|Home | About | Journals | Submit | Contact Us | Français|
Artificial environmental conditions in tissue culture, such as elevated relative humidity and rich nutrient medium, can influence and modify tissue growth and induce spontaneous changes from characteristic organization pattern to unorganized callus. As succulent plants with crassulacean acid metabolism, cacti are particularly susceptible to this altered growth environment. Glycosylated proteins of Mammillaria gracillis tissues cultivated in vitro, separated by SDS-PAGE, were detected with Con A after the transfer of proteins onto the nitrocellulose membrane. The glycan components were further characterized by affinity blotting with different lectins (GNA, DSA, PNA, and RCA120). The results revealed significant differences in glycoprotein pattern among the investigated cactus tissues (shoot, callus, hyperhydric regenerant, and tumor). To test whether the N-glycosylation of the same protein can vary in different developmental stages of cactus tissue, the N-glycans were analyzed by MALDI-TOF MS after in-gel deglycosylation of the excised 38-kDa protein band. Paucimannosidic-type N-glycans were detected in oligosaccharide mixtures from shoot and callus, while the hyperhydric regenerant and tumor shared glycans of complex type. The hybrid oligosaccharide structures were found only in tumor tissue. These results indicate that the adaptation of plant cells to artificial environment in tissue culture is reflected in N-glycosylation, and structures of N-linked glycans vary with different developmental stages of Mammillaria gracillis tissues.
Artificial environmental conditions in tissue culture, such as elevated relative humidity and rich nutrient medium, can influence and modify tissue growth nizsation pattern to unorganized callus.1,2 As succulent plants with crassulacean acid metabolism (CAM), which naturally grow in extremely arid and hot areas, cacti are particularly susceptible to this altered growth environment. The plasticity of changes from organized to unorganized manner of growth in vitro makes this system suitable for studies of plant morphogenesis influenced by tissue-culture conditions. Plants of the Mammillaria gracillis Pfeiff. cactus, propagated in vitro, develop callus in the absence of exogenous growth regulators.2 After being removed to fresh medium, this callus grows as a habituated tissue and spontaneously reorganizes into morphologically normal or malformed hyperhydric shoots. The phenomenon of habituation bears a striking similarity to tumor transformation in crown gall disease, where tumor tissue grows independently of exogenous hormones.3,5 In our study, crown gall tumors were induced to compare habituated and transformed calluses. In most cases, habituation appears to be reversible; habituated cells keep their totipotency, as do genetic tumors, and they can regenerate plants or somatic embryos.2,5 Unlike the habituated callus, cactus tumor cells never express any organogenic potential.
It has been documented that in-vitro cultured cell lines and tissue cultures from various origins exhibit changes in glycan formation that are correlated to changes in nutrient medium.6 Mammalian cells apparently can adapt glycosylation as a result of changes in environment and in physiology.7 Only a few studies have been conducted to reveal whether environmental conditions and developmental stage act upon protein glycosylation in plants. It was found that N-glycosylation of extracellular proteins varies with the developmental stage and organization level of plant tissues cultured in vitro.8,9 Stevens et al.10 reported that the glycosylation profile of endogenous proteins can be altered by plant development and growth conditions. The analysis of N-linked glycans of soluble endogenous glycoproteins from leaves of tobacco plants of different age and under different conditions demonstrated that developmental processes in plants could influence glycosylation.11 However, only modest information about the glycoprotein patterns related to cell differentiation, de-differentiation, and transformation in tissue culture is available so far.
In our previous studies, cellular proteins of different Mammillaria gracillis tissues (shoot, callus, hyperhydric regenerant, and tumor) were separated by one-dimensional (1DE) and two-dimensional polyacrylamide gel electrophoresis (2DE) in order to reveal developmentally specific proteins.2,9 Despite obvious morphological differences between Mammillaria tissue lines, only a few morphogenesis-specific polypeptides were observed. Differences between examined tissues were more pronounced in the glycoprotein pattern detected by Con A on 1DE9 and 2DE lectin blots.12 In the present study, glycoproteins were also detected according to their affinity to lectin Con A, while the glycan component was further characterized by means of affinity blotting with different lectins: GNA, DSA, PNA, and RCA120. The composition of N-glycans released with PNGase A from the 38-kDa glycoprotein of shoot, callus, hyperhydric regenerant, and tumor was investigated by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS). The aim of this work was to find out whether the adaptation of plant cells to artificial environment in tissue culture is reflected in N-glycosylation and whether the structures of N-linked glycans vary with different developmental stages of Mammillaria gracillis tissues.
