Isolation of entomopathogenic nematodes in an apple orchard in Southern Brazil and its virulence to Anastrepha fraterculus (Diptera: Tephritidae) larvae, under laboratory conditions

Entomopathogenic nematodes (EPNs) are a promising alternative to integrated control in many fruit pests. Few studies were made on the relationship of Anastrepha fraterculus natural population with native EPNs population and other biotic and abiotic factors. The aim of this work was to verify the occurrence of endemic nematodes in an apple orchard, concerning environmental conditions and technical procedure, and access isolates virulence to A. fraterculus larvae. The experiment was conducted during a year taking monthly soil samples from an apple orchard, with and without fallen fruits just above the soil. Samples were baited with Tenebrium molitor and A. fraterculus larvae in laboratory. Canopy and fallen fruits were sampled to access the pest infestation. Seventy three EPN isolates were captured, in 23.2% soil samples, more with T. molitor than with A. fraterculus baits. From the 20 isolates tested against A. fraterculus, only five were pathogenic, and they were identified as Oscheius sp. The nematodes were captured during all seasons in a similar frequency. Soil and weather conditions, presence of fruit over the orchard soil, and A. fraterculus pupae in the fruits had no significant influence on the capture. As a conclusion, nematodes of the genera Oscheius are found in an apple orchard of Porto Amazonas constantly along the year, independently of fluctuations in A. fraterculus population, climate conditions and presence of fruit over the soil. Some of the isolates are pathogenic to A. fraterculus.


Introduction
South-American fruit fly, Anastrepha fraterculus (Wied.)(Diptera: Tephritidae), has a wide distribution in America and can be found in almost all parts of South America.In southern Brazil, the species domains and can harm at least 67 hosts, including the exotic ones like apple (Malus domestica Borkh.)(Zucchi, 2000a).In laboratory condition, the adult can survive six months and the female is able to lay more than 400 eggs during lifetime (Salles, 2000).The larvae feed on fruit pulp and induce its earlier ripening and falling (Souza-Filho et al., 2009).On soil, late third-instar larva leaves the fruit and turns to pupa nearby into the soil.Last larval instar and pupal phase takes 15 days at 25 °C (Salles, 2000) and this is the suitable time for many soil biocontrolers to infect A. fraterculus (Rodrigues-Trentini, 1996).
Entomopathogenic nematodes (EPNs) are having promising results on biological control and are an alternative measure for many soil pests control (Stock, 2005).There are three different genus mostly known as EPNs: Steinernema, Heterorhabiditis and Neosteinernema.They are entomopathogenic bacteria vectors (Poinar-Junior, 1990).More recently the genus Oscheius was discovered as EPN (Torres- Barragan et al., 2011).There are few studies on demonstrating Oscheius contribution to biological control, since most species of this genus have facultative-parasite habit (Ye et al., 2011).Oscheius sp. have symbiotic relationship with entomopathogenic bacteria, caring them not inside its body like obligate EPNs do, but in its body surface.When bacteria are present, Osheius sp. are able to infect and kill insect species (Torres-Barragan et al., 2011), but they were not reported infecting A. fraterculus yet.In this work, for convenience, all isolated are called EPN, including Oscheius.EPNs have been studied as biocontrolers candidate of several fruit flies species (Lezama-Gutiérrez et al., 2006;Barbosa-Negrisoli et al., 2009;Rohde et al., 2012;Toledo et al., 2014).However, few studies at orchard condition were found in the literature and even less works reporting EPNs natural occurrence.In a Florida citrus orchard, the increase on EPN population was correlated only with soil temperature (Beavers et al., 1983).McGraw and Koppenhöfer (2009) found a positive correlation between the increase of native EPNs population and the host Listronotus maculicollis Kirby (Coleoptera: Curculionidae) in turfgrass.
The aim of this work was to verify the occurrence of isolate endemic entomopathogenic nematodes strains, characterize their seasonality and relation with biotic and abiotic factors in an apple orchard of Porto Amazonas, PR, Brazil, and select strains capable of infect and kill A. fraterculus in laboratory conditions.
The relationship of the proportion of EPN-positive samples with clay, sand and organic matter soil content, precipitation cumulated and mean air temperature two until six days before sampling, was analyzed through Spearman correlation.
Anastrepha fraterculus population fluctuation was determined by monitoring fruit fly apple infestation.Fruit samples were collected from July 11, 2013 to July 03, 2014.Samples from each sector were composed by twenty fruits (from four plants of a row) every week during harvest season and six fruit (from three plants) biweekly after and before harvest.Fallen fruit were sampled the same way when available.In laboratory, fruit were laid on vermiculite according to Nascimento et al. (2000).Pupae were placed in Petri dishes with filter paper and moist with distilled water until fly emergency.Two days after emergency, flies were registered and stored in ethanol 70% to identification (Nascimento et al., 2000;Zucchi, 2000b;Marsaro Júnior et al., 2013).
The relationship of the proportion of EPN-positive samples with A. fraterculus pupae was analyzed through Spearman correlation.The proportions of EPN-positive samples in the subsets with and without fruit upon the sample site were compared by chi-square tests.The software BIOESTAT 5.0 was used to both tests (Ayres et al., 2007).Additionally, data on EPN capture were split by season and the proportion on EPN-positive samples was compared through a chi-square test (Campbell et al., 1995).

