Sample collection and baculovirus detection

In this work, several populations of sawflies associated with Pinus and Juniperus trees that inhabit temperate forests were sampled. The samples involved populations located between 1 500 and 3 000 m in all mountain systems of Mexico (Figure 1). The location of the sawfly populations (pest outbreaks) was supported by the network of local technicians of CONAFOR (National Forestry Commission), and a number of them were detected in several forest regions. Subsequently, 23 samples of larvae suspected of carrying baculoviral infection were collected between 2011 and 2019. Most of the samples were conform by dead and dehydrated larvae, which were placed in containers either individually or in small groups of individuals with the same appearance. When the collected individuals were still hydrated, they were placed in vials with silica gel, or their confinements were placed on ice for transport to the laboratory. All samples were frozen until microscopic diagnosis. Two of the samples were collected as live larvae but they died during transport to the laboratory.

The samples were examined for the presence of baculovirus polyhedra in the cadavers. For diagnosis, a frotis from the abdominal region of each cadaver was prepared on a slide and observed under an Olympus CX43 microscope at 400X or 1000X with phase contrast or darkfield illumination; in addition to the appearance of each cadaver, the type of microorganism observed, and its relative abundance was also recorded. To identify the sawfly species involved in each field sample, adults obtained in the field or laboratory were processed with appropriate taxonomic keys (^{De Lira et al., 2021}); however, for four samples (1, 2, 9, and 10, Table 1) no adults were collected or brooded from larvae. For other five samples 3, 7, 14, 15, and 23 (Table 1), only the genus was determined due to the lack of appropriate taxonomic keys. The adults derived from these samples did not consistently match the available taxonomic keys, suggesting that they may be new species.

Figure 1 Map showing the locations where the 23 samples of sawfly larvae were collected to search for baculoviral infections. The box on the right shows the list of samples analyzed and the name of the collection sites. 

Molecular detection of baculovirus DNA in larvae

A nucleic acid hybridization test was carried out on DNA from larval cadavers of Zadiprion ojedae Smith collected in Guachochi, Chihuahua, and Zadiprion borjai (^{González et al., 2022}) from Miquihuana, Tamaulipas.

Sample collection

In the field, several sawfly outbreaks were detected in 10 Mexican regions located into 8 Mexican states (Aguascalientes, Chihuahua, Durango, Guerrero, Jalisco, Michoacán, San Luis Potosí and Tamaulipas). From those outbreaks, 23 samples of larvae suspected of being infected by baculovirus were collected (Table 1). These samples included eight species of sawflies (Zadiprio falsus Smith, Z. ojedae, Z. borjai, Zadiprion spp., M. sanchezii Smith, M. cuauhtemoci, Neodiprion omosus Smith, N. autumnalis Smith, and Neodiprion spp.). Other species may be present in the four samples where neither genus nor species was determined because no adults were obtained from the corresponding population, or in the other five samples whose adults were not in accordance with the descriptions of the taxonomic keys, probably because they are new species.

Molecular detection

The DIG-NeabNPVPolh probe hybridized with its homologous sequence from the positive control (NeabNPVPolh sequence) placed in wells 3A and 3B in Figure 2 but did not produce significant signal in the two negative controls: tick DNA (1A and 1B) and DNA from healthy Zadiprion spp. cell culture (wells 2A and 2B). The test showed significant reaction consistent with DNA from three Z. ojedae subsamples from Guachochi, Chihuahua (wells 4A and 4B; 8A and 8B; and 12A and 12B in Figure 2). The remaining samples did not show consistent responses between replications; therefore, they were considered negative.

Figure 2 Hybridization assay of DNA from Diprionidae larval cadavers on a nitrocellulose membrane. The sequence of the NeabNPV Polh gene (741 bp), replicated in the PucIDT plasmid and labeled with digoxigenin, was used as a probe. 100 ng of DNA was used in each well. Each sample was formed from 10 cadavers of Zadiprion ojedae and Z. borjai larvae collected in the field and suspected of carrying baculoviral infection. Hybridization was detected by enzyme-linked immunosorbent reaction using a chromogenic substrate that produces a dark staining. 

Detection of baculoviral infection by light microscopy

Small shiny bodies were found in frotises of larval abdominal tissue observed under the microscope (Figure 3); their brightness, shape and abundance made them compatible with baculoviral polyhedra and different from other shiny objects also present in the frotises. These baculoviral polyhedra were detected in cadavers from 11 field samples (4, 7, 8, 14, 15, 16, 17, 18, 19, 20 and 22) taken from five Mexican states: Chihuahua, Durango, Guerrero, San Luis Potosí and Tamaulipas (Table 1). The cadavers with polyhedra also presented other microorganisms such as bacteria (cocci or bacilli) (Figure 3), microsporidia, or fungi. Frequently, two or more of these groups of microorganisms were found in each cadaver concomitant with polyhedra.