Methanol, acetonitrile, and sodium bicarbonate were obtained from Merck (Darmstadt, Germany) and used without further purification. Distilled and deionized water from Milli-Q water systems (Millipore, Bedford, MA) was used for preparation of the sodium bicarbonate buffer and for sample solutions. Protein Molecular Weight Marker was obtained from Fermentas (St. Leon-Rot, Germany). Con A lectin and peroxidase as well as biotin-labeled lectin RCA120 and streptavidin conjugated to alkaline phosphatase were purchased from Sigma (Steinheim, Germany). Ponceau-S stain, digoxigenin-labeled lectins GNA, DSA, and PNA, peptide-N4-(N-acetyl-β-glucosaminyl)-asparagine amidase from almond (PNGase A), and trypsin were obtained from Roche Applied Science (Mannheim, Germany). ZipTip C18 microcolumns were purchased from Millipore (Bedford, MA). The N-glycans used as standards for MS instrument calibration were obtained from Oxford GlycoSciences. The matrix 2,5-dihydroxybenzoic acid (DHB) was purchased from Sigma.
Mammillaria gracillis plants were propagated in vitro, under 16/8-h light/night photoperiod (light intensity 90 μE/m2sec) at 24°C on solid, MS nutrient medium (0.9% agar, 3% sucrose)13 without any growth regulators. Spontaneously formed callus was detached from the plants and sub-cultivated on the same nutrient medium every 3 wks as a hormone-independent habituated tissue.2 In the callus culture, regeneration of morphologically normal as well as malformed hyperhydric shoots was observed. Regenerated shoots were green, covered with spines, and had a normal growth. Hyperhydrated structures were translucent, light green, more round shaped, and partially covered with softer spines.14 Tumor tissue culture was established from primary tumors induced on shoot explants by Agrobacterium tumefaciens, the wild strain B6S3.2
Total soluble proteins were extracted by grinding 0.5 g of fresh tissue in 1.5 mL of 0.1 M ice-cold Tris/HCl buffer (pH 8.0). in the mixer mill (MM200, Retch, Germany) for 3 min at 30 Hz. The homogenates were centrifuged at 20,000g and 4°C for 15 min. Supernatants were transferred to the centrifugal filter devices (Centricon, 10 MWCO, Millipore) and centrifuged at 5000g and 4°C for 60 min. The filtrate was centrifuged again 10 min at 20,000g at 4°C. Supernatant was collected, and protein content was determined according to Bradford15 using bovine serum albumin as a standard. Samples were denatured by heating for 5 min at 96°C using 0.125 M Tris buffer (pH 6.8), containing 5% (v/v) β-mercaptoethanol and 2% (w/v) sodium dodecyl sulphate (SDS). For the SDS-PAGE, 8 μg of protein was loaded on the gel.
Total soluble proteins were analyzed by SDS-PAGE in 10% T (2.67% C) polyacrylamide gels, with the buffer system of Laemmli.16 The protein bands were visualized by 0.1% Coomassie blue R-250, and gels were scanned as 8-bit grayscale tiff images with an HP Scanjet 2400 scanner (Hewlett-Packard Company, Palo Alto, CA).
The proteins, separated by SDS-PAGE, were electroblotted to a nitrocellulose membrane (0.45 μm, Bio-Rad) in a mini trans blot cell (Bio-Rad) at 60V for 60 min. The transfer buffer was 20 mM Tris-HCl, 150 mM glycine, and 10% (v/v) methanol. The membrane was stained with Ponceau-S stain to confirm the complete transfer of the proteins. The stain was washed off with distilled water. The unoccupied sites of the membrane were blocked by incubating the membrane with 0.1% Tween-20 in TBS buffer (pH 7.5) at 4°C overnight.
