Sugarcane glycoproteins may act as signals for the production of xanthan in the plant-associated bacterium Xanthomonas albilineans

Abstract

Visual symptoms of leaf scald necrosis in sugarcane (Saccharum officinarum) leaves develop in parallel to the accumulation of a fibrous material invading exocellular spaces and both xylem and phloem. These fibers are produced and secreted by the plant-associated bacterium Xanthomonas albilineans. Electron microscopy and specific staining methods for polysaccharides reveal the polysaccharidic nature of this material. These polysaccharides are not present in healthy leaves or in those from diseased plants without visual symptoms of leaf scald. Bacteria in several leaf tissues have been detected by immunogold labeling. The bacterial polysaccharide is not produced in axenic culture but it is actively synthesized when the microbes invade the host plant. This finding may be due to the production of plant glycoproteins, after bacteria infection which inhibit microbial proteases. In summary, our data are consistent with the existence of a positive feedback loop in which plant-produced glycoproteins act as a cell-to-bacteria signal that promotes xanthan production, by protecting some enzymes of xanthan biosynthesis against from bacterial proteolytic degradation.

Introduction

Leaf scald is a bacterial-vascular disease of sugarcane caused by Xanthomonas albilineans. The disease initially manifests itself as a 1–2 mm wide white streak (“pencil-line”) on the leaf, which usually follows the direction of the main veins. This is related to the production of the bacterial toxin albicidin, which inhibits chloroplast DNA replication and blocks chloroplast differentiation.^1,2 The streaks usually enlarge and the affected leaf becomes wilted and necrotic. Necrosis and wilt correlate with the production of a bacterial xanthan-like, exocellular polysaccharide. The white pencil line may also be visible on the leaf sheaths.³ Symptoms of this phase are often observed after sprouting, or in young shoots growing from infected plant cane. These symptoms sometimes disappear, but the plant remains infected. Alternatively, plants may be infected, but grow without showing any symptoms. Mature stalks may suddenly wilt and die, sometimes without the physical manifestation of other symptoms. The bacteria are transmitted through infected cuttings by tools used to cut stalks. There is also evidence of transmission through soil and water. Prevention usually consists of quarantine of varieties introduced from other growing areas. In areas where the disease is endemic, resistance is used to manage the disease. An additional prevention measure is a cold soak/hot water treatment before planting the cane.

X. albilineans may invade the parenchyma between the bundles but it is finally restricted to the xylem,¹ causing reddened pockets of gum similar to a xanthan-like polysaccharide.⁴ The polysaccharide produced by X. albilineans consists of a basal tetramer that is sequentially repeated. This basal tetrasaccharide is composed by two molecules of glucose, one mannose and one glucuronic acid.⁵ The best characterized xanthan is produced by the phytopathogenic, Gram-negative bacterium Xanthomonas campestris pv. campestris, which causes black rot in crucifers. This xanthan is a polymer made of repeats of a pentasaccharide assembled by sequential addition of glucose-1-phosphate, glucose, mannose, glucuronic acid and mannose on a polyprenol phosphate carrier.⁶ Xanthan synthesis occurs at the envelope of X. campestris. Proteins encoded by the gum genes are responsible of the complete synthesis of the xanthan. The gum gene cluster includes glycosyltransferase GumD, which transfers a glucose-phosphate residue from UDP-glucose (UDPglc) to a lipid carrier located at the inner face of the cell membrane; GumM, GumH, GumK and GumI transfer a second glucose unit, mannose, glucuronic acid and a second mannose residue, respectively; GumL, induces pyruvilation of the second mannose; GumF and GumG, which acetylate the mannose residues; GumJ, which promotes translocation of the polymerized units to the outer face of the inner membrane; GumE, which promotes xanthan polymerization in the periplasm and GumC, which forms a complex with the outer membrane protein GumB through the large periplasmic domain of the latter. These complexes form open pores to export the mature xanthan.⁷ Although the gum genes have not been cloned in X. albilineans, the nature of the xanthan-like polysaccharide containing D-glucuronic acid has been conclusively established.^4,5,8 GumK, GumL and GumG are not expressed in X. albilineans.