Virulence to A. fraterculus
An A. fraterculus population from Cena (Centro de Energia Nuclear na Agricultura, Piracicaba -Brazil), was reared following Walder et al. (2014) to obtain late third-instar larvae.They were used in an experiment completely random designed, with five repetitions composed by a nine-centimeter Petri dishes (Rohde et al., 2012) with 12 larvae, in a dark climatic chamber (24±2 °C, U.R. 70±10%).Twenty nematode isolates from Porto Amazonas apple orchard, pathogenic to G. mellonella, were tested split in three trials, each one including a negative control (distilled water) and a positive control with Steinernema carpocapsae CB 02 (provided by Instituto Biológico de Campinas), multiplied as in Voss et al. (2009).The IJ concentration was determined by counting 20-microliter drops with five replicates, using an optical microscope (Voss et al., 2009).Each larva (except in water control) was treated with 100 IJ (Barbosa-Negrisoli et al., 2009) suspended in 0.6 mL distilled water.The number of flies emerged was registered twenty five days after inoculation.Dead insects were dissected under amplification for confirmation of infection by nematode (Barbosa-Negrisoli et al., 2009).The control efficiency was calculated according to Abbott (1925).Data were square-root transformed to reach normality and submitted to analysis of variance and Tukey test using R 3.02 (R Development Core Team, 2008).

Isolates identification
Five isolates were selected based on its virulence to A. fraterculus and sent to identification by Dr. Tesfamariam Mengistu at Florida University.DNA was extracted from five first generation female nematodes from each isolate placed in 1.5-mL Eppendorf tubes each containing 20 μl 0.25 M NaOH and incubated at 25 °C overnight (Floyd et al., 2002).Thereafter, samples were incubated at 99 °C for tree minutes and allowed to cool to room temperature, and then 4 μl 0.25 M HCl, 10 μl 0.5 M Tris-HCl, (pH 8.0), and 5 μl 2% Triton X-100 were added to each tube.Samples were vortexed, spun down, incubated at 99 ºC for tree minutes, cooled to room temperature, and stored at −20 °C until further use.ITS-rRNA gene was amplified using the ITS primers TW81 (5'-GTTTCCGTAGGTGAACCTGC-3') and AB28 (5'-ATATGCTTAAGTTCAGCGGGT-3') (Subbotin et al., 2001) by PCR in an Applied Biosystems 2720 Thermocycler (MJ Research, Waltham, MA, USA).The PCR conditions were: denaturation at 94 °C for five minutes; followed by 35 cycles of 94 ºC for one minute, 52 ºC for one minute, and 72 ºC for one minute; and a final extension at 72 ºC for 10 minutes.Eight microliters of the amplification product was electrophoresed on a 1.8% agarose gel and stained with ethidium bromide.The sizes of amplified products were determined by comparison with a 1 kb molecular weight ladder (Invitrogen).For direct sequencing, PCR products were purified with the QIA quick PCR purification kit (Qiagen) and sequenced.DNA sequences were compared to sequences from the GenBank National Centre for Biotechnology Information (NCBI) using BLAST search with standard algorithm parameters (NCBI, 2015).