The glycoproteins were detected with Con A and peroxidase binding. The bound peroxidase was visualized with 4-chloro naphthol.17,18 Con A specifically binds α-D-mannose and α-D-glucose residues.18,19 Carboxypeptidase Y and transferrin were applied as a positive control, while fetuin and asialofetuin were used as a negative control. For the SDS-PAGE, 2.5 μg of control glycoproteins was loaded on the gel.
The glycan part of proteins was further characterized according to binding of biotin- and digoxigenin-labeled lectins. Applied digoxigenin-labeled lectins were: GNA, specific for terminal mannose18,20; DSA, specific for oligomers of N-acetylglucosamine or galactose-β(1,4)-N-acetyl-glucosamine21,22; and PNA, specific for galactose-β-(1,3)-N-acetylgalactosamine.23–25 The concentration of lectins in incubation solution was 1 μg/mL for GNA and DSA, and 10 μg/mL for PNA. The staining was performed following the manufacturer’s instructions. The membranes were incubated with anti-DIG conjugated with alkaline phosphatase. NBT/BCIP solution was used to determine the presence of alkaline phosphatase. Applied biotin-labeled lectin was RCA120, specific for terminal galactose.22 The concentration of RCA120 in incubation solution was 2.5 μg/mL. The detection procedures included the use of interaction between biotin and streptavidin conjugated to alkaline phosphatase. NBT/BCIP solution was again used for the visualization of the bands. The following control glycoproteins were applied: carboxypeptidase Y was used a positive control for GNA and as a negative control for DSA, PNA, and RCA120; transferrin was used as a positive control for DSA and as a negative control for GNA, PNA, and RCA120; fetuin was used as a positive control for DSA and as a negative control for GNA, PNA, and RCA120; asialofetuin was used as a positive control for PNA and RCA120 and as a negative control for GNA. For the SDS-PAGE, 2.5 μg of each control glycoprotein was loaded on the gel. The membranes were scanned as 8-bit grayscale tiff-images with an HP Scanjet 2400 scanner.
The 38-kDa protein band, which was observed in extracts from all Mammillaria tissues (shoot, callus, hyperhydric regenerant, and tumor), was excised from the Coomassie blue R-250–stained gel. To obtain more proteins, three identical bands of each sample were combined. This method was reported for MALDI analysis of glycoprotein N-glycans from electrophoretic bands corresponding to a few micrograms of proteins from nonmammalian origin.26 It was shown that, depending on the amount of the sample, there is a certain threshold for minor species, which can be lowered by combining two or three protein bands. The bands were first destained with 30% methanol/7.5% acetic acid. Even if gel pieces were incompletely destained, no adverse effect on the subsequent in situ deglycosylation was observed. Oligosaccharide extracts were prepared as previously described by us.12 Briefly, the excised gel pieces were washed in 20 mM NH4HCO3 (pH 7.2) and subsequently reduced and alkylated prior to overnight incubation at 37°C with 12.5 μg/mL trypsin in 25mM NH 4HCO3. The glycopeptides were purified using a ZipTip column, dried in the SpeedVac (SPD 111V evaporator, Savant, Germany), and dissolved in 20 μL 25 mM NH4HCO3 (pH 5.0). In order to inactivate any residual trypsin, the samples were heat treated for 6 min at 95°C. After cooling and a brief centrifugation, 1 μL of PNGase A (5 mU/100 μL) was added and deglycosylation was performed overnight at 37°C.