The two polysaccharides produced by both Xanthomonas species contain glucuronic acid; thus, the ability of Xanthomonas to produce an active UDP-glucose dehydrogenase (the enzyme that produces UDP-glucuronic acid from UDPG) is often identified as a virulence factor.⁹ The expression of the udgH gene, coding for UDPGDH in X. campestris, increases subsequent cell growth. This gene displays a second promoter that responds to stationary phase changes, retaining high levels of expression. Its mutation causes a loss of virulence in X. campestris and X. campestris pv. vesicatoria.¹⁰

Axenically cultured X. albilineans does not secrete xanthans in Willbrink liquid media, in contrast to Xanthomonas campestris,¹¹ which produces xanthan in vitro culture. This ability is based upon the existence of a rpf (regulation of pathogenicity factors) gene cluster that controls virulence, production of extracellular enzymes and synthesis of the xanthan.¹² The rpf gene cluster positively controls both these processes and boosts virulence. Genes within the rpf cluster also encode elements of a regulatory system involving DSF, a diffusible signal factor.¹³ DSF has recently been structurally characterized as cis-11-methyl-2-dodecenoic acid¹⁴ and behaves as a cell-to-cell signal required for biofilm formation, xanthan production and virulence.

Interestingly, X. albilineans depends on reception of a signal from the plant. Thus, characterizing the xanthan-like polysaccharide secreted by this strain requires the use of inoculated sugarcane tissues. We have previously found that several sugarcane glycoproteins, which are produced in response to mechanical injuries or after infection by several pathogens,^4,15 are able to inhibit bacterial protease activities that mainly hydrolyzed UDP-glucose dehydrogenase. This blocks production of UDPglucuronic acid.¹⁶ The polysaccharide moieties of these glycoproteins are composed of a β-1,2-fructosyl fructofuranoside linked to galactitol units through an ether bond.^17,18 It has been proposed that bacterial growth inside the host tissue promotes production of sugarcane glycoproteins; bacterial growth is also thought to increase UDP glucose dehydrogenase activity to assure exocellular polysaccharide production. In turn, expression of sugarcane glycoproteins promotes both production and secretion of the xanthan-like polysaccharide from X. albilineans by inhibiting bacterial proteases, forming a reciprocal feedback loop.

In this study, we test the hypothesis that visual symptoms of leaf scald are related to production of sugarcane glycoproteins, which function as a signal relaying system from sensitive plants to microbial cells to produce bacterial exocellular polysaccharides that will be secreted to the host tissues. We also propose the existence of a correlation between the extent of visual symptoms, the amount of xanthan-like polysaccharide produced and the location of the infecting bacterial populations in plant tissues.

Results and Discussion

Xanthomonas albilineans induces ultrastructural alterations in symptomatic leaves.

In healthy, control plants, leaves exhibited bundle sheath cells that displayed thick cell walls and a high density of chloroplasts with a well-defined lamellar system as well as a high number of starch granules (Fig. 1A). Stone cells surrounding the phloem bundle had an ample lumen, thick and multilayered cell walls and occasional remains of the cytoplasmic content (Fig. 1A and C). These could indicate that their differentiation had not finished. Conversely, docking cells appear as true schlereids (fibers) and displayed an empty cell lumen and enlarged cell walls (Fig. 1B).

Figure 1.

Transmission electron micrographs of cross sections of (A–C) leaves from healthy sugarcane plants, (D–F) leaves from diseased plants without leaf scald symptoms and (G–I) leaves from diseased plants with symptoms where le, lower epidermis; bsc, bundle sheath cells; mc, mesophyll cells; bv, bundle vessels; sc, stone cells; dc, docking cells; ph, phloem; xy, xylem.

The ultrastructure of leaves from diseased plants without visible symptoms of leaf scald displayed marked differences compared to healthy samples. However, the xylem, phloem and the stone cells did not show noticeable changes. However, the chloroplasts of bundle sheath cells (Fig. 1D and F) did not contain starch and the docking cells displayed thickened cell walls causing the almost complete disappearance of the lumen (Fig. 1E).

