Cytological aspects of compatible and incompatible interactions between cashew (Anacardium occidentale L.) seedlings and isolates of Colletotrichum gloeosporioides (Penz.) complex

A strategy in the control anthracnose of cashew (Anacardium occidentale L.) is the management of crop phenology and defense mechanisms of this host. In previous studies, under controlled conditions, the seedling reactions of 5 cashew clones (CAP-14, CCP-06, CCP-09, CCP76 and CCP-1001) to 36 isolates of Colletorichum gloeosporioides Penz. complex (LARS905 to 940) was evaluated. However, good field management requires information about the infection process. This research aimed to clarify cytophysiological aspects of three compatible interactions of this pathosystem (isolates LARS-905 and 910 × CCP-76; LARS-910 × CCP-1001) and an incompatible one (LARS-905 × CCP-1001), using infected leaves/stems and microscopy (light, scanning and transmission electron). No significant differences were found prior to penetration. In the susceptible combinations, 36-66 h after inoculation, a thin primary hypha (TPH) formed in the invaded epidermal cell, widening as a large primary hypha (LPH), which filled the cell lumen simultaneously with accumulation of of yellow-brown lignopolysaccharides. Then, a thin secondary hypha (TSH) developed from the LPH, penetrating adjacent cells before the first became necrotic. In the incompatible interaction, the response of the first invaded cell was faster and more intense, with formation of papilla and lignopolysaccharide-protein-silicon complex usually blocking the pathogen.


Introduction
The first Europeans who arrived in Brazil in the 16th century found a land with exotic plant-fruits, such as cashew trees (Anacardium occidentale L.), transporting them by sea to Africa (Mozambique, Angola, Kenya and Madagascar) and Goa (India). In India, cashew gradually gained commercial importance through the consumption of its nuts, becoming known as the "gold mine of wasteland", since it is a good crop for soil conservation, reforestation and wasteland development (SATAPATHY and BEURA, 2018). Later, in the 20 th century, the North American and European acceptance of cashew nuts, boosted its marketing from India to the world. Brazil's Northeastern semi-arid only entered this international chestnut market in the 1950s, but even more so in the 1970s, mainly in the dry season of the year, after government incentives for the development of cashew nut and juice processing (ALBUQUERQUE; SILVA, 2008).
Despite this, since 2012 the country's participation in the world cashew market is decreasing, not only because of improved technologies provided by incentives from other governments in the market, but due to the strong and prolonged drought in the Northeast region from 2012 to 2016, associated the occurrence of pests and diseases, such as anthracnose (VIDAL, 2016) caused by members of Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. species complex, i.e. C. chrysophilum, C. fragariae, C. fructicola, C. gloeosporioides sensu stricto, C. queenslandicum, C. siamense and C. tropicale (COMÉ, 2014;NORSHIE, 2020;VELOSO et al., 2018). This reduced the cashew plantation area by 144,000 hectares, leading producers to eradicate the dead cashew trees and to trade their wood (VIDAL, 2016). In fact, the losses caused by anthracnose to cashew trees were recognized in all producing countries (KHATOON et al., 2017;MONTEIRO et al., 2015), although there was no published research on the cytology, ultrastructure and physiology of the infection process on such a host.
Studies of the disease in compatible interactions of other pathosystems have shown that the type and sequence of early cytological infection events can differ markedly. In the systems bean-C. lindemuthianum, alfalfa-C. destructivum, pea-C. truncatum, ripe blueberry fruit-C. acutatum and Arabidopsis thaliana-C. higginsianum), the initial colonization of susceptible epidermal cells is via intracellular biotrophic hyphae BAILEY, 1988;O'CONNELL et al., 1993;LATUNDE-DADA et al., 1996;WHARTON;SCHILDER, 2008;O'CONNELL et al., 2012). For other species, such as from C. gloeosporioides complex, different isolates, although sharing similar morphological characteristics, have distinct modes of penetration and colonisation in different hosts (TEBEEST et al., 1978;MORIN et al., 1996;ZULFIQAR et al., 1996;WEI et al., 1997;BARRETO et al., 2007;PANDEY et al., 2012), however, once a compatible interaction has been established, the pathogen switches to a very destructive mode of growth, causing extensive cell wall dissolution and cell death.