EPN capture
Seventy eight pathogenic isolates were captured from a total of 315 soil samples.Seventy tree samples were positive for EPN (23.2%), because some samples were positive in both the trap types.The baits with T. molitor captured more EPNs than the ones with A. fraterculus (Chi-square test, p< 0.05) (Figure 1A).EPNs were captured in 63% T. molitor baits against 15% with A. fraterculus.
EPN-positive samples without fruit over the soil were 21% of the total against 25.7% with some fruit (Figure 1B).The difference was not significant (Chi-square test, p= 0.39).Furthermore, there was no significant correlation between the number of A. fraterculus pupae in the fruits of the sector and the proportion of positive samples (r= 0.52; p= 0.09) (Figure 1C).Ninety one percent of the pupae emerged from fruits gave rise to Anastrepha sp.and, among them, all females were A. fraterculus.In the same way, the proportion of EPN-positive samples among the year seasons was not significantly different (Chi-square test, p= 0.36), besides the differences in air temperature which fell during winter to 16-12 ºC (Figure 2A).
The correlation between cumulated rainfall two to six days before sampling and the proportion of EPN-positive samples was not significant (r= -0.01; p= 0.96).Some rain happened in the six days preceding soil collection in almost all evaluations (Figure 2B).No rain occurred before the sampling in November 2013 and August 2014 and EPNs were captured as well.
There was no correlation between sand or clay content and proportion of EPN positive samples (r= 0.26, p= 0. 62 and r= -0.8, p= 0.10, respectively).In the same way, no correlation was observed between soil organic matter content and the proportion of EPN positive sample (r= -0.39, p= 0.27).

Virulence against A. fraterculus and isolates identification
Among 20 isolates tested, native from Porto Amazonas apple orchard soil, five caused significantly higher mortality of A. fraterculus larvae than the control without nematode: 158, 222, 288, 304 and 319 (Figure 3).Isolates 304 and 319 scored virulence of 73 and 71%, respectively (Figure 3B).For the other tree pathogenic isolates, A. fraterculus mortality ranged from 33 to 54%, also differing from the control (Figure 3A and 3C).Among the five pathogenic isolates, two (158 and 288) had significantly worse results on attacking A. fraterculus larvae than S. carpocapsae CB 02, which killed all the larvae exposed.

