The effect of mutations surrounding and within the active site on the catalytic activity of ricin A chain Catherine J. Marsden, Vilmos Fu¨lo¨ p, Philip J. Day* and J. Michael Lord Department of Biological Sciences, University of Warwick, Coventry, UK Models for the binding of the sarcin–ricin loop (SRL) of 28S ribosomal RNA to ricin A chain (RTA) suggest that several surface exposed arginine residues surrounding the active site cleft make important interactions with the RNA substrate. The data presented in this study suggest differing roles for these arginyl residues. Substitution of Arg48 or Arg213 with Ala lowered the activity of RTA 10-fold. Furthermore, substitution of Arg213 with Asp lowered the activity of RTA 100-fold. The crystal structure of this RTA variant showed it to have an unaltered tertiary structure, suggesting that the positively charged state of Arg213 is crucial for activity. Substitution of Arg258 with Ala had no effect on activity, although substitution with Asp lowered activity 10-fold. Substitution of Arg134 prevented expression of folded pro- tein, suggesting a structural role for this residue. Several models have been proposed for the binding of the SRL to the active site of RTA in which the principal difference lies in the conformation of the second ÔGÕ in the target GAGA motif in the 28S rRNA substrate. In one model, the sidechain of Asn122 is proposed to make interactions with this G, whereas another model proposes interactions with Asp75 and Asn78. Site-directed mutagenesis of these residues of RTA favours the first of these models, as substitution of Asn78 with Ser yielded an RTA variant whose activity was essentially wild-type, whereas substitution of Asn122 reduced activity 37.5-fold. Substitution of Asp75 failed to yield significant folded protein, suggesting a structural role for this residue. Keywords: ricin; ribosome; site-directed mutagenesis; heterologous expression. Ricin is a potent toxin found in the seeds of the Ricinus communis plant. It is a member of a large family of ribosome-inactivating proteins (RIPs) that exist in various tissues of many plants, fungi and bacteria [1]. Ricin consists of a catalytic A chain (RTA) joined by a single disulphide bond to a lectin B chain (RTB) that facilitates both cell surface binding and entry of the catalytically active A chain into the cytoplasm [2]. RTA inactivates eukaryotic ribosomes by catalysing the depurination of a specific adenine residue of 28S rRNA, rendering the ribosome unable to bind elongation factors and hence abolishing protein synthesis [3]. The site of depurination by RTA lies within a highly conserved purine-rich region of 28S rRNA termed the sarcin–ricin loop. Both the NMR structure [4,5] and X-ray structure [6] of an oligoribonucleotide that mimics this region show it to be a compact structure with a stem and four base loop, GAGA, at its centre. Each of the four nucleotides in this GAGA tetraloop are necessary for the action of RTA, thus, implying that the N-glycosidase activity of RTA not only requires specific interactions with the target adenine but also direct recognition of other bases in the tetra- nucleotide [7]. Unlike this tetranucleotide, the sequence of the stem does not appear to be important as long as it is at least three base-pairs in length [8]. Interactions between RTA and the rRNA backbone in this area might be required to maintain the tetraloop in the optimum conformation for catalysis. Both the catalytic role of RTA and its recognition and binding of the substrate RNA have been investigated by chemical modification [9], X-ray crystallography [10–12] and site-directed mutagen- esis [13–17]. Chemical modification showed RTA to be inactivated by treatment with the arginyl-specific reagent, phenylglyoxal [9], although it is likely that the inactivation observed in this study was primarily due to modification of the crucial catalytic residue, Arg180 [18]. Crystal structures of small molecules bound in the active site of RTA have been solved [10,12,19], but to date there are no structures of complexes of RTA with