The N-glycans obtained by in-gel deglycosylation were purified using in-house made graphitized carbon columns, following the procedure described by Küster et al.27 and Šagi et al.28 The purified samples were submitted to dialysis and subsequently concentrated as we reported previously.12
MALDI-TOF MS was performed using 2,5-dihydroxybenzoic acid (DHB).29 The matrix solution was prepared by dissolving 10 mg of 2,5-dihydroxybenzoic acid in 1 mL of acetonitrile:water (7:3, v/v). Vaccum-dried N-glycans were dissolved in 2 mL of water; 0.5 μL of this N-glycan solution was mixed with 0.5 μL of the matrix solution on the stainless steel target and dried at room temperature by a gentle stream of nitrogen. Positive-ion MALDI-TOF mass spectra were acquired in the reflectron mode using a Bruker Reflex III instrument (Bruker Daltonik, Bremen, Germany) equipped with delayed extraction. The acceleration voltage was 20kV and the delayed extraction was set to 2000 nsec. The instrument was externally calibrated using the following commercially available mixture of N-glycans, which was purified in our laboratory: asialo-, agalacto-, biantennary (NGA2), molecular monoisotopic mass 1316.49 Da; asialo-, aga-lacto-, biantennary with core fucose (NGA2F), molecular monoisotopic mass 1462.54 Da; asialo-, galactosylated biantennary with core fucose (NA2F), molecular monoisotopic mass 1786.65 Da; monosialylated, galactosylated biantennary (A1), molecular monoisotopic mass 1931.69 Da. The species of the calibrant were detected as protonated and sodiated pseudomolecular ions.
The electrophoretic pattern of soluble proteins from cactus tissues stained with Coomassie blue R-250 is presented in Figure 1. Most of the polypeptides detected in shoots were also observed in the callus, hyperhydric regenerant, and tumor, although an overall increase of protein content characterized the tumor tissue.
According to the Con A binding pattern, all cactus tissues were characterized by the presence of the 38-kDa glycoprotein band, although the staining intensity was relatively weak in shoot extract (Figure 2). Callus, hyperhydric regenerant and tumor had a common band of approximately 67 kDa, which had a stronger staining intensity in tumor tissue. The 28-kDa band appeared as a strong signal merely in the extract of cactus tumor, while the several glycoproteins in the range of 47–70 kDa were specific for tumor tissue only.
GNA exhibited some similarities with Con A in binding patterns of cactus tissues (Figure 3a). The 38-kDa glycoprotein band was detected in all tissue samples, although it was more pronounced in callus and tumor. The tumor tissue revealed two additional bands of higher molecular masses, 50 and 70 kDa, which were also previously observed with Con A.
The glycoprotein pattern detected with PNA was identical in callus and hyperhydric regenerant (Figure 3b). These tissues were characterized by the presence of the 16-, 24-, 30-, and 70-kDa glycoproteins, which have not occurred in shoot and tumor. All cactus tissues revealed the 48-kDa glycoprotein, while the 40- and 47-kDa ones were observed only in the untransformed tissues. The 50-kDa glycoprotein band was again detected only in cactus tumor. The rest of the bands appeared as weak signals; therefore, they cannot be with certainty considered as the proteins with PNA in their glycan component.
No signal was observed after the treatment with DSA (Figure 3c).
The 70-kDa band was the only glycoprotein that was detected on the membranes treated with RCA120, and it appeared in all cactus samples (Figure 3d).
The MALDI-MS analysis of oligosaccharides released with PNGase A from the 38-kDa glycoprotein revealed a heterogeneous carbohydrate mixture. From the determined m/z values and present knowledge about plant N-glycans, the oligosaccharide structures were predicted. Shoot and callus were characterized by paucimannosidic-type N-glycans. Hyperhydric regenerant and tumor contained complex oligosaccharides of the same composition, while a series of hybrid N-glycans was specific only for tumor tissue.