Figure 3 Macrophotographic section taken from a frotis of the abdominal region of a cadaver of Monoctenus cuauhtemoci. from Ixcateopan de Cuauhtémoc, Guerrero (sample 16 in Table 1). Phase contrast illumination and 400X magnification were used for this image. The circles surround small shiny bodies interpreted as baculiviral polyhedra. Arrows point to bacillary bacterial cells with a shiny body at each end, suggesting that they may belong to some variant of Bacillus thuringienesis. Rectangles surround cocci-like bacteria. 

Biological activity of cadaveric inocula

The cadaveric inocula explored induced symptoms and signs of disease compatible with baculoviral or bacterial infections: test larvae stopped feeding, gradually lost motility and died with generalized darkening of the body or localized in their abdominal area. In contrast, about 85 % of the larvae in the control groups reached the pupal stage or remained healthy until the last treated larva died. Microscopic examination of most of the cadavers resulting from larvae treated with inocula 4 and 14 showed abundant polyhedra (in 17 of 19 and 22 of 22, respectively). According to Table 3, inocula 7 and 8 induced sparse polyhedra in few cadavers (in 1 of 20 and in 6 of 20, respectively). Inoculum 16 induced a moderate frequency of polyhedra in the resulting cadavers (33 out of 145). All macerates induced, in addition to polyhedra, proliferation of other microorganisms distinguishable under light microscopy at 400X or 1000X and phase contrast or darkfield illumination. Such microorganisms were bacilli (small, medium or large), cocci, microsporidia or fungal hyphae. All of these constituted the accompanying microbiota of the polyhedra in the field samples and in the cadavers resulting from this assay, part of which is shown in Figure 3. The accompanying microbiota in most of the larvae treated with inocula 4 and 14 were bacilli. Thirteen of the 20 larvae treated with inoculum 7 showed proliferation of fungal hyphae similar to those observed in the inoculum donor cadaver and bacteria (Table 3). Three other cadavers from larvae treated with inoculum 7 or 8 showed cocci-like bacteria (Table 3). Most of the larvae treated with inoculum 16 showed small to medium-sized bacilli or fungal hyphae.

Detection of baculoviral infections

The microscopic diagnosis of larval cadaveric frotises at 400X and 1000X indicated the presence of small shiny bodies in 11 of the 23 field samples, which were interpreted as baculoviral polyhedra based on their brightness, shape, and relative abundance (Table 1). The samples harboring baculoviral infections comes from 6 of the 11-collection region explored, into five Mexican states. Eight sawfly species were represented in the samples harboring baculoviral polyhedra: Z. falsus, Z. ojedae, Zadiprion borjai, Zadiprion spp., M. sanchezii, M. cuauhtemoci, N. omosus, and N. autumnalis (Table 1), which are distributed in three genera.

The interpretation of those shiny bodies as baculoviral polyhedra is supported by two pieces of evidence provided in this study. 1) Successful induction of disease in healthy larvae inoculated with any of five macerates of cadavers harboring polyhedra and the subsequent proliferation of polyhedra in the cadavers resulting from such inoculations (Table 3). 2) Hybridization of three DNA subsamples from the Z. ojedae population located in Guachochi, Chihuahua (locus 4, 8, and 12 in Figure 2) with the DIG-NeabNPVPolh probe, indicating the presence of baculoviral DNA in the samples. Other larvae from Guachochis’s Z. ojedae population showed polyhedra on microscopic examination (sample 18, Table 1). In agreement with this evidence, the small shiny bodies observed in cadavers of the other 10 samples (Table 1) not used in this hybridization test were also interpreted as baculoviral polyhedra: 4 and 8 (Zadiprion falsus); 7, 14, 15 (Zadiprion spp.); 19 (Z. borjai); 16 and 17 (M. cuauhtemoci); 20 (M. sanchezii); and 22 (N. autumnalis).

The DIG-NeabNPVPolh probe is considered highly specific for detecting the sawfly nucleopolyhedrovirus polyhedrin gene, as its sequence, the same as that of the NeabNPV Polh gene, only differs by 5 % from the homologous sequence of any of the other known nucleopolyhedroviruses in Diprionidae; for example, those of N. lecontei (NeleNPV) and N. sertifer (NeseNPV) (^{Lauzon et al., 2006}). That maximum expected divergence between the probe used here and its target DNA does not represent an impediment to hybridization, as suggested by the successful use of other probes in detecting their respective DNA targets with divergences >20 % (^{Meier-Kolthoff et al., 2014}).

