Biocontrol of Sirex noctilio by the parasitic nematode Deladenus siricidicola: A five season field study in southern Chile

In 2001, the woodwasp Sirex noctilio was detected in Pinus radiata plantations in the Biobio region of southern Chile. Subsequently, an intense biological control program using the female sterilizing nematode Deladenus siricidicola was implemented in 2010. During five seasons between 2012 and 2017, we studied the parasitism of D. siricidicola nematode and its effect on woodwasp populations and infestation of P. radiata in different locations within the Biobio region. Parasitism was assessed by dissecting adult females of S. noctilio obtained from infested P. radiata logs. The total population of S. noctilio was determined by the emergence of individuals from the same logs. The level of damage caused by the S. noctilio pest was determined by establishing plots in stands of P. radiata at an intensity of 1 plot every 5 ha-1. During the study period, parasitism of S. noctilio by the nematode D. siricidicola increased from 29.6% in 2012 to 93.1% in 2016, while pest population decreased 3.4% in the same time period. Infestation increased from 0.3 to 11,6% of trees between 2012 and 2015, but subsequently decreased to 5.9% by 2017. We confirmed establishment of the nematode in the region under study and its natural dispersion to non-inoculated areas. Finally, we determined that the effect of inoculation age (antiquity) on parasitism levels reached 90% after three years of inoculation.


Introduction
The wood wasp S. noctilio Fabricius (Hymenoptera: Siricidae) drills the wood of standing trees belonging to the Pinus genus and occasionally Pseudotsuga, Abies, Larix and Picea species [1], with the most susceptible species being Pinus radiata D. Don [2]. The Sirex noctilio attack begins when females oviposit in the stems of live trees and inject a phytotoxic mucus and spores of the symbiotic fungi Amylostereum areolatum [1] along with the egg and later the fungus is eaten by the wasp's larva as food. The combined action of the mucus and basidiomycete a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Inoculation of Deladenus siricidicola in Pinus radiata
This study was performed between 2012 and 2017 in S. noctilio infested P. radiata plantations on the property of Forestal Mininco S.A., located in different sites within the Biobio Region. The extent of S. noctilio infestation varied annually, with 3,100 ha, 5,800 ha, 11,900 ha, 19,500 ha, 20,000 ha, and 20,000 ha recorded for the years 2012, 2013, 2014, 2015, 2016 and 2017, respectively (unpublished data). Between the months of April and September for the years 2010-2016, D. siricidicola was inoculated in 2,500, 4,000, 11,455, 12,500, 15,000, 15,000 and 10,000 trees naturally infested by S. noctilio, respectively. Each tree was inoculated with 250,000 to 300,000 nematodes, as reported previously [3,8]. The inoculation procedure was performed as described by Bedding [8]. Trees were selected for inoculation based on diameter (greater than 15 cm) and the absence of emergence orifices. The presence of S. noctilio larvae was confirmed through dissection of the stem segment. Trees were debranched and drilled at intervals of 30 cm in two parallel lines along the axe. Cylindrical perforations of 1 cm in diameter were made at a depth of 2 cm. D. siricidicola was applied in each perforation as a gel suspension. This process was carried out between the months of April to September of each year, without rain and at environmental temperatures ranging between 7˚C and 24˚C. The temperature or moisture inside the inoculated trees was not measured.

Selection of sampling sites to evaluate parasitism of S. noctilio by D. siricidicola
With the aim of evaluating the efficacy of the biological control program in the macroarea of the Biobio region in southern Chile, sampling were selected across the zone of S. noctilio colonization, advancing from the south to the north.  (Table 1) and in stand older than ten years regardless of D. siridicola inoculation status. Each SU consisted of 3 wooden logs 1 m in length extracted from 3 randomly selected trees with evidence of S. noctilio infestation. Infestation symptoms include the presence of eggs of S. noctilio on the tree stem, chlorosis, and fading or death of the foliage [2,5]. Asymptomatic trees were not sampled since they were not counted in the infected population. Therefore, the infestation data represented the population of the pest in infected trees, expressed as individuals attacked/ m 3 . We established 78, 85, 92, 104, and 93 SU for the five years between 2012 and 2016, respectively (Table 1, Fig 1). In some localities, the number of sampling points decreased between years due to reduction of the pest, forest fires, and tree harvests.
Selected trees from each SU were manually felled in October of each year. A meter-long log was extracted from the medial section of the stem. Prior to extraction, the presence of S. noctilio was determined by splintering the upper and lower ends of the shaft. When the insect was not detected in the stem shaft, the tree was exchanged for another infected individual.