larger substrate analogues. The structure of the ribosome and of small RNA oligonucleotides show that the target adenine is not in a conformation that is compatible with binding to RTA [6]. Monzingo and Robertus [10] generated three models of complexes of a hexanucleotide (C 1 G 2 A 3 G 4 A 5 G 6 ,whereA 3 is the target for depurination by RTA) and RTA. The Correspondence to J. M. Lord, Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK. Fax: + 44 2476523701, Tel.: + 442476523598, E-mail: mlord@bio.warwick.ac.uk Abbreviations: ApG, adenyl (3¢fi5¢) guanosine; DMEM, Dulbecco’s modified Eagles medium; FMP, formycin monophosphate; RIP, ribosome-inactivating protein; RTA, ricin A chain; RTB, ricin B chain; SRL, sarcin-ricin loop. *Present address: Astex Technology, 436 Cambridge Science Park, Milton Road, Cambridge, CB4 0QA, UK. (Received 25 September 2003, revised 28 October 2003, accepted 10 November 2003) Eur. J. Biochem. 271, 153–162 (2004) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03914.x models were based on the tetraloop structure of such a hexanucleotide, where the conformation of the target adenine had been altered, using the known structures of complexes of RTA with FMP (formycin monophosphate) and ApG (adenyl 3¢fi5¢-guanosine), to facilitate docking intotheactivesiteofRTA. Here, two of the models proposed [10] have been examined (Fig. 1). The third model in which the hexanucle- otide is bound to the active site of RTA in an open conformation is not included in this study as, although several favourable interactions between the hexanucelotide and RTA are maintained, many additional interactions are proposed that involve poorly conserved residues. Each of the two models tested proposed that several arginyl residues were involved in electrostatic interactions with the phos- phdiester backbone of the hexanucleotide. The role of four such arginyl residues (Arg48, Arg134, Arg213 and Arg258) has been examined here. The most significant difference between the two models is the identity of the residues involved in interactions made between RTA and base G 4 of the hexanucleotide (Fig. 1). In this study those amino acid residues (Asp75, Asn78 and Asn122) around the active site cleft of RTA that have a putative role in the binding of G 4 in each of the two models have been examined. Experimental procedures Materials ( 35 S)Methionine was from Amersham, RTB was from Vector Laboratories (Peterborough, UK) and microbridges were from Crystal Microsystems (Oxford, UK). R258 R213 R180 N78 D75 E177 R134 Y123 Y80 N122 R258 R213 R180 N78 E177 D75 R134 Y123 Y80 N122 R258 R213 R180 N78 D75 E177 Y123 Y80 N122 R134 R258 R213 R180 N78 E177 D75 Y123 Y80 R134 N122 A B Fig. 1. Models of hexanucleotide binding in the active site of RTA (based upon [10]). The structures are shown as stereo images with the C 1 G 2 A 3 G 4 A 5 G 6 (where A 3 is the target for depurination by RTA) in red. C 1 of the hexanucleotide is at the bottom left and the target adenine is at the top of each model. The sidechains of RTA are shown in blue. (A) Model 1 has the tetraloop bound in the active site of RTA with G 4 stacked upon the G 2 -A 5 pair and able to make interactions with Asn122. (B) Model 2 is a variation of model 1 with the tetraloop bound in a conformation where G 4 stacks with Tyr80 and makes interactions with Asp75 and Asn78. Drawn with MOLSCRIPT [34,35]. 154 C. J. Marsden et al. (Eur. J. Biochem. 271) Ó FEBS 2003 Creation of ricin A chain variants Ricin A chain variants containing single amino acid sub- stitutions were generated by PCR mutagenesis and standard recombinant DNA techniques. Using the RTA expression plasmid pUTA [14] as a PCR template, appropriate base changes were introduced to encode for the single amino acid substitutions R48A, R134A, R134Q, R213A, R213D, R258A, R258D, N78S, D75A, D75S, D75N, N122A. All substitutions were confirmed by DNA sequencing. Expression and purification of ricin A chain variants A single colony of Escherichia coli JM101 transformed with the pUTA vector [14] containing the appropriate RTA- variant sequence was