Cadaver appearance (rictus mortem)

In the literature, the only description found of baculovirus-induced rictus mortem in these insects indicates that larvae hang from their hind legs (^{Dixon, 2019}). Based on that description and this working group's experience with baculoviral infections in Lepidoptera, we searched for sawfly larval cadavers with that appearance in field. They were expected to be dark, and flattened with liquefied internal tissues, as described for lepidopteran larvae killed by baculoviruses (^{Federici, 1997}). Although larvae with liquefied internal tissues were observed in three of the first samples with polyhedra, subsequent samples showed that this appearance is the least frequent in baculovirus-killed diprionids. Two samples from the same population of Zadiprion spp. (14 and 15; Table 1 and Figure 4) clearly show that the most common consistency of Diprionidae cadavers with polyhedra differs from that commonly observed in Lepidoptera (hanging from their hind legs, liquefied, and flattened).

The first of these samples, number 14 (Table 1; Figures 4A and 4B), consisted of cadavers found on needles of low branches of pine trees. They were either hanging by their hind legs or fell from the highest branches; their bodies were flattened, black in color and had a leathery or hard consistency. Only one of them showed abundant polyhedra, while the other 40 showed few or scarce polyhedra. About 97 % of the cadavers in this sample presented bacilli as the dominant microorganism, so their death was attributed to bacillus and not to virus, despite the presence of the latter in the cadavers.

Figure 4 Appearance of sawfly (Zadiprion spp.) larval cadavers in the field (La Unión de San Isidro Montes de Oca, Sierra de Ixtapa, Guerrero), which carried baculoviral polyhedra. A and B are typical in sample 14; C, D, and E are typical in sample 15 in Table 1. The formers are fully or partially obscured. The latters are only partially obscured, depending on the relative abundance of the accompanying microbiota in the cadavers. 

Cadavers in sample 15 (Table 1; Figures 4C, 4D and 4E) were found on the ground at the foot of pine trees. It is assumed that they climbed down the trunk before dying, possibly to pupate. The cadavers were found dry and hard (not liquefied), retaining their cylindrical shape and, most of them, even their natural color (yellowish), while others showed black areas on their body. Because these characteristics suggest a fungal death, the first 145 cadavers examined from that sample were placed in a humid chamber or small pieces of them were directly deposited on PDA culture media. Subsequently, for the next 42 cadavers, frotises were prepared from their abdominal region; under light microscopy it was observed that polyhedra dominated over concurrent microbiota in more than 90 % of them (38 of 42), and only four cadavers had cocci and bacilli as dominant microbiota.

When comparing the appearance of Zadiprion spp. cadavers in samples 14 and 15 and their accompanying microbiota, it was clear that the flat body and dark color observed in sample 14 were related to the bacilli and not to the virus. Similarly, the hard, cylindrical body of the cadavers in sample 15 was associated with the virus and not with the accompanying microbiota. Another piece of evidence suggesting that baculovirus-killed sawfly larvae remain hard or leathery, but not liquefied, comes from ^{Lucarotti et al. (2007}), who describe how they preserve and process large quantities of baculovirus-killed N. abietis larvae for the formulation of a bioinsecticide.

The liquefied appearance of the cadavers was observed in one individual from sample 4, in 10 from sample 7 and in 10 from sample 8. The liquefied cadaver from sample 4 showed few cocci and many bacteria and polyhedra; while those from samples 7 and 8 showed few polyhedra, but a lot of accompanying microbiota (Table 1). Subsequently, when the macerate of a cadaver from sample 4 was inoculated into healthy larvae, it induced a high frequency of cadavers with polyhedra (17 out of 19) and a large amount of polyhedra in each cadaver; while a macerate from sample 7 induced only one cadaver with very few polyhedra, but abundant accompanying microbiota in all treated larvae (Table 3). Therefore, in cadavers carrying polyhedra, characteristics such as liquefaction, flattening and darkening of the body, and hanging from the hind legs, appear to be associated with the presence, type, and relative abundance of the accompanying microbiota, but not with baculovirus.