According to López and Lucas (2010a), after assessing the reaction of one non-commercial  and four commercial dwarf cashew clones (CCP-06, CCP-09, CCP-76 and CCP-1001), in controlled environmental conditions, against 36 isolates of Colletotrichum, it was found that all the strains, including those from hosts other than cashew, were able to cause lesions on leaves and stems of most clones, albeit to different degrees. The isolates were obtained from anthracnose lesions of cashew and associated host plants in Brazil, and had been previously identified (by cultural, morphological and partial sequences of their 28S ribosomal DNA characters) as members of the C. gloeosporioides complex (López and Lucas, 2010b). While the clone CCP-06 was susceptible to most isolates, the clone CCP-1001 was resistant to many of them, including LARS-905, originally isolated from leaves of a blackberry plant (Rubus brasiliensis) in a cashew orchard, while the isolate LARS-910, originally obtained from cashew leaves, was aggressive to all clones. On the other hand, injury increased the susceptibility to all isolates, indicating that resistance may also be associated with structural barriers that hinder penetration.
Based on these results, this research aimed to investigate the infection process of two isolates of the C. gloeosporioides complex (LARS-905 and 910) on stems/leaves of seedlings of three cashew clones (CCP-76, CCP-1001 and CAP-14), providing information to improve understanding of the host-pathogen interactions, and aid disease management.

Materials and methods
The leaves and stems of cashew seedlings (60-days-old) of the clone CCP-1001 × LARS-905 (member of C. gloeosporioides complex originally isolated from anthracnose lesion of R. brasiliensis leaf, in Maceió-AL), an incompatible interaction, as well the ones of compatible interactions of the clones CCP-76, CCP-1001 and CAP-14 × LARS-910 (member of C. gloeosporioides complex originally isolated from anthracnose lesion of A. occidentale leaf, in Maceió-AL), or CCP-76 and CAP-14 × LARS-905, were investigated by microscopy. Pathogen behaviour on the host surface and the initial penetration process were studied by light and scanning microscopy (LM and SEM). Infection hyphae were visualized in stained semithin and ultra-thin sections of resin embedded samples, by LM and transmission electron microscopy (TEM). Rev. Bras. Frutic., Jaboticabal, 2021, v. 43, n. 2: (e-701) The ultrastructural relationship between the intracellular primary hyphae and host cell cytoplasm was followed to determine the effects on host cells, including the host responses.

Inoculation and whole tissue clearing and fixing
The detached leaves or stem segments were inoculated and incubated, as described by López and Lucas (2010a). At intervals after inoculation (12, 18, 24, 30, 38, 48, 72, 96, 114, 132, 158 and 182 h), tissues beneath inoculation droplets of twelve young stem segments (ca. 5 cm diameter) and three detached young first leaves of cashew seedlings were sliced with a double-edge razor blade. Five thin longitudinal strips of stem segments (ca. 0.5 mm thick) and five leaf discs (ca. 4 mm diameter), collected at each interval, were immediately placed in vials containing ca. 2ml of clarifying solution [0.15 % (w/v) Trichloroacetic acid (TCA) in a 3:1 mixture of Ethanol: Chloroform (EtOH: CHCl 3 )]. Then, these vials were immersed in a water bath (70 o C, ca. 1.5 h), until the tissues became clear. The assay was repeated three times, and cleared tissues were also used in histochemical tests.

Bright field and fluorescence -light microscopy (LM)
After a preliminary bright field observation of the mounted sections (Zeiss MC63 microscope), samples were viewed with a Zeiss Axiophot microscope equipped with Nomarski differential interference contrast (NDIC) optics and/or UV -epifluorescence. Photomicrographs were recorded [Kodak Technical Pan film (80 ASA) or Kodak TMAX-pro film (400 ASA), when UV-epifluorescence was required, and developed in DIAFINE (Acufine Inc., Chicago, U.S.A.), or Kodak Ektachrome daylight (100 ASA) film], according to the manufacturer's instructions, and then digitized.

