EXTERNAL DEVELOPMENT OF THE ENTOMOPATHOGENIC FUNGI Beauveria bassiana AND Metarhizium anisopliae IN THE SUBTERRANEAN TERMITE Heterotermes tenuis

The subterranean termite Heterotermes tenuis is one of the main pests of sugarcane and eucalyptus in Brazil, and the use of entomopathogenic fungi, alone or associated to chemicals, is an efficient and environmentally favorable method for its control. Studies related to the fungal development on these insects are important due to the effect of insect behavior on entomopathogens. The objective of this work was to describe the external development of Beauveria bassiana and Metarhizium anisopliae on H. tenuis using Scanning Electron Microscopy (SEM), determining the duration of the different phases of fungal infection. Two fixation techniques for preparing SEM samples were also evaluated. Worker specimens of H. tenuis were inoculated with a 1 x 10 conidia mL suspension of the fungi and maintained at 25±1C and 70±10% relative humidity. Insects were collected from 0 to 144 hours after inoculation and prepared on SEM stubs for each of the two fixation techniques. The results obtained with the two techniques were compared and duration of the different phases of the infection process were estimated from SEM observations and compared for three fungal isolates. B. bassiana and M. anisopliae have similar development cycles on the termite, but some important differences exist. The penetration, colonization and conidiogenesis phases are relatively faster for M. anisopliae than for B. bassiana, which results in a faster rate of insect mortality. The fixation technique with OsO4 vapor is suitable for preparation of insects to be used in SEM observation of the developmental stages of entomopathogenic fungi.


INTRODUCTION
The subterranean termite Heterotermes tenuis (Hagen) is one of the main pests of sugarcane and eucalyptus in Brazil, and its importance has increased considerably after the use of organo-clorinated insecticides was banned in the middle of the 1980's.The use of entomopathogenic fungi, alone or associated to new chemical active ingredients, has proved to be an efficient and environmentally favorable method for the Scientia Agricola, v.59, n.2, p.267-273, abr./jun.2002 control of the termite.Beauveria bassiana (Bals.)Vuill.and Metarhizium anisopliae (Metsch.)Sorok.are two of the most well studied entomopathogenic fungi (Fernandes & Alves, 1991, 1992;Almeida & Alves, 1995, 1996;Almeida et al. 1998;Moino Jr. & Alves, 1998).
Studies related to the duration of the different phases of fungal development on insects are relatively rare.This is especially important due to the effect of insect behavior on entomopathogens (Bao & Yendol, 1971;Hanel, 1982).These studies can be conducted using bioassays and observation of the different phases of fungal infection, through scanning electron microscopy (SEM) (Neves et al., 1996).The method used in the preparation of SEM samples should avoid damage to the insect and fungal structures involved in penetration, especially when the objective is to document the infection process (Quattlebaum & Carner, 1980;Hunt et al., 1984).
The objective of this work was to describe the external development cycle of B. bassiana and M. anisopliae on H. tenuis using SEM, and to determine the duration of the different phases of fungal infection.Two fixation techniques for preparing SEM samples were evaluated.

MATERIAL AND METHODS
Worker specimens of H. tenuis, collected on cardboard traps (Termitrap ® ) in an area of sugarcane production (Piracicaba, SP -Brazil), were transferred to plastic petri plates (6 x 1.5 cm), lined with filter paper.Termites were inoculated with 0.1 ml of a suspension (1 x 10 9 conidia mL -1 ) of B. bassiana (isolates 447 or 634) or M. anisopliae (isolate E-9).The fungal strains, obtained from the bank of entomopathogens at the Insect Pathology and Microbial Control Laboratory (Entomology Sector -USP/ESALQ), were applied with a Paasche airbrush sprayer operated with a pressure of 1,47 MPa.Ten plates with 20 insects/plate were used for each one of the fungal strains, maintained at 25 ± 1 o C and 70 ± 10% relative humidity.
Ten insects were removed from the plates 0, 6, 12, 24, 48, 72, 96, 120 and 144 hours after inoculation.The insects were killed in Petri dishes with ethyl ether and mounted on SEM stubs.Two stubs with five insects each were prepared for each sampling time for each of the two fixation techniques used: Osmium Tetroxide (OsO 4 ) fixation, and Glutaraldehyde/OsO 4 /Sodiumcacodilate buffer fixation.
For the first fixation technique, the specimens were exposed to OsO 4 vapor for 48 hours in a hermetically closed plastic container with a glass Petri dish containing a liquid film of OsO 4 .After fixation, the insect stubs were dried for 72 hours in a glass desiccator with silica gel (relative humidity at 0%).The last stage of preparation was the gold-palladium coating in a Balzers Evaporator, MED 010, for 120 seconds.
For fixation in Glutaraldehyde/OsO 4 /Sodiumcacodilate buffer, the insects were fixed by immersion for 6 hours in 4% glutaraldehyde with a 0.2 M pH 7.2 sodium-cacodilate buffer.Specimens were then fixed in 1% OsO 4 in a 0.1 M pH 7.2 sodium-cacodilate buffer for 1 hour.Fixation and dehydration were performed in plastic Ependorff tubes.
After fixation, specimens were dehydrated in 30, 50, 70, 90, and 100% acetone series.The insects were finaly washed three times in a 100% acetone solution.The critical point during drying when using CO 2 was followed in a Balzers, CPD 030.The insects were then mounted on stubs and coated with goldpalladium as previously described.As in the previous technique, two stubs with five insects each were prepared for each sampling time.Only isolates 634 (B.bassiana) and E-9 (M.anisopliae) were used for this technique.
The insects were observed with a Zeiss, DMS 940-A scanning electron microscope, and the results obtained with the two preparation techniques were compared.Durations of the different phases of the infection process were estimated from SEM observations and compared for the three fungal isolates.