Evaluation of parasitism of S. noctilio by D. siricidicola
Pinus radiata logs were labeled and the ends were sealed with solid paraffin until adults hatched according to the methods described by Goycoolea et al. [25]. Logs from the same SU were arranged vertically in a breeding chamber comprised of a cardboard drum (1.1 m of height x 0.7 m width) with a metal mesh cover, avoiding contact between the logs and the walls of the drum. The breeding chambers were kept in a shed with semi-shade mesh (50% coverage) and a polyethylene roof for up to 30 days after the emergence of the last adult S. noctilio specimen. Breeding chambers were observed weekly and adult S. noctilio emergence was recorded. S. noctilio specimens were preserved in 70% alcohol along with to the other individuals that emerged from the same chamber Deladenus siricidicola parasitism was determined in the lab by dissecting adult S. noctilio females using methods described by Zondang [15]. Briefly, the abdomen was cut and placed in a clock glass. Distilled water was added and the abdomen was dissected longitudinally under a dissecting scope (10x). The presence of nematodes was observed directly in the tissue, hemocele, and reproductive organs. Percentage of S. noctilio parasitism by D. siricidicola por SU was calculated as a ratio between N˚of parasitized in relation to N˚of dissected.

Effect of inoculation antiquity and geography on parasitism of S. noctilio
To assess the effect of time passed since initial inoculation on parasitism levels, we surveyed parasitism in SU in sites with different inoculation dates. We included SUs from distinct regions (coastal mountain, central valley, and Andes mountain) in consideration of possible geographic differences including Nacimiento from the Nahuebulta mountain chains, Los

Evaluation of the effect of D. siricidicola parasitism on S. noctilio population and P. radiata infestation
We selected the localities Nacimiento, Los Á ngeles, Quilleco, and Mulchén, which were inoculated with D. siricidicola in 2011, to allow for the longest possible temporal analysis of parasitism evaluation, S. noctilio population, and P. radiata damage assessment. Sirex noctilio population was determined according to the emergence of adult specimens in the growth chambers and expressed as the number of S. noctilio specimens per cubic meter of infected trees. The volume was obtained by applying the formula for a cylinder.
Sirex noctilio infestation of P. radiata was evaluated between 2012 and 2017. The infestation evaluation unit was comprised of lineal parcels of 30 trees. Evaluation units were established every five hectares. Sirex noctilio infestation was confirmed according to the signs and symptoms described by Newmann et al. (1987) and Aguilar & Lanfranco (1988) [2,5]. Results were expressed as the percentage of infested trees versus healthy trees (i.e. infestation intensity). The relationship between infestation intensity by S. noctilio and D. siricidicola parasitism was analyzed using the Duncan multiple comparison test as described above.

Sirex noctilio emergence
Total adult S. noctilio emergence (male and female adult specimens) between 2012 and 2016 was 3,374, 2,259, 4,033, 3,100 and 1,597 specimens, respectively. The specimens/m 3 Table 2). The number of locations with D. siricidicola parasitism increased over time, varying from 57% during the first season to 100% in the third season. D. siricidicola parasitism increased with time in all study locations, exceeding 85% from 2014 onward ( Table 2). During 2015, we observed D. siricidicola parasitism of S. noctilio in the non-inoculated locations of El Carmen and Hualqui (Table 2). In addition, we observed D. siricidicola parasitism in several SUs in seven noninoculated sites in 2012 and nine non-inoculated sites in 2013.  Table 3 presents data collected from SUs of the 2011 inoculation area. Between the years 2014 and 2016 parasitism levels where higher, more homogenous, and independent of time since inoculation (Table 3). Parasitism levels in relation to time since inoculation increased progressively from the years 2012 to 2014. For one, two and three years after inoculation, parasitism levels increased from 21.5 to 90.9%, 60.7 to 96.9%, and 91.5 to 90.6%, respectively. From 2015 onward, parasitism levels were higher and more homogeneous than previous years with values around 90% (Table 3).