The absence of starch in chloroplasts of bundle sheath cells was also observed in leaves of diseased plants with visible symptoms of leaf scald (Fig. 1G and H). This symptom could be related the action of albicidin, which blocks chloroplast differentiation.^1,2 The cell walls of the docking cells were thinner that those observed in leaves from asymptomatic plants, and their lumen was wider (Fig. 1H). This may be related to the fragility of the lignified cells walls of large xylem vessels, which appeared somewhat fragmented and broken (Fig. 1I).

SEM ultrastructural analysis of both healthy and asymptomatic diseased leaves did not reveal significant alterations of the leaf structure. Epidermal were arranged normally (Fig. 2A and D), and mesophyll cells did not show any visible deformation (Fig. 2B and E). Bundle sheath cells and the vascular bundles were also intact (Fig. 2C and F). However, ultrastructural SEM examination of leaves with symptoms of leaf scald disease display marked differences. For example, the extracellular spaces of mesophyll cells were densely covered with long fibers from the subepidermal layer to the neighbouring of the vascular bundle (Fig. 2G). These fibers penetrated both xylem and phloem vessels (Fig. 2H); in the latter, bacterial cells could be observed together with a dense network of fibrous material (Fig. 2H). These fibers were apparently associated, forming large filamentous structures near the xylem vessels (Fig. 2I).

Figure 2.

Scanning electron micrographs of cross sections of (A–C) leaves from healthy sugarcane plants, (D–F) leaves from diseased plants without leaf scald symptoms and (G–I) leaves from diseased plants with symptoms. ue, upper epidermis; le, lower epidermis; bsc, bundle sheath cells; mc, mesophyll cells; vb, vascular bundle; ph, phloem; xy, xylem.

Location of pathogenic bacteria in leaf tissues.

Previous reports have indicated that bacteria were mainly confined to the vascular bundle in stalks and leaves;³⁰ conversely, another report suggested that X. albilineans was constrained to the xylem.¹ To address the exact localization of the pathogen within symptomatic leaves, semi-thin sections from necrotic leaves (symptomatic, 14 months-old plants) were treated for IGL with rabbit polyclonal antibodies against X. albilineans. Intense, specific staining was observed in epidermal cuticle, epidermal cell walls, guard cells, subepidermal fibers and pericyclic fibers. However, staining was less intense in chloroplasts and walls of bundle sheath cells, in mesophyll cells layer close to the bundle sheath cells rings and also inside the sieve tubes and companion cells (Fig. 3B and D). Thus, these staining patterns indicate the specific areas where X. albilineans localize inside the leaf blade. Non-specific staining was revealed using a pre-immune rabbit serum, and localized to the same locations described above in control leaves (Fig. 3A and C).

Figure 3.

Micrographs from semi-thin cross-sections of sugarcane leaves, symptomatic cv. Mayarí 55-14 inoculated with X. albilineans. Sections were incubated with an antibody raised against X. albilineans (B and D) or rabbit normal serum (A and C), followed by treatment with 5 nm particles conjugated to goat anti-rabbit IgG. The gold labeling was visualized by light microscopy using silver enhancement, and the background was lightly stained with a dilute solution of toluidine blue (1%). The strong silver-enhanced signal shown in (B and D) indicates the possible localization of X. albilineans in the leaf.

X. albilineans secretes xanthan-like polysaccharides in both asymptomatic and symptomatic leaves.

Some bacteria localized near the xylem displayed long, polar flagella identical to those observed in Xanthomonas (Fig. 4A); however, this fact alone is not proof of the unequivocal identity of the genus. Large filamentous material may self-organize inside the xylem vessels to form cross-linked walls made of a mucilaginous-like material that occlude the lumen of the conducting elements (Fig. 4B). Observation of the phloem by TEM after PATAg staining for polysaccharides showed that some of the material occupying the lumen of several phloem vessels in diseased leaves with leaf scald symptoms was dense to electron (Fig. 4C), whereas the inner layer of the cell wall was mainly composed by cellulose and thus transparent to electrons. This indicates that bacteria are not necessarily restricted to the xylem, as previously reported by Birch.¹ However, IGL in Figure 3B and D revealed bacterial accumulation in the fibers surrounding phloem, supporting the possibility that an exocellular, bacterial polysaccharide could invade the phloem. Bacteria were also abundant and widespread in several other leaf tissues, which explained the relative abundance of fibrous exocellular material in these tissues. These fibers were not apparent in diseased leaves observed with TEM, possibly because of their transparency to electrons, which is similar to that of cellulose. However, an accumulation of polysaccharides has been revealed by PATAg staining (Fig. 4C).