used to inoculate 50 mL of 2YT and grown overnight at 37 °C. This starter culture was used to inoculate 500 mL of 2YT, and the culture was grown for 2 h at 30 °C. Expression was induced by adding isopropyl thio- b- D -galactoside to a final concentration of 0.1 m M for 4 h at 30 °C. Cells were harvested by centrifugation at 2740 g, resuspended in 15 mL of 5 m M sodium phosphate buffer (pH 6.5), and lysed by sonication on ice. Cell debris was pelleted by centrifugation at 31 400 g at 4 °Cfor30min and the supernatant loaded onto a 50 mL CM-Sepharose CL-6B column (Amersham Biosciences). The column was washed with 1L of 5m M sodium phosphate, pH 6.5 followed by 100 mL of 100 m M NaCl in 5 m M sodium phosphate, pH 6.5 and RTA was eluted with a linear gradient of 100–300 m M NaCl in the same buffer. Fractions containing RTA were pooled and stored at 4 °Cata concentration of no more than 1 mgÆmL )1 . Typical yields of purified wild-type RTA and RTA variants were between 10 and 12 mgÆL )1 unless otherwise stated in the text. Crystallization, X-ray data collection and refinement of ricin A chain variants Crystals were grown in the tetragonal space group P4 1 2 1 2 by the sitting-drop method using microbridges (Crystal Microsystems, UK) and the conditions described for wild- type RTA crystallization [11]. Data were collected at 100 K and processed using the HKL suite of programs [20]. Refinement of the structures was carried out by alternate cycles of REFMAC [21] and manual refitting using O [22], based on the 1.8 A ˚ resolution model of wild-type RTA [11] (Protein Data Bank code 1ift). Water molecules were added to the atomic model automatically using ARP [23] at the positions of large positive peaks in the difference electron density, only at places where the resulting water molecule fell into an appropriate hydrogen bonding environment. Restrained isotropic temperature factor refinements were carried out for each individual atom. Data collection and refinement statistics are given in Table 1. Assay of the N -glycosidase activity of ricin A chain variants The activity of each of the RTA variants was determined by assessing their ability to depurinate 26S rRNA of Table 1. Data collection and refinement statistics. Numbers in parentheses refer to values in the highest resolution shell. R213D N122A Data collection Radiation, detector In-house CuKa MAXLAB BL-I711 and wavelength (A ˚ ) DIP2030, 1.54184 MAR IP, 1.0213 Unit cell dimensions (A ˚ ) a ¼ b ¼ 67.3, c ¼ 140.7 a ¼ b ¼ 67.5, c ¼ 140.6 Resolution (A ˚ ) 28 - 1.9 (1.949 - 1.9) 46 – 1.4 (1.436–1.4) Observations 53 233 274 010 Unique reflections 22 494 60 829 I/r(I) 21.6 (4.7) 43.2 (8.1) R sym a 0.038 (0.133) 0.031 (0.121) Completeness (%) 85.6 (43.6) 93.9 (99.3) Refinement Non-hydrogen atoms 2492 (including 2 sulphate 406 water molecules) 2596 (including 2 sulphate, 1 acetate and 510 water molecules) R cryst b 0.154 (0.154) 0.177 (0.176) Reflections used 21 573 (720) 58 381 (4,486) R free c 0.231 (0.257) 0.214 (0.192) Reflections used 921 (24) 2,448 (188) R cryst (all data) b 0.157 0.179 Mean temperature factor (A ˚ 2 ) 24.8 19.8 Rmsds from ideal values Bonds (A ˚ ) 0.021 0.059 Angles ( o ) 1.9 3.0 DPI coordinate error (A ˚ ) 0.145 0.060 PDB accession codes 1uq4, r1uq4sf 1uq5, r1uq5sf a R sym ¼ S j S h |I h,j –<I h >|/S j S h <I h > where I h,j is the jth observation of reflection h, and <I h > is the mean intensity of that reflection. b R cryst ¼ S||F obs |–|F calc ||/S|F obs | where F obs and F calc are the observed and calculated structure factor amplitudes, respectively. c R free is equivalent to R cryst for a 4 % subset of reflections not used in the refinement [24]. Ó FEBS 2003 Mutations affecting the activity of ricin A chain (Eur. J. Biochem. 