Glycobiology Advance Access originally published online on July 21, 2005
Glycobiology 2005 15(12):1386-1395; doi:10.1093/glycob/cwj020
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Monomer/dimer equilibrium of the AB-type lectin from mistletoe enables combination of toxin/agglutinin activities in one protein: analysis of native and citraconylated proteins by ultracentrifugation/gel filtration and cell biological consequences of dimer destabilization
2 Instituto de Química Física Rocasolano, CSIC, 28006 Madrid, Spain; and 3 Institut für Physiologische Chemie, Tierärztliche Fakultät, Ludwig-Maximilians-Universität, 80539 München, Germany
1 To whom correspondence should be addressed; e-mail: d.solis{at}iqfr.csic.es
Received on June 14, 2005; revised on July 15, 2005; accepted on July 18, 2005
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsReferences |
|---|
The biological activity of a lectin is influenced by its quaternary structure. Viscumin is special among the family members of toxic AB-type plant lectins, because it triggers mitogenicity, toxicity, and agglutination. Its activity profile is dependent on the concentration, motivating a thorough inspection of the status of quaternary structure. Over a broad range of protein concentrations (0.01–25 mg/mL), viscumin occurs as a dimer. At high concentrations, the solutions exhibited nonideality, self-association, and polydispersity in sedimentation equilibrium and velocity experiments caused by irreversible aggregation. Calculation of viscumin’s overall shape based on sedimentation velocity data resulted in an elongated dimer form resembling that of crystallized agglutinin. Appearance of monomers was restricted to concentrations in the submicrogram/mL level, as demonstrated by fast protein liquid chromatography gel-filtration analysis. To shift the equilibrium to the monomer for comparative cell biological assays, we performed chemical modification under conditions protecting the lectin activity. Citraconylation was effective to destabilize the dimer. Binding studies by fluorescence-activated cell scan analysis revealed a reduction in cell association upon modification and a tendency for increased sensitivity towards haptenic inhibitors at µg/mL concentrations. Nonetheless, growth inhibition continued to be potent for the ricin-like monomer despite reduced extent of binding. Occurrence of a concentration-dependent monomer/dimer equilibrium appears to achieve the same objectives as the development of two separate protein entities in Ricinus communis, an alternative strategy to emergence of a monomeric toxin, and cell cross-linking dimeric agglutinin.
Key words: agglutinin / cytotoxicity / lectin / mistletoe / ribosome-inactivating protein
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Membrane glycans are considered to represent biochemical signals for the communication of cells with the environment, and lectins serve as effectors to elicit carbohydrate-dependent biosignaling and cellular responses (Solís et al., 2001
; Gabius et al., 2004; Villalobo et al., in press). In this respect, the galactoside-specific mistletoe agglutinin (viscumin) is a notably versatile effector. It is mitogenic for immune and tumour cells at low concentrations (ng/mL; ng/kg body weight), toxic to cells when increasing the concentration, and then also a veritable agglutinin (Endo, 1989; Hajto et al., 1990; Kunze et al., 2000; Gabius, 2001; Gabius et al., 2001; Timoshenko et al., 2001). Together with, for instance, the infamous biohazard ricin, viscumin belongs to the group of AB-type (or type II) ribosome-inactivating proteins (RIPs) (Endo, 1989; Rüdiger and Gabius, 2001; Hartley and Lord, 2004; Stirpe, 2004). These toxins consist of two disulphide–bond-linked subunits A and B. The toxic A chain is a site-specific rRNA N-glycosidase that shuts off protein synthesis (Endo, 1989). The B chain is the lectin part and establishes contact to cell surface glycans to enable internalization of the toxic subunit. Frequently present Gal/GalNAc-containing epitopes are its docking sites so that there is no strict cell-type dependence. Viscumin, for example, recognizes both 
- and ß-galactosides, thereby avidly associating with various types of cells (Gabius et al., 1992, 2001; Lee et al., 1994; Wu et al., 1995; Galanina et al., 1997). On the structural level, these RIPs share homology regarding sequence and folding.
The X-ray crystal structure of viscumin thus presents an overall protein fold similar to that of ricin (Sweeney et al., 1998; Krauspenhaar et al., 1999; Niwa et al., 2003). The A chain is a globular protein that harbours the active site in a centrally located cleft, whereas the B chain folds into two homologous, tandemly arrayed domains (1 and 2), each comprising four subdomains (designated 
, , ß, and 
). Two carbohydrate-binding sites are present in the 1 and 2 subdomains of ricin’s B chain, and galactose binding to the homologous sites has also been observed in the crystal structure of the viscumin–galactose complex (Niwa et al., 2003). In the crystal, two AB monomers following a two-fold symmetry axis face each other at domain 1 of adjacent B chains, and some both polar and nonpolar contacts are established between –', ß–ß', and –' subdomains. This quaternary arrangement may explain the strong effects of certain glycodendrimers on viscumin’s cell binding (André et al., 1999, 2001, 2004), its potency to cross-link glycoproteins and cells (Gupta et al., 1996; Timoshenko et al., 1999), and the high rupture force measured in atomic force microscopy to separate the lectin from its ligand (Dettmann et al., 2000). Because the complete lectin is a stronger elicitor of biosignaling than the isolated B chain (Hajto et al., 1990; Timoshenko and Gabius, 1993), the cross-linking capacity appears to be important for signal induction, the first reason to focus on studying the quaternary structure of the lectin in solution in detail. Incidentally, this structural insight can also be helpful for the design of effective inhibitors of cell binding of this potent toxin.