When the baculovirus was the predominant death agent in the sawfly cadavers, most cadavers conserved its cylindrical shape, their bodies were hard and did not hang from their hind legs. This rictus mortem is different from that induced by the baculovirus in lepidopteran species: which get liquefied, flatted, torn, and commonly, hang from their hind legs (^{Federici, 1997}). The differences most likely arise from the fact that, in diprionids, only mesenteron cells are attacked (^{Harrison et al., 2018}), leaving the rest of the tissues intact (unless degraded by other microorganisms); while in Lepidoptera, almost all tissues are colonized by the virus, which became liquefied by (^{Federici, 1997}). This description of the rictus mortem of sawfly larvae killed by baculoviruses may be useful in future searches for these viruses for controlling forest pests.

Bioinsecticidal potential of baculoviruses

The five inocula tested showed differential potential as bioinsecticides. Although all of them produced death in all treated larvae, the identification of the agent causing death provides a first appreciation of the relative potential of each inoculum as a bioinsecticide. Inocula from samples 4 and 14 caused numerous polyhedra to appear on each cadaver and a high proportion of cadavers with polyhedra (17 of 19 and 22 of 22, respectively; Table 3). Samples 7 and 8 caused a small number of polyhedra in each resulting cadaver and only a few cadavers with polyhedra (1 of 20 and 6 of 20, respectively; Table 3).

Inoculum from sample 16 caused polyhedra proliferation in a moderate frequency of cadavers (33 of 145; Table 3). It was assumed that dead larvae that did not show polyhedra were killed by the microbiota that accompanied the virus in the donor cadavers. The ability of, at least one part of the microbiota accompanying the baculovirus to cause death is discussed below. No species of that microbiota was identified, but it was evident that the fungal hyphae observed on the cadavers in this trial do not belong to Metarhizum or Beauveria genera.

The abundance of polyhedra in the inoculum donor cadavers of samples 4 and 14, on the one hand, and 7 and 8, on the other, maintained a clear correspondence with the abundance of the polyhedra in the cadavers resulting from this test and with the frequency of cadavers bearing polyhedra (Table 3). This correspondence suggests that inocula from samples 4 and 14 have a high potential to be developed as bioinsecticides due to their high infectivity.

The inoculum of sample 14 provided also evidence of having high potential as a natural biocontrol factor in the field where it was collected (La Unión de San Isidro Montes de Oca, Sierra de Ixtapa, Guerrero). Since sample 15 (Table 1) (not included in this test) was taken from the same group of pine trees as sample 14, it is reasonable to assume that both samples harbor the same virus strain and that pathogens found in them performed together at the scene of good control of their host observed in the field. It was, the sawfly population was virtually eliminated by pathogens prior to sampling, and that happened very early in the development of this pest outbreak as suggested both by the appearance of the pupal cocoons found scratching in the ground, and by the evidently reduced extension of the infestation. Only two generations of the pest would have occurred in that place by the sampling time. So early elimination of the outbreak clearly suggests that this inoculum is highly infective on its host.

Both the owner of the field where the sample was collected and the CONAFOR technician (the official in charge of supervising forest health in that area) assured that there was no human intervention in the development of this epizootic. Although several pathogens were detected acting simultaneously (virus, bacteria, and at least one fungus), the virus turned out to be the most frequent in the cadavers collected there.

Regarding the inocula of samples 7, 8, and 16, in the first perception, this test suggests that they have a relatively low potential as bioinsecticides, because they produced a low frequency of cadavers with polyhedra. However, after a detailed look at the result, the expectation of good potential of these inocula as bioinsecticides is maintained for the following reasons: 1) the low frequency of cadavers with polyhedra may be due to the reduced number of polyhedra in donor cadavers; 2) larvae that received inocula 7 or 8 and, most of those receiving inoculum 16, showed abundant accompanying microbiota, but only a few showed some polyhedra suggesting pathogenic capacity in the microbiota; therefore, 3) inevitably, it competed with the virus and could displace it, either due to being more competitive or more abundant.

Based on this reasoning, it seems necessary to carry out further tests to adequately measure the potential of all the inocula obtained, that of the pathogenic microbiota accompanying the viruses, and the interactions between them. Such interactions seem very interesting for the management of these pests because they may involve synergies that should be exploited to achieve greater control of the pest, or antagonisms that should be avoided. In such new trials, it would be desirable to test the viral inocula and the accompanying microbiota, properly purified.

Another reasonable precaution would be to thoroughly search for and evaluate all existing virus strains of each sawfly species. Such evaluation of strains should be done in the stands where their host species is found, and the strain that provides the best control in each location could be used as a means of control in such areas. Failure to do so would run the risk of using, in some areas, a strain with less control potential than that of the resident strain and, due to the flooding application of a less potent exotic strain, the resident strain with greater potential would be displaced. Thus, virus strains with superior potential could be lost due to not having been discovered in a timely manner.