Scanning electron microscopy (SEM)
Strips of leaf tissue (ca. 6 × 6 mm 2 ) beneath the inoculum drop were cut 48 h after inoculation, and attached to a copper specimen stub using a mixture of colloidal graphite and "Tissue Tec" glue (Agar Scientific, UK). Then, the specimens were frozen by placing the holder on the pre-chamber block (-130 o C) of a Hexland CT 1000 Cryo preparation unit (Oxford Instruments, UK). The pre-chamber was evacuated, and the specimens transferred to the cold microscope stage (SEM chamber) were examined using a Philips 505 SEM (4kV, -180 o C). When ice particles were present, the specimens were discarded, and the procedure was repeated. The tissues were then transfered to the pre-chamber, sputter coated with ca. 20 nm of gold, and returned to the microscope cold unit. The tissues were examined using an accelerating voltage of between 7.9kV and 8.1 kV, with the cold stage held at -130 o C to -180 o C. Photographs were taken on Kodak T-Max 100 film and digitized.

Staining of leaf section grids and transmission electron microscopy (TEM)
Ultra-thin sections of the LR-white embedded specimens (obtained as described in 2.2) were cut with a diamond knife using a Reichert-Jung ultramicrotome. They were mounted on Formvar-coated gold slot grids and double-stained with uranyl acetate (UA) followed by lead citrate solution, as described by Reynolds (1963). The grids were floated face down (3 min) on top of UA drops and placed on Parafilm. Just before the end of staining, drops of lead citrate were placed on another Parafilm strip, inside a small Petri dish containing pellets of NaOH to absorb CO 2 . The grids were then washed with a jet of fresh DW and immediately floated face down on top of lead citrate drops (ca. 2 min) and then dipped in 0.1 N NaOH solution followed by a jet of boiled SDW. Grids were finally blotted dry with filter paper, and examined in a Hitachi H7000 TEM operating at an accelerating voltage of 100 kV.

Analysis of the infection process
The cleared sections (TCA: EtOH: CHCl 3 ) to study the time course of infection by LM were either examined directly, after mounting in 15-30 ml of aqueous glycerol (30%) on slides under cover slips, or stained with of 0.025% (w/v) Aniline blue (Ab) in lactophenol (60 min at room temperature), washed in SDW and then mounted in aqueous glycerol (30%). Also, a KOH-Ab technique of Hood and Shew (1996) was used in some cases, to reveal fungal structures by fluorescence. For this, cleared samples were immersed in 1M KOH for 30 min, rinsed in SDW, mounted in 0.05% (w/v) Ab in 0.067 M K 2 HPO 4 (pH 9.0), and examined using excitation filter G365, dichroic mirror FT 396 and barrier filter LP 420.
The percentages of spore germination, appressoria formation and penetration (papilla, large primary and secondary hypha formed), for 100 conidia and appressoria at three different sites of five samples (time and interaction) were recorded. To clarify the mode of penetration and host cell responses, semi-thin sections of resin embedded material were mounted on glass slides at 60 o C and stained (1 min) with a solution of 0.5 % (w/v) Toluidine blue (Tb) in 0.1% (w/v) aqueous Na 2 CO 3 . These sections were examined by LM, with or without NDIC and photomicrographed (as described to SEM).

Assessment of membrane integrity
The viability of infected cells and the functional integrity of their membranes are affected by fungal colonisation. The of living plant cells plasmolysis reflects the semi-permeable nature of the plasmalemma and tonoplast membranes (OPARKA, 1994), since in a hypertonic salt solution, water is withdrawn from the vacuole by osmosis (the protoplast contracts). Then, strips of whole stem tissues, inoculated and collected as above described, were immersed for 15 min in a fresh solution of 0.85 M KNO3 in 5 mM Pb (pH 7.5). For the first 5 min, the vials with samples were placed under vacuum to aid infiltration (BAILEY et al., 1990). Treated tissues were mounted on slides, covered with cover slips and observed immediately by bright field and NDIC microscopy to record the percentage of plasmolysed cells. Photomicrographs were made as described to SEM.
Finally, samples (stem and leaf tissues) were stained for ca. 10 min (at 60 o C) with a mixture of (0.25% w/v) Coomassie blue-Red 250 (CB-R250) in a methanolic solution (MeOH: CH 3 COOH: DW 5:1:4 v/v) just prepared, for detecting proteins at the infection sites. Then, they were washed three times with the same methanolic solution (3 min at room temperature). Due to staining of chloroplasts, a blue background could not be avoided, but proteins stained violet-blue (WOLF and FRIC´, 1981) to purple-red if basic ones (WILSON, 1992).