Other dimeric AB-type RIPs, such as the agglutinins from Abrus precatorius and Ricinus communis (RCA), are categorized as weakly cytotoxic despite their rRNA N-glycosidase activity in a cell-free system (Citores et al., 1993). In contrast, the cytotoxin ricin, abrin, volkensin, and modeccin are invariably monomeric. Thus, viscumin appears to be the only known example of a dimeric RIP-II with strong cytotoxicity. In this respect, previous initial experiments with gel filtration (Olsnes et al., 1982) and electron microscopy (Lutsch et al., 1984) had indicated the possibility for monomer/dimer equilibrium operative at low-protein concentrations, followed by a strong tendency for marked aggregate formation at high concentrations seen by small angle neutron scattering (He et al., 2003). The existence of such an equilibrium implies that viscumin might exert mitogenic and then cytotoxic activity as monomer. Herewith, the study of viscumin affords an attractive natural model to relate dynamic alterations in quaternary structure to cell biological effects. Consequently, we have performed a detailed analysis of the quaternary structure of viscumin using analytical ultracentrifugation and size-exclusion chromatography. With the aim to shift the equilibrium to the monomeric form, site-specific chemical modification with citraconic anhydride, as reagent with only slight effects on lectin activity (Gabius et al., 1992), was performed. Indeed, the stability of the dimers was reduced so that further studies towards our aim to relate quaternary structure to biological properties (cell binding and cytotoxicity) were possible. This work provides evidence for the notion that the occurrence of a dynamic monomer/dimer equilibrium for viscumin is an elegant alternative to satisfy the same requirements met by monomeric/dimeric AB-type protein pairs as present, for example, in R. communis.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Sedimentation-equilibrium analysis
We began analysis of viscumin’s quaternary structure with ultracentrifugation. Preliminary experimental data obtained at a concentration of 0.16 mg viscumin/mL by sedimentation equilibrium at 18,000 rpm could be fitted to a single ideal component (Figure 1A) with a weight-average molecular mass of 104.4 ± 2.3 kDa. However, the Ln(A280 nm) versus r2 plot of these data, which is expected to be a straight line for a single ideal solute, showed a noticeably upwards curvature, indicative of self-association, and/or heterogeneity within the sample (Figure 1B). Therefore, a rigorous analysis using different sample loading concentrations and rotor speeds was required.
|
At 18,000 rpm, the apparent molecular mass (Mapp) decreased with increasing loading concentration (Figure 2A), unequivocally indicating thermodynamic nonideality. Only nonideality can lead to a decrease in apparent molecular weight with concentration so that the measured molecular weight approaches the true molar weight when the concentration approaches zero (Ralston, 1993). The Mapp observed for a 0.01 mg/mL solution of viscumin, the lowest concentration examined, was 119.2 kDa, in perfect agreement with the molecular mass calculated for the viscumin dimer (119.8 kDa). We next examined the influence of the rotor speed on measuring molecular weight. At 15,000 rpm, the decrease in Mapp with increasing loading concentration was significantly lower than at 18,000 rpm, whereas at even slower rotor speeds (10,000 and 8000 rpm), an increase in the apparent molecular weight was observed (Figure 2A). This result argues in favour of the existence of self-association. Furthermore, at increased protein concentrations, the sample was clearly polydisperse, because the measured weight dropped drastically with increasing rotor speed (Figure 2B). Both self-association and polydispersity phenomena were minimized at low-loading concentrations, and—accordingly—the Mapp measured for the 0.01 mg/mL solution was virtually independent of rotor speed (Figure 2B), as expected for a homogeneous sample. A diagnostic graph of the multiple data sets of Mapp versus radial concentration in the cell (Figure 3) confirmed the nonideality, self-association, and polydispersity of the system. Nonoverlapping molecular mass distributions were observed for different loading concentrations, indicating heterogeneity attributable to irreversible aggregation. The upper-limiting molecular weight of the plots was increased at the highest concentrations, a parameter change attributable to self-association, whereas a downward trend in the slope of the plots with increasing loading concentrations can be caused by nonideality. The same behaviour was seen in the presence of either 0.1 M lactose or galactose. At a protein concentration of 0.01 mg/mL, the influence of these three conditions was negligible, and viscumin behaved as a single ideal component with the molecular mass of the dimer. This exclusive presence of the dimer under these conditions was a suitable starting position to spot a chemical modification effecting a shift to the monomer.
|
), 15,000 (n), and 18,000 (u) rpm, and the apparent molecular mass was calculated by fitting the experimental data to a single ideal species model. (B and D) The same data plotted as a function of rotor speed. The loading concentrations were 0.01 (
), 0.1 (
), 1 (,), and 3 (t) mg/mL.
|
Thus, we analyzed viscumin subjected to chemical modification. Indeed, the acetylation of tyrosines (and potentially amino groups via side reactions) or citraconylation restricted to amino groups altered its sedimentation equilibrium behaviour. The fit to a single ideal component of the experimental data obtained for 0.16 mg/mL solutions at 18,000 rpm yielded weight-average molecular masses of 71.1 ± 2.3 kDa and 88.7 ± 3.5 kDa for acetylated and citraconylated viscumin, respectively. These results were clearly below the average molecular mass determined for native viscumin under the same experimental conditions. Focusing on the citraconylated protein, further analysis of its behaviour using different sample loading concentrations (Figure 2C) and rotor speeds (Figure 2D) confirmed that the modified protein also behaved as a polydisperse self-associating system. That no direct evidence of nonideality was detected might be due to the possibility that it could be masked by the heterogeneity of the system, which was inferred at all loading concentrations by a decrease in the measured weight with increasing rotor speed (Figure 2D). Furthermore, an increase in the apparent molecular weight with increasing loading concentrations was consistently observed (Figure 2C), indicating a strong tendency for self-association. At the lowest tested loading concentration (0.01 mg/mL), the observed weight-average molecular mass ranged from 76.8 kDa at 18,000 rpm to 100.8 kDa at 8000 rpm, always significantly below the mass of the viscumin dimer. Thus, the results are consistent with an operative equilibrium between monomers and dimers of citraconylated viscumin. Following the analysis at equilibrium, we proceeded to sedimentation velocity analysis to further characterize the hydrodynamic properties of viscumin and its citraconylated derivative.