Figure 4.

(A) Scanning electron micrograph of a cross section of a diseased leave with leaf scald symptoms. Picture shows a xylem vessel with bacterial cells not only in its neighbourhood, but also penetrating it. Arrow indicates a bacterial cell with a polar flagellum. (B) Scanning electron micrograph of a cross section of xylem vessels of a diseased leaf showing mucilaginous material organized in a reticulate pathway. (C) Transmission electron micrograph PATAg-stained of a cross section of a diseased leaf showing electron-dense material (polysaccharides) inside the phloem vessels (arrows). Xy, xylem; ph, phloem.

Extraction of xanthan-like polysaccharides from sugarcane leaf tissues is very difficult because of the great amount of chlorophylls and protein that could form complexes with bacterial exocellular polysaccharides. Thus, a more feasible approach to study the role of xanthans in Xanthomonas infectivity is to correlate the amount of exocellular polysaccharides produced by bacteria and the sugarcane glycoprotein extracted from sugarcane stalks with the fibrous material invading diseased leaves. Figure 5 shows that bacterial xanthan-like polysaccharide eluted from a Sephadex G-50 column before sugarcane HMMG. This polysaccharide fraction, presumably from bacterial origin, did not contain protein, whereas both HMMG and MMMG contained both protein and polysaccharide moieties. Other sprout stunting diseases of sugarcane also display accumulation of pectinaceous materials derived from the hydrolysis of sugarcane cell walls by hydrolytic, bacterial exocellular enzymes.³² One example is the disease caused by Leifsonia xyli subsp. xyli, a gram-positive coryneform, xylem-limited bacterial pathogen.³¹ The complete absence of L-rhamnose and galacturonic acid in hydrolysates of the first chromatographic eluate (20 to 40 mL) rules out this hypothesis, whereas the occurrence of glucuronic acidand mannose^5,33 permitted us to catalogue this polysaccharide as a bona fide xanthan-like polymer (Fig. 5A) with identical chromatographic behaviour to that described by Blanch et al.³⁴ Appearance of galactitol¹⁸ in hydrolysates of both the second and third eluates (50 to 100 mL) revealed the polysaccharide moiety of sugarcane glycoproteins, HMMG and MMMG (Fig. 5B).

Figure 5.

Elution profile through Sephadex G-50 column of different macromolecules contained in juices extracted from sugarcane stalks from diseased plants with visible leaf scald symptoms. Empty symbols, protein; filled symbols, polysaccharides. Inset (A) Electropherogram of the acid hydrolysates of fractions 20–40 mL, where M, mannose, G, glucose, GA, glucuronic acid, G1p, glucose-1-phosphate. Inset (B) electropherogram of the acid hydrolysates of fractions 40 to 110 mL, where F, fructose and G, galactitol. Numbers near the peaks represent mobility time in min.

The xanthan exopolysaccharide produced by X. fastidiosa in planta has been detected by confocal laser scanning microscopy in the xylem of gravepine petioles using an antiserum against the exopolysaccharide.³⁵ In addition, the numerous tyloses and gels observed in infected plants were similar to those described for L. xyli subsp. xyli,³² but were not immunoreactive. X. fastidiosa GFP-infected petioles incubated with preimmune serum followed by the Alexa Fluor 546 fluorescent conjugate displayed no reactivity either (data not shown).

The amount of xanthan-like polysaccharides correlates with the extent of leaf scald symptoms.

Juices extracted from healthy sugarcane stalks contained no detectable xanthan-like polymers and very low amount of sugarcane glycoprotein. However, discrete amounts of the first product could be extracted from sugarcane stalks without leaf scald symptoms, whereas a lot of xanthanlike macromolecules could be recovered from stalks with visible symptoms of leaf scald. These also contained the highest amount of sugarcane glycoprotein (Fig. 6).