Cycloheximide assays
Fifteen stem segments bordering the youngest leaves, and fifteen fully-expanded first leaves from cashew clone CCP-1001 (60-day-old) were excised, and the cut ends of the excised tissue sealed with warm, molten paraffin wax. Two groups of five stem segments and five leaves (abaxial surface) were placed inside plastic boxes (20 cm × 12 cm × 4 cm), on filter paper moistened either with an aqueous solution of 0.6 mM cycloheximide or with SDW. The samples were immediately inoculated with isolate 905 of C. gloeosporioides and incubated as described by López (1999) and López and Lucas (2010a). The third group of samples were treated with cycloheximide but not inoculated (control), to observe later the possible occurrence of phytotoxicity. After 72 h, the two groups of inoculated stem segments and leaves were collected for microscopic examination.
For assessment of penetration sites beneath appressoria, three sections of each sample were clarified, stained with Ab and mounted as described in the staining item. For each replicate, ca. 100 sites (appressoria formed) were examined by LM, being classified as follows: 1) no attempted penetration and cell wall apposition; 2) wall apposition formation with successful penetration; 3) wall apposition formation with successful penetration; 4) penetration without cell wall apposition. Penetration Rev. Bras. Frutic., Jaboticabal, 2021, v. 43, n. 2: (e-701) was considered successful if encased hyphae overcame papillae or the infection peg (IP) immediately enlarged into the cell lumen. Total sites in each category for the three replicates of the five samples were determined. The leaf samples were further incubated up to 168h after inoculation, and then examined to assess disease severity, as described by López and Lucas (2010). This experiment was repeated three times.

Infection and colonization process on young leaves and stems
There was no observed difference between susceptible (isolate LARS-910 × CCP-76, CAP-14 or CCP-1001; isolate 905 × CCP-76 or CAP-14) and resistant (LARS-905 × CCP-1001) interactions of C. gloeosporioides complex-cashew leaf/stem tissues during the pre-penetration phase. After penetration, the main differences were the timing, number of invaded cells and extent of the host cell response to infection (Tables 1 and 2).
-a A minimum of 100 conidia, appressoria or reaction sites were recorded in each of 3 replicates of 5 samples per interaction per time. b Percentage from the germinated conidia. c Percentage from the appressoria formed and/or attempts of penetration. Papillae (mainly dome-like shape) were observed during cuticular penetration and thin primary hypha (TPH) formation, in all the interactions. d Percentage from attempts of penetration (with or without papillae). LPH formed in the absence of papilla were more common in compatible interactions. e Percentage from the LPH formed; f When multiple penetration attempts occurred in the same cell, its rapid collapse restricted the TPH advance; g Penetration restricted to the first cell containing LPH; h Because of the rapid and intense browning of the invaded cells, SH were not recorded in whole tissues, but in semi-thin resin embedded sections (LM). Gl. Slide = glass slide; conidiat. = conidiation. ---a A minimum of 100 conidia, appressoria or reaction sites were recorded in each of 3 replicates of 5 samples per interaction per time. b Percentage from the germinated conidia. c Percentage from the appressoria formed and/or attempts of penetration. Papillae (mainly dome-like shape) were observed during the cuticular penetration and thin primary hypha (TPH) formation, in all the interactions. d Percentage from attempts of penetration (with or without papillae). LPH formed in the absence of papilla was more common in compatible interactions. e Percentage from the LPH formed; f When multiple penetration attempts occurred in the same cell, its rapid collapse restricted the TPH advance; g Penetration restricted to the first cell containing LPH; h Because of the rapid and intense browning of the invaded cells, SH were not recorded in whole tissues, but in semi-thin resin embedded sections (LM). Gl. Slide = glass slide; conidiat. = conidiation.
In pre-penetration stage, it was possible to see that conidia were longer than the length of stomatal pores of the small cashew leaf epidermal cells (Figure 1a). An extracellular matrix helped the appressoria to adhere firmly to the host surface. Direct penetration through the trichomes and cuticle of epidermal cells (Figure  1b) was the most common infection route, but the fungus also penetrated through stomatal pores ( Figure  2). Intramural invasion was rarely found, except after mechanical injury, when it was common. Mechanical injury invariably resulted in the activation of "quiescent" infections in all interactions and even in mature tissues to produce necrotic lesions.
Although infection pegs (IP) could not be very well visualized with the leaf clearing/stain technique (LM), they were evident in semi or ultra-thin sections. Penetration of trichome cells was always associated with a hypersensitive response, being also described in the interactions for northern jointvetch-C. gloeosporioides f.sp. aeschynomene, tea-G. theasinensis and alfalfa-C.