The structures and composition of paucimannosidic N-glycans, which were observed in shoot and callus, are given in Table 1. Signals detected at m/z 887.01, 909.02, and 925.05 were correlated with [M+Na]+ and [M+2Na-H]+ ions as well as with [M+Na+K-H]+ adducts of the truncated structure, Hex1dHex1HexNAc2Pent1, respectively (Figure 4a). Truncated structure of the same oligosaccharide with one additional mannose, Hex2dHex1HexNAc2Pent1, was observed at m/z 1087.08 as a [M+Na+K-H]+ ion (Figure 4a). Two oligosaccharides bearing one fucose (Fuc) each were identified: a truncated structure, Hex2dHex1HexNAc2, detected as a [M+H]+ ion at m/z 895.07 as well as [M+Na]+, [M+2Na-H]+, and [M+Na+K-H]+ adducts at m/z 917.06, 940.05, and 955.05, respectively (Figure 4a) and an intact structure, Hex3dHex1HexNAc2, detected as a [M+H]+ and [M+Na]+ ions at m/z 1057.98 and 1079.05, respectively (Figure 4b). The [M+Na]+ ion related to the truncated structure with one pentose unit, Hex2HexNAc2Pent1, was observed at m/z 903.06 (Figure 4a). Ion detected at m/z 1046.05 (Figure 4b) could not be assigned to any plant N-glycans.
Hyperhydric regenerant and tumor were characterized by common complex N-glycans, whose structures and composition are listed in Table 2. Signals detected at m/z 1136.13, 1158.11, and 1174.11 were correlated with [M+Na]+, [M+2Na-H]+, and [M+Na+K-H]+ ions of the short complex N-glycan devoid of typical plant glycoepitopes, Hex3HexNAc3, respectively (Figure 5a). Among the complex-type oligosaccharides, two structures with one fucose residue were observed: Hex3dHex1HexNAc3 was detected at m/z 1282.19 as a [M+Na]+ ion and at m/z 1304.17 as a [M+2Na-H]+ ion, while the Hex3dHex1HexNAc4 structure, was detected as a [M+Na]+ ion at m/z 1485.27 (Figure 5a). The signals observed at m/z 1428.25, 1450.23, and 1465.23 corresponded to [M+Na]+, [M+2Na-H]+, and [M+Na+K-H]+ ions of the complex oligosaccharide with two Fuc moieties, Hex3dHex2HexNAc3, respectively (Figure 5a).
The structures and composition of three hybrid N-glycans, which were detected only in tumor tissue, are presented in Table 3. A hybrid-type N-glycan with one Fuc attachment, Hex4dHex1HexNAc2, was detected at m/z 1257.16 as a [M+K]+ ion (Figure 5a). Additionally, two oligosaccharides with one pentose were observed. Thus, the ions detected at m/z 1227.16 and 1249.14 corresponded to [M+Na]+ and [M+2Na-H]+ adducts of the Hex4HexNAc2Pent1 structure (Figure 5a). The oligosaccharide Hex5HexNAc3Pent1 was observed at m/z 1592.29 and 1614.27 as [M+Na]+ and [M+2Na-H]+ ions (Figure 5b). Ion detected at m/z 1367.22 (Figure 5a) could not be assigned to any plant N-glycans.
Glycoproteins of cactus tissues were detected with Con A, which binds specifically to α-D-mannosyl residues and with low affinity to α-D-glucosyl residues.19 After electroblotting of proteins onto the membrane, the strong Con A–positive signal at the position of 38 kDa was noticed in all samples except in the shoot, where it was relatively weak. Some of the Con A–reacting glycoproteins were also detected with GNA, which recognizes terminal D-mannose groups, especially those with Man-α(1-3)-Man units present in high-mannose-type glycans. Crosslinking with GNA indicates that these cactus glycoproteins have high-mannose- or hybrid-type N-glycans. The 38-kDa glycoprotein, detected with GNA, was present in all tissues, being more intensive in the callus and tumor extracts, which corresponds to the results obtained after incubation with Con A. The number of Con A- and GNA-reacting glycoprotein bands increased in tumor tissue. This is in line with observation reported for different morphological stages of sugar beet tissue lines.8 The 70-kDa glycoprotein was the only band present in cactus tissues on membranes treated with RCA120, which demonstrates the presence of the Gal-β(1–4)-GlcNAc sequence. Since the 70-kDa band was also detectable with GNA in tumor tissue, this result can be explained by a hybrid nature of this N-glycan. A similar result was reported for N-glycoproteins specific for different stages of microspore and pollen development in tobacco.18 PNA, which specifically recognizes the Gal-β(1,3)-GalNAc sequence present in O-glycans, revealed the identical glycoprotein pattern of callus and hyperhydric regenerant, and gave the highest number of detected glycoproteins. Interestingly, the 50-kDa glycoprotein from cactus tumor, which gave strong signal with PNA, also reacted with GNA. This result suggests that this protein might posses both N- and O-glycosylation sites, although PNA and GNA may alternatively recognize different glycoproteins of similar molecular size, which are not resolved in SDS-PAGE.