Sedimentation-velocity analysis
Sedimentation-velocity studies of native viscumin revealed a major peak together with evidence for the presence of aggregated material. The sedimentation coefficient of the main component and particularly of the aggregated material increased with increasing protein concentration (Figure 4A), reaching values of 6.4 S and 10.1 S, at a concentration of 2 mg/mL. This behaviour confirmed the tendency of self-association noted above. By extrapolation to zero-protein concentration, a so value of 6.0 S (6.27 S for s020,w, expressed for the standard solvent of water at 20°C) was calculated for the main component (Figure 4A, inset).
|
An estimation of the overall shape of the molecule was obtained using the Sednterp program. The partial specific volume and degree of hydration of viscumin were calculated from the amino acid (Huguet et al., 1996, 1998) and carbohydrate (Stoeva et al., 1999) compositions to be 0.7239 mL/g and 0.3586 g/g, respectively. An average carbohydrate composition for the viscumin dimer of 14 Man, 8 NAcHex, 2 Fuc, and 2 Xyl residues, respectively, which contributes to reach a molecular mass of 119,771.6 Da, was used for the calculation. The radius of a compact anhydrous sphere with equivalent mass and volume, which is the minimal possible radius for the nonhydrated protein (R0), is 3.25 nm, and the minimum frictional coefficient (f0) 6.14 x 10–8 g/s. Taking into account the obtained s020,w value of 6.27 S, the ratio between the experimental and minimum frictional coefficients, f/f0, is 1.43. Based on this frictional ratio, viscumin appears to be hydrodynamically equivalent to an oblate ellipsoid with axial dimensions of 13.2 nm (2a) and 2.4 nm (2b) or to a prolate ellipsoid with axial dimensions of 21.9 nm (2a) and 4.3 nm (2b). Of note, the obtained s020,w value is higher than the maximal possible sedimentation coefficient predicted for a particle with the mass of a viscumin monomer. To figure out further parameters of the impact of chemical modification initially detected above, viscumin derivatives were similarly analyzed.
Chemical modification of tyrosines and/or lysines led to a marked alteration of the sedimentation-coefficient distribution c(s) towards lower s values (Figure 4B). Sedimentation coefficients of 4.4 S and 4.35 S were obtained for 0.25 mg/mL solutions of acetylated and citraconylated viscumin, respectively, as compared with 6.1 S for the native protein. Approximation of the protein shape to an ellipsoid of revolution, considering s020,w values of 4.6 S and 4.56 S and the molecular mass of the dimer, yielded minimum axial a/b ratios of 13, with obviously unreasonable protein dimensions. It is thus the only option to use the molecular mass of the monomer for a meaningful calculation, and a frictional ratio f/f0 of 1.24 was obtained. According to this ratio, both the acetylated and citraconylated viscumin derivatives appeared to be hydrodynamically equivalent to prolate ellipsoids with axial dimensions of 10.9/11.5 nm (2a) and 4.3/4.2 nm (2b), respectively, or to ellipsoids with an equatorial axis equal to that obtained for native viscumin but with a half-length polar axis. Accordingly, the behaviour of the derivatized protein appeared to resemble that expected of a monomer. This set of experiments thus fully confirmed and extended the data obtained by sedimentation-equilibrium analysis. To examine the behaviour of highly diluted solutions not covered by analytical ultracentrifugation, we performed an fast protein liquid chromatography (FPLC) gel-filtration analysis.
Gel-filtration analysis
The elution time of viscumin at concentrations between 0.05 and 25 mg/mL was 22.6 ± 0.2 min, very close to that observed for the dimeric R. communis agglutinin RCA (22.2 ± 0.2 min) (Figure 5). Of note, analysis of a 0.5 µg/mL solution of 125I-viscumin revealed a broadening of the peak with a shift displacement of the peak maximum upwards to 25.7 min, approximating the elution time observed for monomeric ricin (26.4 ± 0.1 min), and the consistent appearance of a small leading shoulder of this main peak, centred at 22.6 min. This behaviour was not caused by radio iodination of the protein, because the solution set to a concentration of 0.5 µg/mL of 125I-viscumin eluted as a single peak at 22.8 min when in the presence of 0.5 mg/mL of unlabelled protein (Figure 5). Further analysis of increasingly diluted solutions of unlabelled viscumin, from 30 to 2 µg/mL, made feasible by monitoring the protein elution at 226 nm, showed a progressive displacement of the peak maxima from 23.1 ± 0.3 to 25 ± 0.4 min (Figure 5, inset). Effectively, the results appeared to reflect the dissociation of viscumin dimers into monomers at the eluted protein concentrations. Assuming a typical dilution factor of ten, as estimated from the ratio of width at half height of the peak to the injection volume, dissociation became visible at eluted concentrations below 3 µg/mL.
|
The ultracentrifugation data given above had indicated that chemical modification affected the quaternary structure. Performing gel filtration, both acetylation and citraconylation of viscumin resulted in an evident increase in the elution volume (Figure 5), the derivatives reaching retention times of 24.6 ± 0.4 min and 24.3 ± 0.4 min, respectively. As control, no effect of either tyrosine iodination or tryptophan oxidation on the chromatographic behaviour of viscumin was observed, excluding a general effect of modification on quaternary structure (not shown). Thus, citraconylation afforded an opportunity to pinpoint any correlation between the quaternary structure and biological effects of viscumin. We thus examined parameters of cell binding, that is, extent of cell positivity, intensity of staining and sugar inhibition, and cytotoxicity with viscumin and its derivative.