In both interactions of the cashew-C. gloeosporioides complex system, epidermal cells beneath melanized appressoria initially responded to the IP (24-48 h) by changing the existing wall chemistry with a continuous deposition of yellow-orange (LM) electron-dense (TEM) wall material, forming halos and papillae of different shapes and sizes (Figure 3). Where the IP extended (48-74 h) through the epidermal cell wall (ECW) and papilla, it formed a single (Figure  4) or sometimes forked (Figure 7a) thin primary hypha (TPH). In compatible interactions, the first cell containing an encased TPH always remained alive in the earlier stages, as indicated by its ability to plasmolyse normally (Figure 3).    In both interactions, when a papilla with a hypersensitive response (HR) was not formed quickly enough, it was possible to see the post-penetration events. However, in the resistant combination, only a small proportion of infected cells containing TPH did not immediately restrict them, since the majority are able to plamolyse too (Figure 5a). Usually, a HR was quickly induced in the first invaded and adjacent cells, restricting the development of the fungus (Figure 5b). On the other hand, in compatible combinations, the TPH immediately enlarged from the IP, forming the large primary hypha (LPH) that totally filled the lumen of the invaded cell (Figure 6a), which sometimes lost organization and stopped to plasmolyse, just before a thinner secondary hypha (TSH) had been formed and grew through thickened and electron-dense cell walls of adjacent cells (Figure 6b). In such necrotic cells, the granulation containing the yellow-orange pigment apparently burst and spread, simultaneously to the degeneration of organelles and membrane.
In the resistant interaction, however, the formed papillae could extend through the adjacent cells more quickly than the encased-TPH developed (Figure 7a), and they very rarely enlarged to a LPH, being possible to detect HR as soon as it is eventually formed (Figure  7b).   , stained with uranyl acetate/lead citrate. a) Note the removal of the appressorium during the resin embedding procedure; 66 h after the inoculation, an encased thin primary hypha (TPH) penetrated through the epidermal cell wall (ECW) and through the formed papilla (P), which extended to another adjacent cell, and the invaded cells could still plasmolyse (×7,000, Bar= 4.0 µm). b) 74 h after inoculation, a rapid hypersentive response (HR) was observed restricting the large primary hypha (LPH) emerging from a previously invaded cell. Note the electron-dense granulation inside the cell and the more electron dense walls of the hypersensitive and adjacent cells (× 5,000, Bar = 3.0 µm).
TSH was not observed in this interaction. In the later stages of the compatible infections (114-182 h), acervuli and conidiation could be observed, and superficial hyphae, or hyphae that emerged from collapsed cells, grew across the surface and directly penetrated subepidermal parenchyma tissue (Figures  8-9). Pandey et al. (2012a), studying onion leaves infected with isolate of the C. gloeosporioides complex, also showed that 6 h after the inoculation, a germ tube emerged from the conidium of this fungus and a globular shaped appressorium formed at the tip, maturing for 24-48 h; then, an infection hypha emerged through a pore at its base and penetrated directly the host cuticle, forming a primary hypha, to whom the onion leaf cell also responds by producing the papilla. In such leaves, the authors also could see, as in the present work, simultaneous penetration of the host through stomata. According to the authors, from 48-72 h, primary hyphae in onion cells started to branch out to form secondary hyphae within the epidermal cells followed by a massive growth of both intra-and intercellular hyphae leading to the development of small whitish and water-soaked lesions, and at 72-96 h after inoculation, intra-and intercellular hyphae radiated from cell to cell, resulting in the formation of acervuli. After that, typical onion anthracnose symptoms, with salmon coloured mucilaginous spore matrix were observed on the infected leaf surface, as was also seen in the infected leaves of susceptible cashew studied here. In addition to these studies, when the same group (PANDEY et al. 2012a) worked with leaves of Mangifera indica susceptible to one isolate of the C. gloeosporioides complex, it was found that hyphal invasion occurred intra and intercellularly until 72 h after the inoculation, with rapid collapse of the mesophyll cells and swelling of epidermal cell walls, and almost all the initially invaded cells became necrotic 96 h after inoculation, which corroborates with what was observed in this study.