Figure 6.

Quantification of bacterial xanthans (filled rectangles) and sugarcane glycoproteins (HMMG + MMMG, empty rectangles) extracted from healthy and diseased sugarcane stalks without or diseased with visible leaf scald symptoms. Values are the mean of five different replicates. Vertical bars give standard error where larger than the symbols.

During infection, X. campestris pv. campestris produces a gum that has been described as a xanthan.³³ This pathogen bacterium is able to produce xanthan even in culture.³⁶ Other pathogenic bacteria, such as Xylella fastidiosa³⁷ and X. albilineans,⁴ also produce xanthan-like polymers.⁸ However, X. albilineans does not produce these polymers in culture.³⁸ Recently, we found that these bacteria actively produce proteases that rapidly hydrolyze UDP glucose dehydrogenase, the enzyme responsible for the production of glucuronic acid. This is the most characteristic monomer of the xanthan macromolecule.³⁹ Xylella fastidiosa, another pathogenic bacterium that produces xanthan-like gums³⁷ also synthesized proteases but, in contrast to that described for X. albilineans, it did not impede xanthan biosynthesis. This is likely due to the fact that they behave as extracellular proteases secreted into the culture broth. Proteases produced by strains of Xylella fastidiosa from citrus and grape, belong to the serine- and metallo-protease group, respectively.⁴⁰ However, some of the glycoproteins produced by sugarcane after infection act as powerful inhibitors of protease activity.¹⁶ It remains possible that this underlies the restriction of production of the xanthanlike polymer to the infective, symptomatic status.³⁴ According to this rationale, it has been shown that stalk segments of sugarcane experimentally infected with X. albilineans not only produce sugarcane polysaccharides (HMMC and MMMC), but also produce polysaccharides of molecular mass higher than those found in HMMC.⁴¹ It remains unclear why a resistant cultivar enhances the production of a chemical factor that assures the activity of an infectivity factor produced by the pathogen. It is possible that glycoprotein production was a non-specific reaction against mechanical or chemical injury, which would induce non-specific effects and inhibit bacterial proteases. Resistant cv. compensates this effect through an overproduction of hydroxycinnamic and hydroxybenzoic acids, which inhibit bacterial growth; as well as overproduction of lignin, which difficult pathogen entry.⁴²

In conclusion, our data support the hypothesis that the production of fibrous material, presumably a xanthan-like polysaccharide that invades leaf tissues of sugarcane plant with visible symptoms of leaf scald could be due to inhibition of bacterial proteases by sugarcane glycoproteins, thereby allowing production of UDP glucuronic acid and subsequent incorporation of glucuronate to the xanthan structure.

Material and Methods

Plant material.

Plants of commercial Saccharum officinarum (L.) cv. Mayarí 55-14 field grown in the Alfonso XIII Royal Botanical Garden (Complutense University, Madrid, Spain) were used. This cultivar was selected because of its high resistance to leaf scald. Raw material was obtained from healthy or scalded asymptomatic and symptomatic 14-mo-old plants. This resistant cv., better than other susceptible to scald, provides greater frequency and quantity of asymptomatic plants, absolutely required to establish the quantitative relationship between the amount of the xanthan-like polysaccharide and the occurrence of necrotic scald symptoms.

Leaves of 14-mo-old sugarcane plants were infected with a cells suspension of X. albilineans NCPPB 887 in Willbrink medium. Each leaf (three per treatment) was inoculated only with 1 mg dry weight of X. albilineans. 0.5 mL of culture medium from bacteria containing the indicated amount of dry inocule were inoculated through the main vein of the leaf using a syringe with a sterile needle. Control leaves were supplied with 0.5 mL of sterile distilled water with no culture under the same experimental conditions. Plants were maintained for eleven days after inoculation in green house conditions (30°C, 70% relative humidity), prior to leave excision.

Purification of sugarcane juice glycoproteins, HMMG and MMMG.