Cell binding and cytotoxicity
Viscumin and the chemically modified derivative bound to cells in a galactose-dependent manner. Because the glycomic profiles of cells depend on their histogenesis, we selected five different lines to study binding patterns covering the phenotypes of different carcinomas, B-lymphoblastoid cells, and fibroblasts. When running systematic experiments with increasing lectin concentrations, clearly decreased avidity of the derivative was observed for the breast carcinoma and fibroblast lines. Exemplarily, the disparate behaviour is shown at one concentration in Figure 6A. Albeit less pronounced, binding of native viscumin surpassed that of the derivative also in the other three cases (Figure 6). Cell association was nearly completely inhibitable by competition with lactose and asialofetuin, and we document the relative inhibitory capacity of 50 mM lactose on lectin binding at a nonsaturating concentration (Figure 6B). In line with the detected lower level of binding, the sensitivity towards inhibition appeared to be higher for the derivative than for native viscumin. Used at 100 ng lectin/mL, this lactose concentration invariably blocked binding of the treated lectin, although being on average less efficient for native viscumin (not shown). Evidently, preferential presence of the dimer favours binding and its resistance to competitive inhibition. As a means to probe cell entry, we determined cytotoxicity comparatively. The scope for the interpretation of the results, however, was inevitably limited by the strong growth inhibition seen even at low concentrations of viscumin. In detail, this parameter of native and citraconylated viscumin against various cell lines was evaluated by measuring cell growth after incubation with lectin under different culture conditions with similar results. To detect any evidence for differential potency, we titrated the biological activity and then routinely performed experiments at 0.1 µg/mL. Recalling the FPLC data, native viscumin dimers will at this concentration already dissociate to a significant extent. Inevitably, viscumin’s potent toxicity yet precluded to run cytotoxicity assays at increased concentrations, where toxicity invariably reached high levels. Extent of viscumin toxicity at this selective concentration was perceptibly dependent on the cell type (Table I), resulting in about 56% inhibition of the proliferation of NIH 3T3 fibroblasts but only 24% inhibition on SW620 cells, obtained from a lymph node metastasis of a primary colon cancer. Colon cancer cells were less sensitive than fibroblasts to viscumin under these conditions. Compared with native viscumin, the citraconylated lectin maintained toxicity and exhibited a rather similar potency against fibroblasts and also against human SW480 colon adenocarcinoma cells. The lower level of cell binding appeared to be compensated by more efficient post-binding effects to reach similar toxicity levels in these cases (Table I). In the other cases, compensation was not as effective, but the results still showed that the chemical destabilization of the dimer will affect toxicity to a smaller extent as expected from effects on binding and sensitivity to lactose-dependent inhibition. The monomer is thus an active toxin, and dimerization maintains toxicity.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Viscumin can be singled out from other AB-type toxins because of its quaternary structure. To resolve the question how viscumin can cover the functional spectrum of monomeric/dimeric AB-type proteins found in other plants, we scrutinized its quaternary structure. We have shown here that viscumin forms a dimer in solution over a broad range of protein concentrations (0.01–25 mg/mL), at which it displays strong agglutinin and toxic activities (André et al., 2001
; Gabius et al., 2001). Viscumin solutions exhibit nonideality, self-association, and polydispersity attributable to irreversible aggregation, as evidenced by sedimentation equilibrium and velocity experiments. When diluting the concentration up to 0.01 mg/mL, the protein behaves as a single ideal component with the molecular mass of the dimer. Estimation of the overall shape of the protein based on sedimentation velocity data indicates that the viscumin dimer is hydrodynamically equivalent to a prolate ellipsoid of revolution with axial dimensions of 21.9 nm (2a) and 4.3 nm (2b). This elongated shape, which possibly underlies the observed nonideality, resembles the shape of the protein in the crystal structure (15.7 x 6.3 x 4.8 nm3) (Sweeney et al., 1998; Krauspenhaar et al., 1999; Niwa et al., 2003). Viscumin’s size at further lowered protein concentrations was examined by FPLC gel-filtration analysis.
At very low concentrations, viscumin monomers were detected (eluted concentrations <3 µg/mL). Using conventional gel filtration on Sephacryl 200, the apparent complete dissociation of viscumin dimers at a loading concentration of 7.5 µg/mL might thus be explained by the dilution of the protein during chromatography, fractions with dimers and monomers both being cytotoxic against mouse 501.1 cells (Olsnes et al., 1982). Moreover, according to our results, the appearance of monomers at protein concentrations of 0.04 mg/mL in an earlier electron microscopic study by negative staining of viscumin (Lutsch et al., 1984) is likely a consequence of the sample processing, because no monomers are present in solution at this concentration, as independently detected previously (Sweeney et al., 1998). The quoted study also revealed a propensity of viscumin’s B chain to dimerize in contrast with ricin’s B chain. Thus, viscumin appears to have a unique ability for monomer/dimer equilibrium among AB toxins. To learn more about the properties of a ricin-like monomer, a shift of the dynamic equilibrium towards the monomer would be helpful. As a means to destabilize the dimer, we exploited chemical modification.
Different gel-filtration behaviour of the lectin was the consequence of acetylation with N-acetylimidazole or citraconylation, with elution times approaching that of a monomer. A thorough analysis of citraconylated viscumin by sedimentation equilibrium revealed polydispersity and self-association compatible with an operative equilibrium between monomers and dimers. Furthermore, the modified proteins’ shape was hydrodynamically equivalent to prolate ellipsoids of dimensions close to those of a monomer. Because the side chain of Tyr68 is involved in few contacts at the dimer interface, acetylation of this residue might compromise the integrity of the dimer. In addition to tyrosines, primary amines could also be acetylated by N-acetylimidazole and, in fact, selective citraconylation of amino groups was sufficient for destabilizing the dimer. Although no contacts at the interface involving a lysine residue were visible in the crystals—obtained at acidic pH (Niwa et al., 2003)—it is conceivable that a salt bridge between Lys115 of one monomer and Asp26 of the neighbouring B chain is established at neutral pH. Its presence and impairment by chemical modification could then have a bearing on the stability of the dimer in solution.
The destabilization of the dimer by citraconylation afforded the opportunity to examine the consequences of shifting the equilibrium in quaternary structure on cell binding. Assays with native cells of different histogenesis documented a tendency for less avid binding and increased sensitivity towards haptenic inhibition with lactose. Of note, chemical modification was performed under conditions protecting the carbohydrate-binding site. Evidently, the monomer will thus still be able to bind glycans, but its potency to establish firm initial contacts is clearly reduced. These properties are fully in line with a functionality as signaling effector and vector for endocytic uptake, akin to monomeric ricin’s potency to act as toxin and as immunomodulator in the ng/mL range (Hajto et al., 1990; Gabius et al., 1992; Licastro et al., 1993). Compared with RCA, the strength of glycan binding is thus expected to be reduced, and, indeed, the rupture force to break viscumin’s contact to asialofetuin is considerably smaller than for RCA (Dettmann et al., 2000).
Because the A-subunit does not contain lysine groups, it should not undergo any significant modification (apart from the potential citraconylation of the terminal amino group). Observed changes in cytotoxicity can therefore be attributed to the change in the quaternary structure, although effects caused by the modification of the B-chain cannot be excluded. At any rate, the overall conformation of this subunit did not seem to be affected by mild citraconylation, as intimated by the preservation of the sugar-binding ability and substantiated by near-ultraviolet circular dichroism spectroscopy (data not shown). However, the replacement of positively charged amino groups by the negatively charged derivatives definitely modifies the charge distribution pattern at the protein surface and will disrupt previously existing long-range Coulombic interactions engaging the lysine residues, with predictably unfavourable consequences on the stability of the protein. This destabilization may matter for the intracellular trafficking of the toxin after binding to the cell surface and endocytosis.