Therefore, the formation of wall appositions and hypersensitive cell death are responses that could contribute to the epidermal resistance of cashew against the invasion of the studied isolates of the C. gloeosporioides complex, besides the individual clonal phenology. An important fact to note is that the isolates used in this study were previously identified through their morphophysiological and genetical (rDNA large subunit) aspects as a member of C. gloeosporioides complex, as well as 34 other isolates with aggressiveness against cashew trees (López and Lucas, 2010 b ), even if obtained from lesions in other associated hosts (crosspathogenicity). However, the diversity of survival strategies and life cycles of Colletotrichum species are characteristics responsible for their high taxonomic complexity, and for this reason the reclassification of species of this genus was established (CANNON et al. (2012). By 2018, 218 species had already been recognized in the same Genera, and the majority were subclassified in 12 species complexes and 2 others were considered individual species (JAYAWARDENA et al., 2016;GUARNACCIA et al., 2017;VELOSO et al. 2018), so that, currently, C.gloeosporioides is considered a species complex for different hosts. In the case of cashew trees, Comé (2014) and then Veloso et al. (2018) evidenced, by multilocus genetic analysis, six and then seven species, respectively, of this complex, capable of infecting this host, and it is also probable that both isolates used in the present study, although belonging to the same complex, may currently correspond to phylogenetically different species, requiring confirmation by the multifocus analysis. On the other hand, for the purpose of this work, Ana Maria Queijeiro López & John Alexander Lucas in which an interaction incompatible with an isolate originally from another host plant and three compatible ones, with an isolate obtained originally from cashew anthracnose lesions, this information does not affect the histological results reported here. Table 3 shows the colour reaction of infection sites in stems and leaves of infected cashew seedlings stained as described in the methodology.

Histochemical studies on papillae in cashew leaves and stems
Under bright field illumination, wall appositions and the lumen of hypersensitive infected cells were yellow-orange (Figure 3), and after aniline blue (Ab) treatment and examination under UV-light, the inner surface of the cell walls beneath appressoria and papillae, in both interactions, gave positive results for callose. However, callose itself was not the only component of the papillae, since granular deposits in the recentlyformed outer layer did not fluoresce with Ab and stained purple (bright field) with Tb-O treatments, a result negative for lignin-like compounds, but positive for pectin, cellulose, hemicellulose (O'BRIEN et al., 1964;HERR, 1992), proteinaceous material (CLOUD and DEVERALL, 1987) or opaline silica (HEATH, 1979). The other histochemical tests for lignin evidenced that phenolic material of different types was detected (Table  3). Autofluorescence, a characteristic of lignin-like compounds, was not observed at any of the penetration sites or in the lumen of hypersensitive cells. Table 3. Histochemical colour reactions of papillae, halos and hypersensitive cells of cashew leaves or stems inoculated with isolates LARS-905 and 910 of the C. gloeosporioides complex.

Expected Colour Reaction
Aniline blue (AB) Callose = green-yellow fluorescence + AB-Toluidine blue (TB) Callose = green-yellow fluorescence + Different dyes increased the colour intensity of the infection sites to dark orange (bright field), and according to the phloroglucinol: HCl, phloroglucinol: CaCl 2 àHCl and chlorine-sodium sulphite tests, this could correspond to the presence of syringyl-lignin with syringyl-propane groups. In such cases, papillae and hypersensitive cells (48-120 h after inoculation) treated with KMnO 4 -HCl-NH 4 OH should stain purplered, but this was not seen. The diazotised-p-nitroaniline and p-nitrobenzene tetrafluoroborate-tests were positive for phenols (mono or polyhydroxy) with a para or ortho position free, but not necessarily lignin. At infection sites, the ferric chloride, vanillin and nitroso reactions were positive for flavonols, pectin bound to tannins or condensed tannins.
Whereas most reports in the literature assume that proteins stain blue-violet with Cb-R250, the wall appositions and hypersensitive cells of cashew leaves stained red-purple. Rosenthal et al. (1988) and Wilson (1992), however, found that histones, rubrophilin (a brain protein), collagen and other HRGPs and PRPs, which have a high affinity for tannins, dye red with Cb-R250. Woof et al. (1968) reported that proteins linked to di-or polyhydroxyphenols, independent of the pH of the interaction, produce different spectra with the individual compounds. Hagerman and Butler (1989) asserted that efforts must be made to remove all tannin from plant extracts before colorimetric tests with Cb, since the dye reacts to tannin as well as to protein. After mechanical or biotic wounding, as caused by insects or the fungus Botryodiplodia theobromae (syn. Lasiodiplodia theobromae), resin ducts of cashew trees, that are also present in their leaves (METCALFE and CHALK, 1950), secrete a gum formed by arabinogalactan proteins (AGPs), phenols and enzymes (polyphenoloxidase, peroxidase, chitinase and proteinase inhibitors) (MARQUES and XAVIER-FILHO, 1991;PAULA and RODRIGUES, 1995), and although an aqueous solution of this gum (33%) did not affect the mycelial growth in vitro of C. gloeosporioides, it inhibited several microorganisms (MARQUES et al., 1992).