Geographical distribution on D. siricidicola parasitism
Deladenus siricidicola parasitism of S. noctilio was determined in four localities representing different geographical zones. No parasitism was observed in Nacimiento from the Nahuebulta mountain chain during the years 2012 and 2013. Los Angeles from the South Central Valley also showed no parasitism during 2012. From the year 2014 onward we observed parasitism in all localities with increased levels of parasitism in comparison to earlier years reaching values ranging between 90.8 and 100% (Table 2). Increases in parasitism levels were generally of greater magnitude between 1 to 2 and 2 to 3 years after inoculation compared to later time points (Table 4). The largest increases were reported in the South Central Valley (59%) during the third year and in the Nahuelbuta mountain chain (90.7%) during the second year (p�0.05) ( Table 4).

Effect of D. siricidicola parasitism on S. noctilio populations and P. radiata infestation
The average population of S. noctilio for the four localities during five seasons between 2012 and 2016 was 408.7, 476.5, 881.5, 544.4, and 394.5 specimens/m 3 ( Table 5). The S. noctilio population increased from 2012 to 2014. By 2016, the population fell below initial levels, representing an overall 3.4% decrease in the pest density (Table 5). During the third year of assessment (2014), the highest emergence of S. noctilio was observed during the entire study period. During the same season we observed an increase in D. siricidicola parasitism levels which then stabilized at 90% in all four localities (Table 2). Between 2015 and 2016, the S. noctilio population decreased (Fig 3).
Total S. noctilio infestation of P. radiata ranged from 0.3 to 5.9% between the first and last year of assessment (Table 6). For all four localities, S. noctilio infestation peaked during 2015, coinciding with the season in which the woodwasp population levels started to decrease ( Table 5, Fig 3). In each subsequent year, infestation levels decreased (Table 6). S. noctilio Table 2. Deladenus siricidicola parasitism of female S. noctilio by locality during five seasons.  infestation levels started to decrease during 2016, two years after D. siricidicola parasitism levels stabilized in 90% and one season after of the initial S. noctilio population decrease ( Table 6).

Discussion
Evaluating D. siricidicola parasitism of S. noctilio revealed a continuous increase in the frequency of D. siricidicola detection and parasitism, reaching 100 and 93.6%, respectively, from 2014 onward. D. siricidicola nematode samples where obtained from naturally infested trees suggesting establishment of the nematode in the area as well as the probable natural dispersion through flight periods of the females as suggested by Taylor [20]. The mean parasitism levels reported during the last three seasons of evaluation (93.6%, 96.5%, and 93.1%) are closer to those reported in South Africa (96% in Cape Locality) and Brazil (90% in Encruzilhada Do Sul) than to the 100% parasitism levels reported for Victoria, Australia two years after inoculation [8]. The parasitism levels observed in this study were also higher than those reported in the US by Williams and Hajek [19], which ranged from 20.5-28.1% and 13.6-17.6%. The wide fluctuation in parasitism levels between locations during the first two seasons of assessment (0 to 89%) suggests that D. siricidicola was in the middle of the colonization process. This situation changed in 2014 when parasitism levels were higher (up to 100%) and more homogeneous. The apparent absence of D. siricidicola during 2012 and 2013 could be The different letters indicate statistically significant differences between the mean from each location within each annual increase (p�0.05) https://doi.org/10.1371/journal.pone.0207529.t004 Table 5. Sirex noctilio adult emergence obtained for the SUs from Los Á ngeles, Mulchén, Nacimiento, and Quilleco for five seasons. due to several factors. Consistent with the density dependence question raised by Bedding [8], the low levels of inoculation after one year coupled with low Sirex populations could have rendered D. siricidicola parasitism undetectable. Deladenus siricidicola parasitism was determined under a dissecting scope, a sensitive method to detect the presence of a nematode. Assuming that parasitism existed in the field, the number of samples may have been an insufficient representation of a population with low nematode presence. Moreover, D. siricidicola inoculum may vary in their parasitic ability as reported by Yu et al [7] in Canada, Williams and Hajek [19] in the US, and Bedding [8] in Australia. However, the viability of inoculum as well as the inoculation technique was similar to those reported in successful cases. The interaction between nematodes and the symbiotic fungus A. areolatum within S. noctilio was not analyzed in the present study. It's has been reported that nematode reproduction in D. siricidicola can be negatively affected by a unique mechanism of parasitism of adults and eggs by the fungus A. areolatum [26]. This discovery provides a possible explanation why Deladenus does not survive in culture when Amylostereum is fast growing or when the ratio of nematodes to fungus is inordinately biased toward the fungus [26]. This fungus has been shown to interfere with biological control of wood wasp, but in this work, the presence of A. areolatum was not observed. The insect Ips grandicollis has also been reported to effect biocontrol of wood wasps [27] but to our knowledge, this insect has not been reported in Chile.