As mentioned in the Introduction, the viscumin dimer harbours some contacts between adjacent B-chains, including the hairpin loop of the 1 subdomain. In ricin and RCA, this loop is stabilized by a disulphide bridge. Notably, the substitution of one of the cysteines involved in this bridge, Cys40, to serine in viscumin endows the loop with increased flexibility. This parameter has been suggested to be relevant for dimerization (Sweeney et al., 1998; Krauspenhaar et al., 1999; Niwa et al., 2003). In contrast, dimerization of RCA occurs through the A-chains, and the RCA dimer is covalently stabilized by a disulphide bridge between adjacent A-chains (Sweeney et al., 1997). Although the interface region underlies structurally critical substitutions, the close relationship between the two Ricinus proteins and viscumin is underscored by their similar reactivity pattern to ß-lactoside analogues in chemical mapping of the binding site (Rüdiger et al., 2000). The monomer/dimer equilibrium of viscumin can thus be considered as an alternative strategy to having separate monomeric and dimeric type-II RIPs available in a plant. Within the galectin family of endogenous ß-galactoside-specific lectins, presence of separate monomers and dimers as well as of modules for reversible aggregation upon ligand contact appears to reach the same objectives in a sophisticated manner (André et al., 2003, 2005; Kopitz et al., 2003; Ahmad et al., 2004).
In summary, our study combining analysis of viscumin’s hydrodynamic characteristics, site-specific chemical modification to destabilize dimers and cell biological assays reveals this lectin to act as potent toxin as monomer, like ricin, with increases in binding avidity to cells as dimer maintaining the strong toxicity to guarantee protection against predators. Dynamic monomer/dimer equilibrium thus qualifies as a suitable strategy to combine functions of monomeric/dimeric AB-type protein pairs in a single protein.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Isolation and radio iodination of viscumin
The galactoside-specific lectin was purified from supernatants of extracts of dried mistletoe (Viscum album) leaves by affinity chromatography on lactosylated Sepharose 4B, obtained by divinyl sulfone activation, as crucial step, and analyzed for purity and activity by one- and two-dimensional gel electrophoresis, an enzyme-linked lectin-binding assay and haemagglutination (Gabius et al., 1992
; André et al., 1999). The protein concentration was determined by the Lowry assay using concanavalin A (Sigma, St. Louis, MO) as standard (Jiménez et al., 2000). The toxin (2.5 mg) was labelled with 0.2 mCi of 125I, in the presence of 0.1 M lactose to protect the sugar-binding sites, using Iodogen (Pierce, Rockford, IL), according to the manufacturer’s recommendations. The molecular weight of the lectin was not altered by any degradation during labelling, as ascertained by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography.
Preparation of antiviscumin antibody
To avoid raising cross-reactive antibodies against the glycan chains covalently linked to the protein, the purified lectin was subjected to periodate oxidation and reduction of the resulting vicinal aldehydes by NaBH4 treatment, as described (Hajto et al., 1989). Before immunization of the rabbits, the toxic lectin was treated with 2% (v/v) formaldehyde. The immunoglobulin G fraction obtained using the periodate- and formaldehyde-treated lectin as antigen was purified by affinity chromatography on protein A-Sepharose 4B (Pharmacia, Freiburg, Germany), and antibody specificity was ascertained by immunospotting, immunoblotting, and enzyme-linked immunosorbent assay (ELISA) assays with preimmune serum as control, as described (Hajto et al., 1989). In these assays, the reactivity of the antibody was evaluated to be 10–15% lower for citraconylated viscumin than for the native lectin.
Chemical modifications
The reactions were carried out in the presence of 0.1 M lactose to protect the carbohydrate-binding sites from modification. Tyrosine iodination was carried out in 0.1 M sodium phosphate buffer, pH 7.0, with 0.05 N resublimed iodine in 0.1 M KI, in a proportion of 11 µL per mg of protein (Ishiguro et al., 1977). Acetylation of tyrosine moieties with N-acetylimidazole was carried out in 5 mM sodium phosphate buffer, pH 7.2, containing 0.2 M NaCl (phosphate-buffered saline [PBS]), using a 20-fold excess of the reagent (Gabius et al., 1992). Amino groups were modified in Tris–HCl 0.1 M, pH 8.0, using 1.3 µL of citraconic anhydride per mg of protein (Gabius et al., 1992), and tryptophan residues underwent oxidation with N-bromosuccinimide at pH 4.0 (Sandvig et al., 1978). Excess reagents were removed by exhaustive dialysis against PBS at 4°C. Preservation of the carbohydrate-binding ability of the chemically modified lectin was confirmed by affinity chromatography on an asialofetuin-Sepharose 4B column.
Analytical ultracentrifugation
Sedimentation equilibrium experiments were performed by centrifugation of 80-µL samples adjusted to different protein concentrations at 8000, 10,000, 15,000, and 18,000 rpm and 20°C in an Optima XL-A analytical ultracentrifuge (Beckman Instruments, Fullerton, CA), equipped with UV–visible optics and an An50Ti analytical rotor. Data were collected using 12-mm path length double-sector six-channel centerpieces with quartz windows. Weight-average molecular weights, Mw, were calculated with the XLAEQ program, using the signal conservation algorithm (Minton, 1994).
Sedimentation velocity experiments were carried out in the same ultracentrifuge at 45,000 rpm and 20°C, using 400-µL samples. Up to 20 scans were carried out every 5 min, and the corresponding buffer signal was subtracted. Sedimentation coefficients were calculated using the programs XLAVEL, supplied by Beckman, and SEDFIT (Schuck, 2000), yielding similar results. Solvent density and viscosity at 20°C were calculated using the Sednterp software.