Phenol oxidation of syringaldazine in the presence of H 2 O 2 produces the intensely purple tetramethoxyazo-pmethylenequinone. This histochemical test for peroxidase was positive in all the reaction sites and hypersensitive cells, 48-120 h after inoculation, with all the interactions. Walls of unchallenged cells or non-inoculated controls did not exhibit peroxidase activity. Although some authors have associated oxidation of syringaldazine with lignin formation (LEWIS and YAMAMOTO, 1990), the oxidase activity detected is not necessarily specific to lignin residues, since laccase from extracellular mucilage of C. graminicola complex conidia has also been detected with this substrate (ANDERSON and NICHOLSON, 1996). This glycoprotein mucilage, which has high affinity for binding condensed tannins, protects spores from inhibition of germination by polyphenols (NICHOLSON et al., 1986). While laccase can oxidise dihydroxyflavonols (ANDERSON and NICHOLSON, 1996), plant peroxidases are implicated in the crosslinking of cell wall components (polyphenols, HRGP and pectin), synthesis of o-quinones (HARRISON et al., 1995) and conversion of dihydroxyphenols to free radicals. The later may undergo reactions with cellular constituents and have toxic activity against microorganisms (BELL, 1981).
Oxidation and condensation of o-dihydroxyphenolic compounds (such as catechin, gallocatechin and their proanthocyanidin derivatives) is generally associated with resistance and characterised by the formation of brown o-quinones throughout the cell walls and collapsed protoplasts (BELL, 1981). In this study, necrotic tissues in susceptible interactions exude an intense brown pigment from the collapsed cells. When drops of NaOH (0.5-10 N) were placed on the sections, rather than dissolving the yellow-orange pigmentation at infection sites, the healthy tissue became blue (dihydroxyphenol derivatives). Oxidative bleaching (H 2 O 2 : CH 3 COOH) for 30 min also only caused a partial loss of the pigment intensity to pale yellow and none of the sections showed fluorescence. Increases in the time of treatment up to 90 min caused an overall tissue degeneration. WHEELER et al. (1978) found that treatment with H 2 O 2 : CH 3 COOH, for up to 4 h, did not completely remove dihydroxyphenol and catechol melanins from the matrix and outer cell wall of microsclerotia of Verticillium dahliae.
When the complex tannin-protein-quinone is formed, plant cell walls remain resistant to hydrolysis for extended periods of time (HARBORNE, 1980). Plant o-quinones (melanins) by themselves, showed extremely weak electron density in ultrastructural studies with French beans (HEATH, 1979). However, as the condensed tannins, which bind together starch or pectic compounds and form hydrogen bonds with structural proteins (MULLINS and LEE, 1991), quinone residues form covalent bonds with free amino groups of HRGPs. Tannins and o-quinones can be deposited in the cell walls, serving as scavengers for electron dense silicates, inactivating microbial extracellular enzymes (BELL, 1981).