RESULTS AND DISCUSSION
Fixing insects in glutaraldehyde and OsO 4 was more efficient in preserving the characteristics of both insects and fungi when compared to the fixation technique in OsO 4 vapor only.The fungal structures were well preserved in the specimens fixed with glutaraldehyde technique (Figure 1 A, B, C), whereas with the osmium vapor technique, the fungal conidia became wrinkled and contracted (Figure 1 D, E, F).The insect body parts with stronger cuticle, such as the head, were less susceptible to alterations caused by the fixation technique.Thus, when the objective of the study is the quantification or visualization of pathogen structures in certain areas of the insect body, as in the present study, the osmium vapor technique can be used, because of its lower cost and degree of difficulty than the glutaraldehyde technique.Also, the OsO 4 technique does not cause loss of the features to be studied.On the other hand, when more detailed characteristics are to be observed, as in taxonomic studies, the glutaraldehyde technique is preferable, because the aesthetic results are better.
In the present study, observation of germination and penetration of the conidia was performed on specimens fixed by the osmium vapor technique, which allowed the visualization of a greater number of structures than the glutaraldehyde technique.Despite the aesthetic advantages, the glutaraldehyde technique results in some loss of structures caused by immersion of the specimens in different liquids (acetone, glutaraldehyde, buffer, and osmium tetroxide).Little variation in timing and duration of the phases was observed in the infection process for the 2 B. bassiana isolates.After application of B. bassiana, conidia adhered to the insect integument within 0 to 6 hours after inoculation (Figure 2 A, B).Initiation of conidial germination happened between 12 and 48 hours after inoculation (Figure 2 C).The majority of the observed conidia were on the insect legs and head, but some conidia were also observed on the thoracic or abdominal segments (Figure 2 D).According to Boucias et al. (1996), B. bassiana formed germ-tubes and penetration structures between 12 and 24 hours after inoculation in Reticulitermes flavipes.
The formation of haloes around fungal conidia and germination tubes on the insect cuticle was observed with conidia adhesion and germination, but mainly during the penetration process (Figure 2 C).The occurrence of these haloes seems to be related to production and excretion of exoenzymes by the entomopathogen during the infective process.The enzymatic action of the fungus B. bassiana on Heliothis zea and Curculio caryae has been observed before with SEM (Smith et al., 1981;Champlin et al., 1981).The combination of proteolitic enzymes and chitinase produced by the fungal mycelium digest the insect cuticle, facilitating the penetration of the insect integument.
A thickening of the extremity of the germ-tube, characterizing the formation of appressoria, was observed during penetration by the fungus (Figure 2 D).
In some areas such as the insect head, extensive growth of germ-tubes was observed, probably due to resistance to fungal penetration in these areas with more heavily sclerotized cuticle.
The phase of host colonization occurred between 72 and 120 hours, and most of the insects died between 72 and 96 hours after inoculation.The insects killed by B. bassiana had a pinkish coloration characteristic of oosporein activity, which is common in insects infected by this fungus (Vining et al., 1962).
Mycelial extrusion from the cadavers happened between 96 and 120 hours after inoculation, mainly in the