Gel filtration
Gel filtration was carried out with a Superose 12 HR 10/30 column (Pharmacia Biotech, Uppsala, Sweden; void volume, 7.5 mL) equilibrated with PBS, 0.02% NaN3, containing 0.1 M lactose to prevent interactions of the lectin with the agarose-based matrix. The flow rate was 0.5 mL/min, and the elution was monitored at 280 nm or alternatively at 226 nm for the most diluted samples. Ricin and the agglutinin RCA, obtained from R. communis seeds (Jardín Botánico, CSIC, Madrid, Spain) (Nicolson et al., 1974), were fractionated as controls under similar conditions. When radio-iodinated viscumin was chromatographed, 300 µL of fractions were collected and their radioactivity was measured in an LKB MiniGamma counter.
Flow cytofluorimetry
Cell pellets of the human mammary carcinoma line MDA-MB-435 established from pleural effusion of a patient, kindly provided by Dr. W. Kemmner (Berlin, Germany), the human colon adenocarcinoma lines SW480 from a primary tumor and SW620 from a lymph node metastasis of the primary colon cancer (American Type Culture Collection, Rockville, MD), the human B-lymphoblastoid line Croco II established from a patient with M4 myelomonocytic leukemia and propagated as described (Gabius et al., 1991), and the NIH 3T3 fibroblast line were carefully washed with Dulbecco’s PBS solution containing 0.1% carbohydrate-free bovine serum albumin to remove any inhibitory serum glycoproteins and to saturate nonspecific protein-binding sites on the cell surface. Afterwards, cells were incubated for 30 min with 0.5–1 µg viscumin/mL at 4°C to avoid uptake by endocytosis. Following thorough washing to remove unbound lectin, the staining profiles of the cell populations were monitored by fluorescence-activated cell scan analysis, using standard equipment, including the software package CellQuest Pro (Becton-Dickinson, Heidelberg, Germany), with a polyclonal rabbit immunoglobulin G fraction against viscumin (10 µg/mL) and a goat anti-rabbit immunoglobulin G fluoresceinthicocorbamyl-conjugate (Sigma, Munich, Germany; 1:100 dilution of the commercial solution) in the second step, as described (André et al., 2003, 2004). Omission of the incubation step with the lectin from the standard protocol served as control to assess any occurrence of lectin-independent staining and carbohydrate-dependent binding was ascertained by using increasing lactose concentrations or a mixture of 0.2 M lactose with 0.5 mg asialofetuin/mL as inhibitors of galactose-dependent binding. Similar controls were performed to account for any binding of the lectin’s N-glycan chains to receptors on the tumour cell surface which had been detected by neoglycoproteins (Gabius et al., 1987, 1991).
Cell growth assay
Cells (1 x 104 cells per assay for adherent cells and 5 x 104 for cells in suspension) were seeded in 96-well plates and cultured for 24, 48, and 72 h either with AIM V® medium (Invitrogen, Karlsruhe, Germany) without fetal calf serum or with RPMI-1640 medium containing 5% (v/v) heat-inactivated fetal calf serum, both supplemented with antibiotics and 100 ng viscumin/mL. Control cells were cultured in parallel under identical experimental conditions but without lectin addition. Cell growth was comparatively assessed by the application of 0.5 mg/mL of the chromogen MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide).
|
Acknowledgements |
|---|
We thank DGICYT (BQU2000-1501-C02-02 and BQU2003–03550-C03-03), the Mizutani Foundation for Glycoscience, and the Verein zur Förderung des biologisch-technologischen Fortschritts in der Medizin e. V. for financial support.
|
Abbreviations |
|---|
FPLC, fast protein liquid chromatography; PBS, phosphate-buffered saline; RIP, ribosome-inactivating protein
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsReferences |
|---|
Ahmad, N., Gabius, H.-J., André, S., Kaltner, H., Sabesan, S., Roy, R., Liu, B., Macaluso, F., and Brewer, C.F. (2004) Galectin-3 precipitates as a pentamer with synthetic multivalent carbohydrates and forms heterogeneous cross-linked complexes. J. Biol. Chem., 279, 10841–10847.
André, S., Cejas Ortega, P.J., Perez, M.A., Roy, R., and Gabius, H.-J. (1999) Lactose-containing starburst dendrimers: influence of dendrimer generation and binding-site orientation of receptors (plant/animal lectins and immunoglobulins) on binding properties. Glycobiology, 9, 1253–1261.
André, S., Pieters, R.J., Vrasidas, I., Kaltner, H., Kuwabara, I., Liu, F.-T., Liskamp, R.M.J., and Gabius, H.-J. (2001) Wedgelike glycodendrimers as inhibitors of binding of mammalian galectins to glycoproteins, lactose maxiclusters and cell surface glycoclusters. Chembiochem, 2, 822–830.[CrossRef][ISI][Medline]
André, S., Liu, B., Gabius, H.-J., and Roy, R. (2003) First demonstration of differential inhibition of lectin binding by synthetic tri- and tetravalent glycoclusters from cross-coupling of rigidified 2-propynyl lactosides. Org. Biomol. Chem., 1, 3909–3916.[CrossRef][ISI][Medline]
André, S., Kaltner, H., Furuike, T., Nishimura, S.-I., and Gabius, H.-J. (2004) Persubstituted cyclodextrin-based glycoclusters as inhibitors of protein-carbohydrate recognition using purified plant and mammalian lectins and wild-type and lectin-gene-transfected tumor cells as targets. Bioconjug. Chem., 15, 87–98.[CrossRef][ISI][Medline]
André, S., Kaltner, H., Lensch, M., Russwurm, R., Siebert, H.-C., Fallsehr, C., Tajkhorshid, E., Heck, A.J.R., von Knebel-Döberitz, M., Gabius, H.-J., and Kopitz, J. (2005) Determination of structural and functional overlap/divergence of five proto-type galectins by analysis of the growth-regulatory interaction with ganglioside GM1 in silico and in vitro on human neuroblastoma cells. Int. J. Cancer, 114, 46–57.[CrossRef][ISI][Medline]
Citores, L., Ferreras, J.M., Iglesias, R., Carbajales, M.L., Arias, F.J., Jimenez, P., Rojo, M.A., and Girbes, T. (1993) Molecular mechanism of inhibition of mammalian protein synthesis by some four-chain agglutinins. Proposal of an extended classification of plant ribosome-inactivating proteins (rRNA N-glycosidases). FEBS Lett., 329, 59–62.[CrossRef][ISI][Medline]
Dettmann, W., Grandbois, M., André, S., Benoit, M., Wehle, A.K., Kaltner, H., Gabius, H.-J., and Gaub, H.E. (2000) Differences in zero-force and force-driven kinetics of ligand dissociation from ß-galactoside-specific proteins (plant and animal lectins, immunoglobulin G) monitored by plasmon resonance and dynamic single molecule force microscopy. Arch. Biochem. Biophys., 383, 157–170.[CrossRef][ISI][Medline]
Endo, Y. (1989) Mechanism of action of ricin and related toxic lectins on the inactivation of eukaryotic ribosomes. Adv. Lectin Res., 2, 60–73.