Stalks from healthy or scalded asymptomatic and symptomatic 14-mo-old sugarcane plants, cv. Mayarí 55-14, field-grown, were mechanically crushed immediately after cut. Crude juice was clarified with saturated sodium carbonate to precipitate soluble proteins that were subsequently removed by centrifugation at 10,000x g for 20 min at 4°C. The pellet was discarded and the supernatant was filtered through filter paper and dialyzed overnight against distilled water. Dialyzed samples were stored at −26°C and then used to separate bacterial exopolysaccharides, high molecular-mass-glycoproteins (HMMG) and mid-molecular mass-glycoproteins (MMMG).

For sugarcane glycoproteins separation, 10 mL of clarified juice were filtered through a Sephadex G-10 column (15 cm × 2.5 cm) embedded in distilled water. Elution was also carried out with MilliQ-grade water. The void volume (22 mL) was discarded, whereas the following 14 mL contained both HMMG and MMMG.^18,19 These were identified as glycoproteins by separation by HPLC and carbohydrate and fluorescence emission analyses. Two filtration processes were required to obtain 28 mL of these glycoproteins. An aliquot of 25 mL of this fraction was then loaded onto a Sephadex G-50 column (30 cm × 2.5 cm). The void volume (40 mL) was discarded and the next 20 mL (40 to 60 mL) contained HMMG; whereas MMMG eluted from 60 to 110 mL.²⁰ Eluted fractions were monitored for carbohydrates according to Dubois et al.²¹ and for protein according to Lowry et al.²²

For separation of the xanthan-like polysaccaride, aliquots of 10 mL of the remaining post-dialyzed supernatant were precipitated with 14.4 mL of iso-propyl alcohol containing 3% (w/v) KCl with continuous shaking at room temperature. The mixture was maintained at 4°C for 2 h and then centrifuged at 14,000x g for 20 min at 2°C.⁵ The supernatant was discarded and the pellet containing precipitated polysaccharides was dissolved in 10 mL of 10 mM sodium phosphate buffer, pH 6.8, to be chromatographed. Polysaccharides were then separated on a Sephadex G-10 (15 cm × 2.5 cm) column pre-equilibrated with 10 mM sodium phosphate buffer, pH 6.8, connected to a G-50 column (30 cm × 2.5 cm) pre-equilibrated as described above. Elution was carried out using the same buffer. A part of the void volume (the initial 20 mL) was discarded. Polysaccharides were eluted in fractions of 5 mL. The volume eluted from 20 to 40 mL contained polysaccharides with a molecular mass higher than that of HMMC. Fractions were monitored for carbohydrates and protein as described above.

Acidic hydrolysis and sugar extraction.

HMMGs and xanthan-like polysaccharides obtained from stalk segments of control, infected without symptoms or infected with symptoms plants and purified by filtration through Sephadex G-50, were hydrolyzed with 6 N HCl at 50°C overnight and then dried under reduced pressure. Each residue was ground with 3 mL of cold 80% (v/v) ethanol, stored for 2 h at 2°C and centrifuged at 19,000x g for 5 min at 2°C. After complete evaporation, precipitates from inoculated and control segments were used for capillary electrophoresis (CE) analysis. Standards were treated equivalently before electrophoresis analysis. Fractions of HMMG obtained from Sephadex G-50 filtration were lyophilized and dissolved in 500 mL of 10 mM sodium borate buffer, pH 9.2 and used for CE analysis.

Capillary electrophoresis.

A P/ACE MDQ Instrument from Beckman Coulter (Fullerton, CA) was used for CE analysis. Detection of saccharides was monitored at 200 and 280 nm. New capillaries were conditioned with 1 M NaOH for 10 min at 60°C and with 0.1 M NaOH for 10 min at 60°C. Equilibration of the capillary was then performed by washing with 25 mM sodium borate buffer, pH 9.2 for 30 min at 25°C and finally with the same buffer for 30 min at 25°C under an applied voltage of 17 kV. Regeneration of the capillary surface between runs was performed by rinsing it with the following sequence of buffers: 0.1 M NaOH for 10 min, Milli-Q grade water for 10 min and 25 mM sodium borate buffer, pH 9.2, for 15 min. The buffer used as electrolyte was 25 mM sodium borate buffer, pH 9.2.²³ Standard solutions (1.0 mgmL⁻¹) and biological samples were injected under pressure (about 40 nL at 0.5 psi for 5 s) and separated at 11 kV using 25 mM borate buffer, pH 9.2, as electrolyte. Data were acquired using 32 Karat™ (v 7.0) software.