271) 155 purified yeast (Saccharomyces cerevisiae) ribosomes. For each reaction, 20 lg of yeast ribosomes were incubated at 30 °C for 1 h with the relevant RTA variant in 25 m M Tris/HCl (pH 7.6), 25 m M KCl, 5 m M MgCl 2 , in a total volume of 20 lL. Reactions were stopped by the addition of 100 lLof2· Kirby buffer [25] and 80 lLofH 2 O, and rRNA was obtained by precipitation after two phenol– chloroform extractions. Four micrograms of rRNA were treatedwith20lL of acetic-aniline for 2 min at 60 °C, and rRNA was precipitated and resuspending in 15 lLof 60% de-ionized formamide/0.1· TPE (3.6 m M Tris, 3 m M NaH 2 PO 4 ,0.2m M EDTA) and heated at 65 °Cfor 5 min. Ribosomal RNA fragments were separated on a 1.2% agarose, 0.1· TPE, 50% (v/v) formamide gel. rRNAs were quantified from digital images of ethidium bromide-stained gels, using IMAGEQUANT software, and depurination was calculated by relating the amounts of the small aniline-fragment and 5.8S rRNA and expressing values as a percentage. Reassociation and quantification of ricin A-chain variants Purified RTA (100 lg) was mixed with 100 lgofRTB (Vector Laboratories) and made up to a final volume of 2 mL with NaCl/P i containing 0.1 M lactose and 2% (v/v) 2-mercaptoethanol. This was dialysed for 24 h against 1 L of NaCl/P i containing 0.1 M lactose, followed by a further 36 h against 5 L of NaCl/P i . Reassociated holotoxin was separated from free RTA on a 0.5 mL immobilized a-lactose column. The dialysate was loaded onto the column three times and the column was then washed with 10 mL of NaCl/P i before eluting bound holotoxin (and free RTB) in 5 mL of NaCl/P i containing 75 m M galactose. Eluted protein was dialysed for 16 h against 1L of NaCl/P i to remove the galactose, before quanti- fying against know quantities of RTA on a silver-stained SDS/polyacrylamide gel using Molecular Dynamics IMAGEQUANT version 3.3. Cytotoxicity assay Cells were plated out in a volume of 100 lLin96-wellplates at a density of 1.5 · 10 5 cellsÆmL )1 and incubated for 18 h at 37 °C. Toxin dilutions (100 lL) in DMEM were added in quadruplicate and the plates were incubated for 4 h at 37 °C. Protein synthesis was measured by incubating the plates for 90 min at 37 °C in the presence of 1 lCi of ( 35 S)methionine in 50 lLofNaCl/P i per well. Proteins were precipitated by washing three times with ice-cold 5% (v/v) trichloroace- tic acid and, after the addition of 200 lL of scintillant (OptiPhase ÔSupermixÕ) to each well, plates were counted in a Wallac 1450 MicroBeta Trilux liquid scintillation counter. Results N -glycosidase activity of RTA variants containing arginine substitutions RTA variants in which arginyl residues 213 or 258 were substituted with Ala, or Asp or in which Arg48 had been substituted with Ala, expressed to levels equivalent to wild- type RTA and were readily purified to homogeneity. Each of these RTA variants had the same stability to digestion by trypsin as wild-type RTA (data not shown). Substitu- tion of Arg134 with either Ala or Gln resulted in barely detectable expression levels and, as such, these mutants could not be purified. To assess the effect of substitutions made at each of the arginyl residues on catalytic activity of RTA, the ability of each of the purified RTA variants to depurinate yeast ribosomes was compared to that of wild- type RTA. Conversion of Arg213 to either alanine or aspartate (R213A and R213D, respectively) reduced the D 50 (concentration giving half maximal depurination) by 10-fold and over 100-fold, respectively (Fig. 2A). Substitu- tion of Arg258 with Ala (R258A) had no effect on activity whereas substitution of Arg258 with Asp (R258D) lowered the activity niinefold (Fig. 2B). Finally, substitution of Arg48 with Ala (R48A) lowered the activity by 10-fold (Fig. 2C). Cytotoxicity of RTA variants containing arginine substitutions Each of the RTA variants described above were reassoci- ated with RTB and the cytotoxicity of each of the resultant ricin variants was compared to wild-type ricin. Ricin R213A and ricin R213D were 12-fold and more than 2000- fold less cytotoxic than ricin, respectively (Fig. 3A). Ricin R258A was equally as cytotoxic as ricin (Fig. 3B) whereas ricin R258D (Fig. 3B) and ricin R48A (Fig. 3C) were 10-fold less cytotoxic. The reductions in cytotoxcity compared to native ricin were comparable to the decrease in catalytic activity against ribosomes in vitro for each of the