides and cultural methods but their efficiency was low due to the cryptic zones of the plants where the insects hide and also because their impermeable waxy cover complicates the penetration of insecticides (Sagarra and Peterkin, 1999;Kairo et al., 2000;Chong et al., 2015). For these reasons, biological control by means of parasitoid wasps has been gaining acceptance. Among those natural enemies, Anagyrus kamali Moursi (Hymenoptera: Encyrtidae) has shown remarkable results in controlling PHM (Kairo et al., 2000).
In 1996, an initiative proposed by CARICOM (Caribbean Community) and Venezuela was approved to design a long-term sustainable program for control of PHM in the Caribbean region. Over time, most the Caribbean countries developed successful biological control programs (Kairo et al., 2000). Venezuela was an exception because of lack of implementation of insect mass production in laboratories, among other causes (Cermelli et al., 2002). For those reasons, there is a need to study biological alternatives to reduce the populations of PHM in Venezuela.
Entomopathogenic nematodes (EPNs) have been used in biological control programs for crops worldwide because of their effectiveness, time of response, innocuousness to mammals, and relative simplicity of mass production (Kaya et al., 2006). Some laboratories in Venezuela work actively on these organisms (San-Blas et al., 2015), which are currently available in small quantities for biological control programs (San-Blas, Luzardo, Larreal, Portillo andBastidas, 2017, San-Blas, Luzardo, Portillo, Fuenmayor andBastidas, 2017) and production of these biocontrol agents is expected to rise shortly (San-Blas et al., 2019).

PHM breeding
PHM-Infected fruits of Annona cherimola Mill. (Magnoliales: Annonaceae) plants were collected from the field and carried to the laboratory in plastic boxes (1 L). Females were detached from the fruits and ovisacs separated from females. Ovisacs were placed in an Eppendorf tube (1 ovisac per tube) and incubated at 28°C. Daily the tubes were checked until hatching was observed. The new-born nymphs were collected using a small brush and set over a pumpkin (bought from the local market) previously disinfected with 10% sodium hypochlorite and placed in a clean plastic box with a perforated lid (28°C; 75% RH; 12 hr light). The life cycle was monitored every 48 hr and the procedure was repeated when the pumpkin was consumed by the insects. In this stage, all instars were present, so the insects were collected and separated by instars on the day of the experiment.

Nematode culture
The nematode H. amazonensis was cultured in the fourth instar larvae of Galleria mellonella (Lepidoptera: Pyralidae) following Dutky et al. (1964) technique. The infected larvae were incubated at 25°C. Infective juveniles were collected in White traps (White, 1927) and stored at 20°C until the experiment proceeded (no more than 2 weeks).