Gabius, H.-J. (2001) Probing the cons and pros of lectin induced immunomodulation: case studies for the mistletoe lectin and galectin-1. Biochimie, 83, 659–666.[Medline]
Gabius, H.-J., Engelhardt, R., Hellmann, T., Midoux, P., Monsigny, M., Nagel, G.A., and Vehmeyer, K. (1987) Characterization of membrane lectins in human colon carcinoma cells by flow cytofluorometry, drug targeting and affinity chromatography. Anticancer Res., 7, 109–112.[ISI][Medline]
Gabius, S., Joshi, S.S., Gabius, H.-J., and Sharp, J.G. (1991) Establishment, characterization and determination of cell surface sugar receptor (lectin) expression by neoglycoenzymes of a human myeloid marker-expressing B lymphoblastoid cell line. Anticancer Res., 11, 793–800.[ISI][Medline]
Gabius, H.-J., Walzel, H., Joshi, S.S., Kruip, J., Kojima, S., Gerke, V., Kratzin, H., and Gabius, S. (1992) The immunomodulatory ß-galactoside-specific lectin from mistletoe: partial sequence analysis, cell and tissue binding, and impact on intracellular biosignalling of monocytic leukemia cells. Anticancer Res., 12, 669–676.[ISI][Medline]
Gabius, H.-J., Darro, F., Remmelink, M., André, S., Kopitz, J., Danguy, A., Gabius, S., Salmon, I., and Kiss, R. (2001) Evidence for stimulation of tumor proliferation in cell lines and histotypic cultures by clinically relevant low doses of the galactoside-binding mistletoe lectin, a component of proprietary extracts. Cancer Invest., 19, 114–126.[CrossRef][ISI][Medline]
Gabius, H.-J., Siebert, H.-C., André, S., Jiménez-Barbero, J., and Rüdiger, H. (2004) Chemical biology of the sugar code. Chembiochem, 5, 740–764.[CrossRef][Medline]
Galanina, O.E., Kaltner, H., Khraltsova, L.S., Bovin, N.V., and Gabius, H.-J. (1997) Further refinement of the description of the ligand-binding characteristics for the galactoside-binding mistletoe lectin, a plant agglutinin with immunomodulatory potency. J. Mol. Recognit., 10, 139–147.[CrossRef][ISI][Medline]
Gupta, D., Kaltner, H., Dong, X., Gabius, H.-J., and Brewer, C.F. (1996) Comparative cross-linking activities of lactose-specific plant and animal lectins and a natural lactose-binding immunoglobulin G fraction from human serum with asialofetuin. Glycobiology, 6, 843–849.
Hajto, T., Hostanska, K., and Gabius, H.-J. (1989) Modulatory potency of the ß-galactoside-specific lectin from mistletoe extract (Iscador) on the host defense system in vivo in rabbits and patients. Cancer Res., 49, 4803–4808.
Hajto, T., Hostanska, K., Frei, K., Rordorf, C., and Gabius, H.-J. (1990) Increased secretion of tumor necrosis factor-, interleukin-1, and interleukin-6 by human mononuclear cells exposed to galactoside-specific lectin from clinically applied mistletoe extract. Cancer Res., 50, 3322–3326.
Hartley, M.R. and Lord, J.M. (2004) Cytotoxic ribosome-inactivating proteins from plants. Biochim. Biophys. Acta, 1701, 1–14.[Medline]
He, L., André, S., Siebert, H.-C., Helmholz, H., Niemeyer, B., and Gabius, H.-J. (2003) Detection of ligand- and solvent-induced shape alterations of cell-growth-regulatory human lectin galectin-1 in solution by small angle neutron and X-ray scattering. Biophys. J., 85, 511–524.
Huguet, M., Stoeva, S., Schwamborn, C., Wilhem, T.S., and Voelter, W. (1996) Complete amino acid sequence of the A chain of mistletoe lectin I. FEBS Lett., 399, 153–157.[CrossRef][ISI][Medline]
Huguet, M., Stoeva, S., and Voelter, W. (1998) Complete amino acid sequence of the B chain of mistletoe lectin I. Biochem. Biophys. Res. Commun., 246, 596–601.[CrossRef][ISI][Medline]
Ishiguro, M., Taira, E., Mashima, K., Funatsu, G., Funatsu, M., and Kikutani, M. (1977) Iodination of ricin. Agric. Biol. Chem., 41, 299–307.
Jiménez, M., André, S., Gabius, H.-J., and Solís, D. (2000) Calibration of colorimetric protein assays for quantitation of plant AB toxins. Anal. Biochem., 284, 418–420.[CrossRef][ISI][Medline]
Kopitz, J., André, S., von Reitzenstein, C., Versluis, K., Kaltner, H., Pieters, R.J., Wasano, K., Kuwabara, I., Liu, F.-T., Cantz, M., and others. (2003) Homodimeric galectin-7 (p53-induced gene 1) is a negative growth regulator for human neuroblastoma cells. Oncogene, 22, 6277–6288.[CrossRef][ISI][Medline]
Krauspenhaar, R., Eschenburg, S., Perbandt, M., Kornilov, V., Konareva, N., Mikailova, I., Stoeva, S., Wacker, R., Maier, T., Singh, T., and others. (1999) Crystal structure of mistletoe lectin I from Viscum album. Biochem. Biophys. Res. Commun., 257, 418–424.[CrossRef][ISI][Medline]
Kunze, E., Schulz, H., Adamek, M., and Gabius, H.-J. (2000) Long-term administration of galactoside-specific mistletoe lectin in an animal model: no protection against N-butyl-N-(4-hydroxybutyl)-nitrosamine-induced urinary bladder carcinogenesis in rats and no induction of a relevant local cellular immune response. J. Cancer Res. Clin. Oncol., 126, 125–138.[CrossRef][ISI][Medline]
Lee, R.T., Gabius, H.-J., and Lee, Y.C. (1994) The sugar-combining area of galactose-specific toxic lectin of mistletoe extends beyond the terminal sugar residue: comparison with a homologous toxic lectin, ricin. Carbohydr. Res., 254, 269–276.[CrossRef][ISI][Medline]
Licastro, F., Morini, M.C., Bolognesi, A., and Stirpe, F. (1993) Ricin induces the production of tumor necrosis factor- and interleukin-1ß by human peripheral blood mononuclear cells. Biochem. J., 294, 517–520.