Sirex noctilio/m
Overall, the estimated population of S. noctilio decreased by 3.4% from 2012 to 2016. The population increased and peaked in the third season, then progressively decreased from the fourth season until the fifth, reaching population numbers lower than initial levels in 2012 (Table 5). This is consistent with the results of Williams and Hajek [18] in the US, where the densities of S. noctilio were 1,000 Sirex/m 3 and 300 Sirex/m 3 in 2007 and 2012, respectively.
Infestation increased from 0.3 to 11.6% of trees between 2012 and 2015 followed by decreases in subsequent years reaching as low as 5.9%. The decline in infestation occurred one year after the pest population decreased. The inflection points in the curves of population density and infestation occurred after one season and two seasons, respectively, following the stabilization of D. siricidicola parasitism levels of S. noctilio at 90% in each of the four localities.
The flight period was detected between November and April in this study consistent with the range of flight period from October to May reported by Ruiz [28] in the southern hemisphere and with the October to April period indicated by Iede et al. [29] in Brazil. The 2:1 male to female emergence ratio reported during the first two seasons of the study indicated that the wood wasps were in the process of colonization [28] and reached stability during the last season analyzed.
In this study no effect of the zone on the level of parasitism was determined. The similarity in levels of parasitism observed between Nahuelbuta, the Central Valley (north and south), and the Andean Precordillera as of 2014 was unexpected. Deladenus siricidicola parasitism was to increase more slowly in the central valley as the product of a larger pest population; however, parasitism in the Central Valley (Los Angeles) was comparable to that of the other zones. This is due to the high parasitic capacity of D. siricidicola at high densities of the pest [8,19,30] Notably, climatic variations between zones were not significant, with average temperatures for the period under study of 14.1˚C, 15˚C, and 14˚C for the Precordillera, Central Valley, and Nahuelbuta, respectively.
Catastrophic levels of damage caused by S. noctilio were expected in the study area, particularly in the Biobio region. However, economic losses reported to date have not significantly compromised the forest industry in Chile (personal communication). These results are promising in terms of controlling S. noctilio populations, especially given the observed penetrance of the nematode D. siricidicola. Nonetheless, long-term monitoring of S. noctilio must be considered to detect and prevent potential outbreaks. During the last year of the study, the general level of parasitism decreased from 96.5 (± 0.4) in 2015 to 93.1 (± 0.9) in 2016. This observation is critical to understanding the dynamics of the nematode and wood wasp relationship and the long-term biocontrol of the pest. Achieving high levels of parasitism may come at the cost of reducing nematode presence in the environment due to reduction of the wood wasp host and its role as a dispersion medium for the nematode. Given that the success of nematodes as a biocontrol measure is density-dependent [8], drastically reducing the wood wasp population would likely limit the range and propagation of the nematode, which could affect the sustainability and long-term success of the biocontrol program.

Conclusions
The nematode D. siricidicola, parasite of the woodwasp S. noctilio, has been successfully established in the study area of the Biobio region in southern Chile, dispersing naturally into pine plantations where the nematode was not actively inoculated. Parasitism levels of S. noctilio by the nematode D. siricidicola increased progressively from inoculation until the third year following inoculation to eventually stabilize around 90% of parasitism. Three years after inoculation, D. siricidicola was established in similar levels in all three regions including Pre-Andean sites, the Central Valley, and the Nahuelbuta mountain range in the Biobio region. Our work demonstrates the reduction of S. noctilio populations and associated decrease in P. radiata infestation after parasitism by the nematode D. siricidicola reached 90% in all localities and serves as a useful example of the scale and timeframe over which biological control of S. noctilio is possible.