substitutions except for ricin R213D, whose cytotoxi- city was reduced by 20-fold more than the catalytic activity of RTA R213D. The X-ray structure of RTA R213D In the absence of structural data for each of the RTA variants discussed, it is possible that the reduction in activity might be attributed to structural changes of the enzyme. The most substantial decrease in catalytic activity was seen when a substitution was made at Arg213 (R213D). The crystal structure was solved to determine whether the reduction in activity of this RTA variant could be attributed solely to the change in charge and size of the sidechain of this single residue. The structure of RTA R213D is essentially identical to that of recombinant wild-type RTA with an root mean square deviation (RMSD) from the Ca atoms of the wild- type crystal structure [11] of 0.33 A ˚ . The electron density in the area local to the substitution is shown in Fig. 4. The positions of catalytic residues and all other residues local to the substitution site do not differ significantly from the wild- type crystal structure. Binding of the GAGA tetraloop to RTA In order to better understand the specific interactions that RTA makes with the GAGA tetraloop, the models of RTA–substrate interactions proposed by Monzingo and Robertus [10] have been examined. The first model examined here has the sequence C 1 G 2 A 3 G 4 A 5 G 6 (where A 3 156 C. J. Marsden et al. (Eur. J. Biochem. 271) Ó FEBS 2003 is the target for depurination by RTA) modelled into the active site with the first base C 1 of the hexanucleotide forming a Watson–Crick pair with the last base G 6 .This both closes the tetraloop and allows G 2 and A 5 to be stacked upon this base pair forming a non-Watson–Crick pairing (Fig. 1A). The target adenine residue A 3 projects out of the tetraloop and is positioned in the same location as indicated by the crystal structures of both the FMP and ApG complexes [10], forming stacking interactions with the aromatic groups of Tyr80 and Tyr123. The first model has G 4 stackedontheG 2 -A 5 pair of the hexanucelotide allowing two hydrogen bonds to be formed between the base G 4 and the sidechain of Asn122. In the second model (Fig. 1B) the structure of the hexanucelotide is, on the whole, unchanged and the majority of the interactions that were seen in the first model are maintained. However, the interaction between G 4 andAsn122cannolongerbemade as G 4 is in an altered position making a continuous curved stack of aromatic groups, Tyr123, A 3 , Tyr80, G 4 .Inthis Fig. 3. Dose dependent cytotoxicity assay of ricin variants, ricin R213A and ricin R213D towards Vero cells. Vero cells were challenged with increasing concentrations of toxin at 37 °C for 4 h. Remaining protein synthesis after this time was measured by the incorporation of ( 35 S)methionine. Symbols indicate the mean of four replicate samples, error bars represent the SD. (A) ricin, d, ricin R213A, j;andricin R213D, m.(B)ricin,d; ricin R258A, j, and ricin R258D, m;(C)ricin, d; ricin R48A, j. Fig. 2. Assessment of the N-glycosidase activity of RTA variants con- taining arginine substitutions. Isolated yeast ribosomes (20 lg) were incubated with RTA or RTA variants for 60 min at 30 °C. rRNA was isolated and 4 lg was aniline-treated and electrophoresed on an ag- arose/formamide gel. rRNAs were quantified from digital images using IMAGEQUANT software. The depurination was calculated by relating the amounts of small diagnostic aniline-fragment [3] and 5.8S rRNA and expressing values as a percentage. Symbols indicate the experi- mental data, error bars represent the SD and solid lines represent best- fitted curves. (A) RTA, d; RTA R213A, j, and RTA R213D, m.(B) RTA, d; RTA R258A j, and RTA R258D, m; (C) RTA, d;RTA R48A, j. Ó FEBS 2003 Mutations affecting the activity of ricin A chain (Eur. J. Biochem. 