As described previously, examination of leafsamples after conventional preparation for electron microscopy (involving heavy metal staining) revealed typical electron-dense deposits on the inner surface of the cells walls adjacent to appressoria. The appressorial walls, which contain polymeric melanins, appeared less electron dense than the ECWs of the infected leaves (Figure 1b). This may indicate that not only accumulation of the common organic constituents (proteins, carbohydrates and phenols) restricted the passage of electrons through the specimen after staining, but probably also deposits of some elements of high atomic weight. Silicon is a common element in most soils where cashew is grown, and it can be easily assimilated and stored as amorphous opaline silica (SiO 2 .nH 2 O). It is energetically cheap for defence, especially against herbivores, and high amounts of it have been detected in ash of cashew gum (9.3 mg/g ash) and of other Anacardiaceae gum (PINTO et al., 1995). The base of the soft trichomes of young leaves of some Anacardiaceae accumulates silicon (LUCAS; TEAFORD, 1995), besides tannins and alkyl-salicylic acid derivatives (anacardic acids). Table 4 show the effect of cycloheximide on the formation of cell wall appositions and tissue colonization. Detached leaves or stem segments from cashew clone CCP-1001 became susceptible to the pathogen when floated on cycloheximide (0.6 mM) solution immediately prior to inoculation, and the tissue collapsed and exuded brown pigment in advanced stages of infection. When stem segments where floated on SDW for 72 h, ca. 77% of the epidermal cells beneath appressoria showed papillae, ca. 10% showed penetrations without papillae, ca. 7% showed penetrations through the appositions and ca. 6% showed no wall apposition and no penetration ( Table 4). From the reaction sites of cycloheximide treated tissue, only 4 % of the penetration attempts showed appositions, while ca. 81% showed penetration without apposition, ca. 6% showed penetration through apposition and ca. 9% showed no wall apposition and no penetration. Therefore, cycloheximide may have inhibited the synthesis of proteins or enzymes involved in the formation of wall appositions in cashew tissues, and such appositions appear to be important for the active process of epidermal resistance.
Taken together, results of the different histochemical tests suggested that polysaccharides (callose, pectin, hemicellulose and cellulose), proteins (such as hydroxyproline rich glycoproteins, HRGPs), peroxidase and phenolics/tannins are produced and added to stored silicon in response to attempts of penetration in all the observed interactions. The formation of specific inhibitory metabolites at the site of wall apposition, might account for the unsuccessful penetration attempts, and probably varies in timing and intensity, permitting some attempts to succeed.

Conclusions
The cyto-histochemical studies described here, regarding the compatible interactions (LARS-910 × CCP-76, CCP-1001 and CAP-14; LARS-905 × CCP-76 and CAP-14 ) or resistant combination (LARS-905 × clone 1001) involving isolates of the C.gloeosporioides complex and leaves and stems of cashew seedlings, revealed that prior to penetration there are no significant differences between them. In compatible combinations, after the formation of melanized appressoria (24-48h after inoculation), penetration generally occurred from the infection peg (IP) through the cuticle and epidermal cell walls and trichomes. In the resistant combination, the same thing happened, although with a lower frequency and with a longer period of time to form the appressoria (48-74h). Less commonly, direct penetration of the fungus germ tubes through open stomata, without production of appressoria and without affecting the host's cuticle, was observed in both combinations. Very rarely, intramural invasion took place, except after mechanical injury, when it was common. Mechanical injury invariably resulted in the activation of "quiescent" infections in all interactions and even in mature tissues to produce necrotic lesions.
After penetration, no special infection structure was observed in both interactions, although in the compatible combination, the thin primary hypha (TPH) from the developed IP, enlarged in the first invaded epidermal cell, forming a broad primary hypha (LPH) 36-66 h after inoculation, even when eventually a papilla was later formed around the penetration point. In these early stages of colonization, the invaded cells were activated Rev. Bras. Frutic., Jaboticabal, 2021, v. 43, n. 2: (e-701) to trigger various defense mechanisms, albeit apparently slowly. Then, a yellow-brown lignopolysaccharideprotein-silicon complex filled the lumen of the invaded cells, and thin secondary hyphae (TSH) developed from the LPH, starting penetration of the healthy adjacent cells (66-96h) just before the colonized cell underwent necrosis (hemibiotrophy). In the most advanced stages of infection (114-182 h), the majority of colonized epidermal cells collapsed, and acervuli, setae and conidia were formed.
Histochemical tests suggested that, in both compatible or incompatible interactions, papillae are formed by polysaccharides (callose, pectin, hemicellulose and cellulose), PRPs or HPRGs, phenolics and silicates, with peroxidase activity detected in response to penetration in all of them. In the incompatible interaction (LARS-905 × CCP-1001), the treatment of the studied tissues with cycloheximide inhibited the synthesis of wall appositions and related resistance mechanisms. Therefore, the rapid and intense production of papillae and pigmented lignopolysaccharides, that may also signal adjacent cells to react to the fungus before it reaches them, are important defense mechanisms in the resistant interaction.
Such studies, associated with the phenology of the studied cashew clones in the edafoclimatic conditions of the regions where they are grown in the Brazilian Northeast, should enable the best integrated management of the damaging anthracnose disease.