Discussion
Tenebrio molitor baits captured significantly more EPN than A. fraterculus ones.This difference is probably due to the bigger soil volume in T. molitor baits.Besides, T. molitor moves inside the soil, which made the infection more likely.The proportion of EPN-positive samples obtained was considered low, when compared with other works using G. mellonella as bait: samples with nematodes reached 50% (Rio and Cameron, 2000); Campbell et al. (1995) found nematode-positive soil samples between 25 and 38%.Besides the possible difference in EPN density in the soils, for most EPNs G. mellonella is better than T. molitor because of its higher susceptibility and body size (Boff et al., 2000), what can explain at least partially the lower frequency observed here.
The presence of a fallen fruit over the soil is expected to increase arthropods abundance nearby, because of fruit fly larvae and pupae came from the fruit (Salles, 2000) and other insect species that feeds on the decomposing fruit are stimulated.These arthropods are possible hosts for EPNs.So, it was expected an increase in EPN capture when soil samples is taken underneath a fallen apple, but it was not observed here (Figure 1B), despite fruit infestation by A. fraterculus (Figure 1C).This trial didn't regard the total host range in the apple orchard, so the nematodes could have other suitable host or food source, even because A. fraterculus is not present during the entire year.Furthermore, a great number of arthropods live in orchards and are possible nematode hosts during all seasons, regardless fallen fruit (Glazer et al., 1996).
EPN-positive samples were not different across year seasons.Temperatures between 16-10 °C as observed in Porto Amazonas during winter usually reduces EPN efficacy (Klein, 1990;Mejia-Torres and Sáenz, 2013).However, winter temperatures seemed to be high enough to assure nematode capture during the entire year, even because samples were kept at 25 ºC in the laboratory, what favored the infection.The homogeneous occurrence of EPNs along the year corroborates Campbell et al. (1995) findings on Heterorhabditis sp.constancy during all seasons, and Puza and Mracek (2005) on Steinernema affine (Gerdin and Beddind).The capture of EPNs over different seasons, and recoveries in the same area during and after winter, indicates its population is well established.
Soil moisture is one of the most critical factors on nematodes movement and survival (Kaya, 1990), as observed in a Citrus sp.grove where higher humidity close to the roots was responsible by higher number of EPN-positive soil samples (Glazer et al., 1996).However, in the present work there was no correlation between rainfall and the proportion of samples with EPN captured, probably because sample soil humidity was corrected in laboratory (Voss et al., 2009), so this influence could have been diminished.It's prudent to have in mind that between-work differences could be related to the species considered.However, results for Oscheius sp. are still few or inexistent.
Different textures of soil didn't affect the capture of nematodes.It was expected that sandy soil could have more EPNs captured, especially because of the pore sizes that enhance nematode efficiency (Kaya, 1990).Hoy et al. (2008) observed that Rhabditis sp.bacteria feeders were found in most vegetable cultivated soil, beside the differences in sand content from 29.7 to 49.1%.In the same way, organic matter was not an important factor in EPN capture.Organic matter richer soils have more microorganism activity, which means more food sources for facultative EPNs like Oscheius.Bacterial activity in the soil correlated positively with number of the facultative-EPN Rhabditis sp.collected (Hoy et al., 2008).The soil collected in this trial was rich in organic matter (3.8 -5%) what can be enough to ensure good microorganism activity.
Five isolates from this trial were found to be close related to two accessions of Oscheius sp.(MCB and TEL-2014).Besides other accessions were found to be more identical to isolate 222, e-values for TEL-2014 and MCB pairing were lower (0.0), because the query cover was higher.The smaller the e-value, the higher the chance of matches not to be random, but due to relativeness.The accession MCB was already supported as entomopathogenic (Serepa and Gray, 2014).The isolate Oscheius sp.TEL-2014 was collected in a loam sandy soil on G. mellonella baits in South Africa (NCBI, 2015).
In South America, some Oscheius species [Oscheius colombiana and Rhabditis (Oscheius) pheropsophi n. sp.] were already found associated with insect in Colombia and southeast Brazil, respectively (Smart and Nguyen, 1994;Stock et al., 2005).However, it is possible that the captured isolates are new species since there is no report of this genus in south Brazil.The identification of these new facultative insect-parasitic nematodes is important since they can be natural pest enemies, and also can help on equilibrating soil insect populations.Further studies are necessary on the species-level identification, morphometric characteristics, associated bacteria identification and the relationship between them, involving biotic and abiotic factors.
It was observed differences in virulence to A. fraterculus between 304 and 297 as well as 319 and 297 (Figure 3B).Assuming the isolates are the same species (if not, it would explain the differences), the difference in virulence can be explained by the kind and amount of bacteria each isolate carries.In Oscheius sp., the association between nematode and bacteria is symbiotic, not mutualistic as in Heterorhabiditis and Steinernema genera.Although the associated bacterium was not studied in this work, it is known that Oscheius-associated bacterium doesn't live inside its body, but on its surface (Torres- Barragan et al., 2011).Four different bacteria species were associated with Rhabditis (Oscheius) blumi Sabhaus.Only two were pathogenic to G. mellonella and with different virulence (Park et al., 2011).So, it is likely to be similar in the isolates captured here.Still, Oscheius sp.isolates which were less virulent to A. fraterculus can have other suitable food source at field since they have facultative-parasitic behavior.
It was expected that EPNs captured in the orchard would be more virulent to the same orchard pest A. fraterculus than S. carpocapsae CB 02, which was cllected in Florida (USA).However, it didn't happen.Similar observation was already reported in other insects (Grewal et al., 2002).Rodrigues-Trentini (1996) also found high virulence of S. carpocapsae to A. fraterculus larvae.Steinernema carpocapsae is an obligate parasite, with a wide range of hosts, and needs to be highly effective to survive, while Oscheius sp., as explained above, can feed in sources other than live insects.
As a conclusion, nematodes of the genera Oscheius are found in an apple orchard of Porto Amazonas.Some of them are pathogenic to Anastrepha fraterculus.Facultative EPNs population is constant along the year in the orchard, independently of fluctuations in A. fraterculus population, air temperature, pluvial precipitation and presence of fruit over the soil surface.

Figure 1 .
Figure 1.Relative frequency of EPNs -positive soil samples when the bait was Tenebrio molitor or Anastrepha fraterculus larvae (A), and when samples were collected without or with an apple over the soil (B).In chart (C) the number of Anastrepha sp.pupae obtained from apples collected in the canopies or orchard ground.Time is about one 'Eva' apple production cycle in Porto Amazonas, PR.

Figure 2 .
Figure 2. Mean air temperature measured in Lapa-PR (A) and rainfall cumulated in Porto Amazonas-PR (B), 2, 4 and 6 days before soil sampling (sampling day included) from Jul 11, 2013 to Aug 6, 2014, and percentage of samples with entomopathogenic nematode (EPN) captured (right axis) in soil from an apple orchard in Porto Amazonas.Weather data provided by Simepar.

Figure 3 .
Figure 3. Mortality of Anastrepha fraterculus larvae following inoculation of different nematode strains isolated from soil in Porto Amazonas, PR, and Steinernema carpocapsae CB 02, in three trials performed in Petri dishes (100 IJ.larva -1 ) (A, B and C).Mortality was transformed by square root for statistical analysis, resulting in coefficient of variation 9.58, 8.75 and 6.26 for A, B and C, respectively; columns with the same letter above are not different (Tukey test, α=0.05).