Virulence, pathogenicity, and mortality determination of different instars of PHM to Heterorhabditis amazonensis
Multi-well plates (six wells) were filled with 250 mg sterile sand per well. Randomly, 40 µL of a nematode suspension containing either 0, 10, 40, 60, 80, or 100 IJs/insect were placed over the sand. In every well, five insects were placed and covered with Parafilm ® and then aluminum foil. The procedure was repeated for 1st (crawlers), 2nd, and 3rd nymph instars and adult females. The plates were placed in humidity chambers (a plastic box with moistened tissue paper in the bottom) and incubated at 28°C. After 96 hr, the dead insects were counted and dissected. The number of nematodes (J4 and adults) found inside the insect cadavers was recoded to evaluate nematode infection parameters (see below). The experiment was repeated four times. As no mortality was found in the control treatments (concentration = 0), no adjustment of mortality was done.
The percentages of mortality and host infection (number of nematodes found inside the cadaver out of the total nematodes set in the experiment) were angular transformed (arcsine of the square root of the proportions presented in degrees) due to the binomial nature of the data and ANOVA tests were performed to assess differences of host infection of each insect instar at varying IJ concentrations; if differences were statistically significant, a family LSD test was done. Data in figures and tables were presented untransformed (% ± s.e.m.). The penetration rate of infective juveniles per insect was determined using Glazer and Lewis (2000) formulae for different concentrations. The results were angular transformed and treated as above. Data in the figures were presented untransformed (% ± s.e.m.).
To evaluate differences between number of invader nematodes per insect instar and per nematode concentration, the resulting numbers of nematodes found inside the insect cadaver were used untransformed (including 0 values) and a Poisson distribution confirmed (data not shown), followed by a Mood's median test (Siegel and Castellan, 1988). Data were presented as medians (first quartile, third quartile). Probit calculations were performed to assess lethal concentration of number of nematodes to kill 50 (LD 50 ), 75 (LD 75 ), and 90 (LD 90 ) percent of the insects per instar.
This difference (developmental stage/nematode concentration) has been observed previously and been used for biological control programs to targeting special insect instars (Schroeder et al., 1996;Kakouli-Duarte et al., 1997). In general, the use of EPN to control other pseudococcids such as Dysmicoccus. vaccinii, D. brevipes, Planococcus citri, and P. ficus resulted in higher mortality rates of mature females (Stuart et al., 1997;van Niekerk and Malan, 2012;le Vieux and Malan, 2013;Ferreira et al., 2015).
The variation in mortality rates may be attributed to the size of the instars (Stuart et al., 1997;Bastidas et al., 2014). In our results, the second and third nymphal stages were also susceptible to H. amazonensis, but less so compared to adult females. There was also significantly less infectivity of EPNs to the 'crawlers'. In micro hosts (less than 5 mm), the penetration, development, and reproduction of EPNs are compromised due to a series of factors related to the sizes of both the insect and the nematode (Bastidas et al., 2014).
In this case, the size of the PHM nymphal stages limited the penetration and infection rates of the nematodes. For example, the penetration rates of H. amazonensis against PHM females (Fig. 2) did not vary significantly according to the concentration (F (4,95) = 2.16; p = 0.079). However, the number of individuals able to invade insects was higher (median   (Table 1). In contrast, the percentage of nematodes able to invade nymphal stages (second and third instars and crawlers) did not show significant differences between the concentrations of nematodes applied to the insects (F (4,95) = 1.08; p = 0.369 and F (4,95) = 0.62; p = 0.647, respectively) (Fig. 2). The median number of IJs found inside the host cadavers did not show any relation between the nymphal stages and the concentration (χ 2 = 8.46; d.f. 4; p < 0.076 in the second and third instars and χ 2 = 4.49; d.f. 4; p < 0.076 for crawlers) and ranged between 4 and 12 individuals per insect (second and third instars) and one IJ per crawler (Table 1). In fact, the number of nematodes capable of penetrating the nymphs was the same even though the EPN concentrations increased. This constant pattern of nematode penetration has been observed previously when the host size is reduced, because there is no available space for more nematodes inside the insect body (Gouge and Hague, 1995;Bastidas et al., 2014).
In terms of the potential to use H. amazonensis as biological control of PHM, it is evident that females should be the only stage considered to be part of a pest management program using EPN according to the probit models. Because the number of nematodes necessary to kill a given percentage of the insect population must be economically feasible. In our results, the lethal concentrations for females (Fig. 3A) ranged between 34.45 and 91.47 nematodes/insect (LD 50 and LD 95 respectively) ( Table 2), whereas the number of nematodes required to kill the nymph stages was too high to be considered as biocontrol mean (Fig. 3B, C) ( Table 2). The number of H. amazonensis required to kill 50 or 95% of the adult females was similar to results reported previously for pseudococcids (van Niekerk and Malan, 2012;le Vieux and Malan, 2013;Ferreira et al., 2015), with the exception of D. vaccinii which needed 500 IJ/ insect of H. bacteriophora to reach 80% of mortality (Stuart et al., 1997).
The spread and crop damage of PHM is increasing in the tropics, requiring more research aimed at insect control. Some biological control    techniques are being successfully used to reduce PHM populations in some countries but EPNs have never been tested. Cryptolaemus montrouzieri Mulsant (Coleoptera: Coccinellidae) is an Australiannative ladybird which was successfully released in the Caribbean and India but in Egypt, the beetles are unable to survive winter conditions (Kairo et al., 2000;Mani et al., 2011). Another important biological control agent is the parasitoid Anagyrus kamali Moursi (Hymenoptera: Encyrtidae), which has also been released in some Caribbean islands and Egypt with goods results, in the absence of native enemies (Kairo et al., 2000). As EPNs are ubiquitous organisms (Griffin et al., 1990), and have proved to be effective to control some nymphal (second and third) stages and adults females of PHM, EPNs should be considered promptly in integrated pest programs to control this serious pest.