Lutsch, G., Noll, F., Ziska, P., Kindt, A., and Franz, H. (1984) Electron microscopic investigations on the structure of lectin I from Viscum album L. FEBS Lett., 170, 335–338.[CrossRef]
Minton, A.P. (1994) Conservation of signal: a new algorithm for the elimination of the reference concentration as an independently variable parameter in the analysis of sedimentation equilibrium. In Schuster, T.H. and Laue, T.H. (eds), Modern Analytical Ultracentrifugation. Birkhäuser, Boston, pp. 81–93.
Nicolson, G.L., Blaustein, J., and Etzler, M.E. (1974) Characterization of two plant lectins from Ricinus communis and their quantitative interaction with a murine lymphoma. Biochemistry, 13, 196–204.[CrossRef][Medline]
Niwa, H., Tonevitsky, A.G., Agapov, I.I., Saward, S., Pfuller, U., and Palmer, R.A. (2003) Crystal structure at 3 Å of mistletoe lectin I, a dimeric type-II ribosome-inactivating protein, complexed with galactose. Eur. J. Biochem., 270, 2739–2749.[ISI][Medline]
Olsnes, S., Stirpe, F., Sandvig, K., and Pihl, A. (1982) Isolation and characterization of viscumin, a toxic lectin from Viscum album L. (mistletoe). J. Biol. Chem., 257, 13263–13270.
Ralston, G. (1993) Introduction to Analytical Ultracentrifugation. Beckman Instruments, Fullerton, CA.
Rüdiger, H. and Gabius, H.-J. (2001) Plant lectins. Glycoconj. J., 18, 589–613.[CrossRef][ISI][Medline]
Rüdiger, H., Siebert, H.-C., Solís, D., Jiménez-Barbero, J., Romero, A., von der Lieth, C.-W., Diaz-Mauriño, T., and Gabius, H.-J. (2000) Medicinal chemistry based on the sugar code: fundamentals of lectinology and experimental strategies with lectins as target. Curr. Med. Chem., 7, 389–416.[ISI][Medline]
Sandvig, K., Olsnes, S., and Pihl, A. (1978) Chemical modifications of the toxic lectins abrin and ricin. Eur. J. Biochem., 84, 323–331.[ISI][Medline]
Schuck, P. (2000) Size-distribution analysis of macromolecules by sedimentation velocity ultracentrifugation and Lamm equation modeling. Biophys. J., 78, 1606–1619.
Solís, D., Jiménez-Barbero, J., Kaltner, H., Romero, A., Siebert, H.-C., von der Lieth, C.-W., and Gabius, H.-J. (2001) Towards defining the role of glycans as hardware in information storage and transfer: basic principles, experimental approaches and recent progress. Cells Tissues Organs, 168, 5–23.[CrossRef][ISI][Medline]
Stirpe, F. (2004) Ribosome-inactivating proteins. Toxicon, 44, 371–382.[Medline]
Stoeva, S., Maier, T., Soler, M.H., and Voelter, W. (1999) Carbohydrate chains and their binding sites in mistletoe lectin I. Pol. J. Chem., 73, 125–133.
Sweeney, E.C., Tonevitsky, A.G., Temiakov, D.E., Agapov, I.I., Saward, S., and Palmer, R.A. (1997) Preliminary crystallographic characterization of ricin agglutinin. Proteins, 28, 586–589.[CrossRef][ISI][Medline]
Sweeney, E.C., Tonevitsky, A.G., Palmer, R.A., Niwa, H., Pfüller, U., Eck, J., Lentzen, H., Agapov, I.I., and Kirpichnikov, M.P. (1998) Mistletoe lectin I forms a double trefoil structure. FEBS Lett., 431, 367–370.[CrossRef][ISI][Medline]
Timoshenko, A.V. and Gabius, H.-J. (1993) Efficient induction of superoxide release from human neutrophils by the galactoside-specific lectin from Viscum album. Biol. Chem. Hoppe Seyler, 374, 237–243.[ISI][Medline]
Timoshenko, A.V., Gorudko, I.V., Kaltner, H., and Gabius, H.-J. (1999) Dissection of the impact of various intracellular signaling pathways on stable cell aggregate formation of rat thymocytes after initial lectin-dependent cell association using a plant lectin as model and target selective inhibitiors. Mol. Cell. Biochem., 197, 137–145.[CrossRef][ISI][Medline]
Timoshenko, A.V., Lan, Y., Gabius, H.-J., and Lala, P.K. (2001) Immunotherapy of C3H/HeJ mammary adenocarcinoma with interleukin-2, mistletoe lectin or their combination: effects on tumor growth, capillary leakage and nitric oxide (NO) production. Eur. J. Cancer, 37, 1910–1920.[CrossRef][ISI][Medline]
Villalobo, A., Nogales-Gonzalez, A., and Gabius, H.-J. (in press) A guide to signaling pathways connecting protein-carbohydrate interaction with the emerging versatile effector functionality of mammalian lectins. Trends Glycosci. Glycotechnol.
Wu, A.M., Song, S.-C., Hwang, P.-Y., Wu, J.H., and Pfüller, U. (1995) Interaction of mistletoe toxic lectin-I with sialoglycoproteins. Biochem. Biophys. Res. Commun., 214, 396–402.[CrossRef][ISI][Medline]