The greatest number of glycosylated tumor proteins was detected with Con A, while the PNA revealed the greatest number of bands in untransformed unorganized cactus tissues. Lectin-binding assay demonstrated that the glycosylation of cellular proteins is changed, if the characteristic organization pattern of Mammillaria shoot is lost.
Results obtained by treatment with Con A and GNA clearly showed that the 38-kDa protein is glycosylated in all cactus tissues. Since the signal was not of the same intensity in all samples, we suspected that the glycosylation profile of this protein might vary among the examined tissues. Therefore, the 38-kDa glycoprotein was selected for further analysis.
The PNGase A was applied to release the oligosaccharides from the 38-kDa glycoprotein because this enzyme is suitable for releasing the N-glycans with the core α1,3-linked fucose, appearing in insect and plant glycoproteins.30,31 The analysis of the resulting sugar mixtures revealed the paucimannosidic-type N-glycans with β1,2-xylose and/or α1,3-fucose linked at the proximal N-acetylglucosamine in shoot and callus samples. These structures are the most common mature glycans N-linked to plant glycoproteins. They are made of a core Man3GlcNAc2, which is common to all of the N-linked glycans in eukaryotic organisms, but with α1,3-linked fucose and/or β1,2-linked xylose. This paucimannosidic-type N-glycans are typical products of the plant N-glycosylation machinery and have been described for a wide range of plant glycoproteins.11,32,33 The presence of paucimannosidic N-glycans, carrying fucosyl and/or xylosyl residues, indicates the vacuolar localization of the 38-kDa glycoprotein in the shoot and callus. Paucimannosidic-type N-glycans have been found only on vacuolar glycoproteins but not in any other compartment of plant cells; in contrast, complex-type N-glycans are associated with Golgi and extracellular glycoproteins.34 Vacuolar glycoproteins containing paucimannosidic-type N-glycans on their mature form, and complex N-glycans cross the same Golgi stacks before targeting their final destination. Paucimannosidic-type N-glycans result from the post-Golgi removal of at least terminal N-acetylglucosamine residues from complex-type N-glycans. However, these modified oligosaccharides may also be formed from larger complex-type glycans by successive action of different exoglycosidases.32 Post-Golgi modification of the 38-kDa glycoprotein N-glycans seems to be missing in tumor and hyperhydric regenerant.
Oligosaccharide mixtures from hyperhydric regenerant and tumor revealed completely different patterns of N-glycan structures. The oligosaccharides common for both tissues were mono- and bi-antennary N-linked complex oligosaccharides with core α1,3-linked fucose. This result suggests a different subcellular localization of the 38-kDa glycoprotein in hyperhydric regenerant and tumor from the vacuolar localization in shoot and callus. Tumor revealed the highest versatility of detected N-glycans in comparison to other cactus tissues. Three hybrid-type N-glycans with structures containing α1,3-linked fucose or β1,2-linked xylose, which are typical for plant glycoproteins,35 was the specificity of the tumor tissue. Two types of N-glycan structures found in tumor demonstrate that the 38-kDa glycoprotein could have two N-glycosylation sites, out of which one site is not glycosylated in shoot, callus, and hyperhydric regenerant, but only in tumor tissue. It is possible that N-glycans at a certain site are fully susceptible to the sugar-processing enzymes (which results in formation of complex N-glycans), while the N-glycans at the other site are partially or not at all accessible to the processing enzymes (which results in presence of hybrid or high-mannose N-glycans, respectively).