I H G F E D C B A
head, and also in the membranous area of the labrum (Neves & Alves, 2000).However, little conidiogenesis occurred on the abdomen.Unlike H. tenuis, and due to feeding habits and behavior of C. cumulans, this insect harbors a large number of microorganisms in its digestive system.These microorganisms may compete with the fungal pathogen, preventing optimal growth of the entomopathogen.Bipiramidal crystals, possibly of double oxalate of magnesium and ammonium, were observed encrusted on the integument of an infected insect, close to the area of the penetration of the germ-tubes (Figure 2 E).These crystals have been observed before in Bombyx mori (Amaral & Alves, 1979) and their role in the pathogenicity of certain fungi is well documented in the literature (Roberts & Krasnoff, 1998).These authors observed oxalate crystals on the surface of other insects killed by B. bassiana.Oxalic acid may be an important toxin in the haemolymph of insects infected by Beauveria species.The presence of these crystals is also common in the general cavity of mites (Tetranichidae and Eryophiidae) infected by Beauveria and Metarhizium strains (M.A. Tamai & S.B. Alves, personal communication).
Observation of the development of M. anisopliae on H. tenuis revealed many similarities with the events reported for B. bassiana.Conidial adhesion to the integument happened immediately after inoculation and between 0 and 6 hours (Figure 3 A), whereas conidial germination occurred between 12 and 24 hours after inoculation (Figure 3 B).Penetration happened between 24 and 48 hours after inoculation.The average time for the penetration of M. anisopliae in N. exitiosus was 48 hours (Hanel, 1982).
Similar to B. bassiana, appressoria were also formed in M. anisopliae, characterized by a thickening of the extremity of the germ-tubes (Figure 3 B, C).This thickening is probably due to the translocation of the conidial cytoplasmatic content to facilitate the enzymatic synthesis necessary for the penetration phase.The formation of haloes on the insect cuticle at the germination and penetration points was also observed for M. anisopliae.An enzyme of the chymoelastase group (Pr1) has been implicated in cuticle degradation (St. Leger et al., 1988).According to these authors, M. anisopliae uses up to 78% of the protein synthesis for the production of this important enzyme in the process of penetration.
The colonization phase of M. anisopliae occurred between 48 and 72 hours after inoculation.M. anisopliae killed the insects 48 to 72 hours after inoculation, and the mycelium extrusion happened between 48 and 120 hours after inoculation.The mycelial extrusion was more intense for M. anisopliae than for B. bassiana.It was present initially, in the intersegmental areas (Figure 3 D,  E), and later in other areas (Figure 3 F), resulting in a process of cuticle degradation along the whole body of the insect (Figure 3 G).Conidiogenesis happened between 96 and 144 hours after inoculation and, as happened for B. bassiana, the whole body was covered by M. anisopliae conidia 120 hours after inoculation (Figure 3 H, I).
The timing for the different phases of the cycle of B. bassiana and M. anisopliae on H. tenuis are superimposed without a well defined separation (Figure 4).The M. anisopliae isolate E-9 killed the insect faster than the B. bassiana isolates 447 and 634, due to faster penetration and colonization.Almeida & Alves (1995), selecting strains of entomopathogenic fungi for the control of H. tenuis, concentrated their efforts on the genus Beauveria, because the strains of M. anisopliae tested had low pathogenicity.Later, the M. anisopliae isolate 1037, from Solenopsis sp.fire ant, provided promising results in the control of H. tenuis and C. cumulans (Alves et al., 1997).The ongoing search for more pathogenic and virulent isolates of entomopathogenic fungi can profit from SEM studies, like the present study, that identify isolates with faster penetration and colonization.
Despite the standardization in fungal application, the inocula that reaches the insects are variable.This variation may interfere with the extent of colonization of the insect, the speed of mycelium growth, and conidiogenesis, resulting in differences in the mortality of the insects.The rates of conidia germination and penetration of the insect integument have been shown to be related to the virulence of the fungi and the susceptibility of the insect host (Pekrul & Grula, 1979;Fargues, 1984).This study allowed the observation of the different phases of the disease cycle, and further demonstrated the importance of understanding these phases in selecting isolates for biological control of insects.
Figure 2 -Scanning electron microscope micrographs of the development of B. bassiana on H. tenuis.A) Conidia adhered to the tegument (1000X, 0 h after inoculation); B) Conidium adhered to the basis of a seta (2000X, 24 h); C) Germinating conidium, with halo formation in the cuticle (4000X, 72 h); D) Penetration of a germ-tube into the cuticle (3000X, 72 h); E) Details of the bipiramidal crystals, possibly of oxalate (2000X, 96 h); F) Extrusion of the mycelium at intersegmental region (500X, 96 h); G) Conidiophore formation and conidiogenesis on the termite head, with degradation of the cuticle (800X, 120 h); H and I) Details of the conidiogenesis process (4000X and 2000X, 120 h, respectively).

Figure 4 -
Figure 4 -Duration, in hours, of the different developmental phases of B. bassiana and M. anisopliae on workers of the subterranean termite H. tenuis.