Transmission electron microscopy.

The microstructure of sugarcane leaves was examined using conventional TEM as described elsewhere in reference ²⁴. Briefly, samples were fixed in 2% glutaraldehyde (v/v) in 0.1 M phosphate buffer, pH 7.2, post-fixed with osmium tetroxide, washed, dehydrated in acetone and finally embedded in Epon. Thin sections were counter-stained with uranyl acetate and lead citrate. Electronic micrographs were produced at 80 kV by TEM (Carl Zeiss EM 902). Image analysis was performed using Image Tool v2.00. Alternatively, samples were stained for specific detection of polysaccharides (PATAg) after inclusion, following the protocol described by Thiéry and Rambourg.²⁵

Scanning electron microscopy.

For ultrastructural studies, we selected leaves 3 and 4 from the bottom of six different plants, at a middle position on the stalks. The ultrastructure of sugarcane leaves was examined by conventional scanning electron microscopy (SEM). Samples were fixed in 2% glutaraldehyde (v/v) in 0.1 M phosphate buffer, pH 7.2, post-fixed with osmium tetroxide, washed, dehydrated in acetone, critical-point dried, sputter-coated with gold/palladium and scanned at 20 kV by SEM⁵ using a JEOL JSM 6400 (Japan) electron microscope. Digital images were obtained by using built-in INCA software (Oxford).²⁶

Immunogold labeling (IGL).

Small pieces (1.0 mm × 1.0 mm) of sugarcane leaves were taken 11 days after inoculation at the point of inoculation. Later, these samples and control, non-inoculated leaves were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.4) containing 0.1% (v/v) Triton X-100 for 24 h at 4°C. Samples were washed with 0.1 M sodium cacodylate buffer (pH 7.4), dehydrated in an ethanol series and infiltrated in LR White acrylic resin (London Resin Company Ltd., Berkshire, England) for 3 days at 4°C with gently shaking, followed by polymerization at 60°C for 24 h.²⁷ Semi-thin (1 µm) sections were taken from the resin-embedded samples using a Reichert Ultracut E ultramicrotome and collected on Superfrostplus electrostatically charged slides (Erie Scientific Company, Portsmouth, NH), used for immunogold labeling.

IGL buffer²⁷ was used for the blocking stage of the IGL procedure and for diluting the antibody solutions in order to prevent non-specific binding during IGL with rabbit polyclonal antiserum. Briefly, sections on slides were immersed in blocking buffer for 30 min and immunogold labeled for 50 min with a dilution (1:100 v/v) of rabbit antibody raised against X. albilineans, kindly provided by Dr. P. Rott.^28,29 After washing with IGL buffer, sections were treated for 45 min with anti-rabbit IgG gold (5 nm, diluted 1:50 v/v), from Sigma-Aldrich Inc. Sections were then washed twice with IGL buffer and distilled water and, after a short drying period, they were silver-enhanced by use (10 min) of the IntenSE M Kit from Amershan (UK) and background stained with 1% (w/v) toluidine blue O in 1% (w/v) borax solution (diluted 1:300). Sections were visualized in an Olympus BX 51 microscope. In each IGL preparation, a control to recognize non-specific binding of the gold-conjugated antibody was run in parallel, substituting the anti-X. albilineans primary antibody with a rabbit preimmune serum (Sigma-Aldrich Inc.).

Abbreviations

CE

capillary electrophoresis

HMMG

high molecular mass glycoproteins

HPLC

high performance liquid chromatography

IGL

immunogold labeling

MMMH

mid molecular mass glycoproteins

PATAg

periodic acidthiocarbohidrazide-argent

SEM

scanning electron microscopy

TEM

transmission electron microscopy

UDPG

UDP glucose

UDPGDH

UDP glucose dehydrogenase