Molecular cloning and characterization of cDNA coding for β1,2N-acetylglucosaminyltransferase I (GlcNAc-TI) from Nicotiana tabacum showed that the transfer of bisecting β1,2-xylose and α1,3-linked core fucose requires the presence of at least one terminal GlcNAc.36 As the N-glycans identified in all cactus tissues contain both epitopes, the higher proportion of GlcNAc-containing glycans in hyperhydric regenerant and tumor may reflect differences in N-acetylglucosaminidase activities that govern the biosynthesis of paucimannosidic type glycans, after maturation of the N-glycans in the Golgi compartment. A shift in glycosylation of endogenous proteins in organized tissue (shoot) toward vacuolar-type oligosaccaharides could also be caused by increased induction of vacuolar proteins in this tissue in comparison to the hyperhydric regenerant and tumor. The observed changes in glycosylation pattern between fully organized cactus shoot and partly organized or unorganized tissues such as hyperhydric regenerant and tumor are most likely related to the differences in glycosylation in different developmental stages. Developmental processes of plants, such as senescence, can cause changes in glycosylation pattern.11 The developmental stage of tobacco leaves influenced the protein N-glycoslyation, with a higher proportion of GlcNAc-containing glycans in older leaves compared to younger ones, which suggests that N-glycans are probably being processed gradually during plant tissue maturation. In our experiment, complexity of the N-glycan structures increased with the loss of tissue organization, being the greatest in tumor. This demonstrates that the loss of the characteristic organization pattern of tissue affects the-glycosylation profile of cellular protein. It is well known that habituated and hyperhydrated plant tissues have disturbed cell differentiation, which indicates that habituation and hyperhydricity are ways of tissue juvenilization.3,4 A shift in glycosylation of the 38-kDa protein in hyperhydric regenerant and TW tumor toward complex-type N-glycans with a high GlcNAc content might be caused by the rejuvenilization process of unorganized and only partially organized cactus tissues. The results indicate that the adaptation of plant cells to artificial environment in tissue culture is reflected in N-glycosylation, and structures of N-linked glycans vary with different developmental stages of Mammillaria gracillis tissues.
Identification of the 38-kDa protein in parallel with glycan analysis was performed. We applied the method that combines peptide and glycan mass mapping previously described by Kolarich and Altmann,26 which is particulary devised for glycoproteins that may be found in plants and invertebrates. The band was excised from the protein extracts of all investigated tissues and subsequently submitted to the tryptic digestion. The peptides were analyzed by MALDI-TOF MS. However, a database search gave a result of low similarity with known plant proteins, and the protein could not be identified, most probably because the Mammillaria gracillis genome has not been sequenced so far. Further efforts will be, however, invested in this direction in the future.
The combination of 1DE and MALDI-TOF MS has proved to be a sensitive technique to gain information on global glycosylation as well as on individual variation of N-glycan structures from cactus tissues grown in vitro. Results obtained by affinity blotting with different lec-tines demonstrated that the glycosylation of cellular proteins is changed if the characterisitc organization pattern of Mammillaria shoot is lost. MALDI analysis of shoot and callus oligosaccharides revealed plant-typical paucimannosidic-type N-glycans, which were also present in other cactus tissues. In hyperhydric regenerant and TW tumor, increased number and greater diversity of detected structures were detected, including larger mono- and bi-antennary N-linked complex. The tumor tissue N-glycan pattern was particularly heterogeneou; three oligosaccharides could be assigned as hybrid-type N-glycans. Results obtained in this study indicate that the N-glycosylation pattern of the same protein is highly dependent on the organization level of the plant tissue and can be correlated to the specific morphogenic status of Mammillaria gracillis tissues in in-vitro culture.
The financial support of this work was provided by: the Ministry of Science Education and Sports of the Republic of Croatia, within the project Proteins and Sugars in Plant Development to M.K.-R.; and by Deutsche Forschungsgemeinschaft within Sonderforschungsbereich 492, project Z2, to J.P.-K. B.B. was supported at the University of Münster by the Federation of European Biochemical Societies (FEBS) short-term fellowship in 2004.