271) 157 model alone G 4 is proposed to make two hydrogen bonds with Asn78 and a further hydrogen bond with Asp75. In order to test these models, substitutions were made at each of the residues proposed to be involved – Asp75, Asn78 and Asn122. N -glycosidase activity of RTA variants with substitutions at Asp75 and Asn78 Substitution of either Asp75 or Asn78 with Ala resulted in very low expression levels and purification of these variants was not achieved. However, although a very low expression level was again seen on substitution of Asp75 to Ser, substitution of Asn78 to Ser (RTA N78S) produced a protein that expressed as well as wild-type RTA and had equal stability to digestion by trypsin as wild-type RTA (data not shown). Further substitutions were made at Asp75 to Asp, Arg and Gln. However, all were expressed at very low levels, and purification to homogeneity, to allow quantitative activity assays of these RTA variants, was not achieved. To assess whether substitution at Asn78 with Ser changed the catalytic activity of RTA, the N-glycosidase activity of RTA N78S against yeast ribosomes was determined and compared to that of wild-type RTA (Fig. 5A): there was less than a twofold reduction in activity between them. Cytotoxicity of RTA N78S RTA N78S was reassociated with RTB and its cytotoxicity was compared to ricin (Fig. 5B). Ricin containing RTA N78S was approximately twofold less cytotoxic than wild- type ricin. This difference in cytotoxicity was comparable to the small reduction in N-glycosidase activity seen for RTA-N78. N -glycosidase activity of RTA variants with substitutions at Asn122 Conversion of Asn122 to Ala (RTA N122A) produced an RTA variant that expressed to high levels (equivalent to wild-type RTA). RTA N122A had equal stability, based on sensitivity to trypsin digestion, as the wild-type protein (data not shown) and was readily purified to homogeneity. To assess whether the substitution at Asn122 for Ala had caused any change in the catalytic activity of RTA, the N-glycosidase activity of RTA N122A against yeast ribo- somes was determined and compared to that of wild-type RTA (Fig. 6A). The reduction in activity of this RTA N122A was 37.5-fold. Cytotoxicity of RTA N122A RTAN122AwasreassociatedwithRTBanditscytotoxi- city was compared to ricin (Fig. 6B). Substituting Ala for Asp at residue 122 in RTA reduced the cytotoxicity of RTA by 30-fold (correlating well with the in vitro N-glycosidase activity). The X-ray structure of RTA N122A In order to confirm that the reduction in catalytic activity of RTA N122A could be attributed solely to the change in charge and size of the sidechain of this single residue, the crystal structure of RTA N122A was solved. The RMSD from the C a atoms of the wild-type RTA crystal structure [11] was 0.33 A ˚ indicating that the structure of RTA N122A is essentially identical to recombinant wild-type RTA. The electron density in the area of the substitution is shown in Fig. 7. The positions of catalytic residues and all other residues local to the substitution site do not differ from the wild-type crystal structure. Discussion When the structure of ricin A-chain was solved to 2.5 A ˚ [26], it was suggested that a number of arginyl residues were likely to be responsible for the binding of the rRNA substrate. Several of these arginyl residues have subse- quently been proposed to form an arginine-rich binding motif [27]. Furthermore, when RTA is treated with the arginyl specific reagent, phenylglyoxal, it is readily inacti- vated leading to the proposal that this inactivation R213D R213D Fig. 4. Electron density of RTA R213D in the vicinity of residue 213. The backbone and sidechains of the R213D substitution are shown as stereo images in thick ball and stick and the position of the Arg213 side-chain of the wild-type enzyme is overlayed and shown in thin ball and stick. The SIGMAA [33] weighted 2mF o -DF c electron density using phases from the final model is contoured at 1 r level, where r represents the rms electron density for the unit cell. Contours more than 1.4 A ˚ from any of the displayed atoms have been removed for clarity. Drawn with MOLSCRIPT [34,35]. 158 C. J. Marsden et al. (Eur. J. Biochem. 271) Ó FEBS 2003 involved the modification of arginyl residues 196, 213, 234 and 235 [9]. Of theses residues, only Arg213 is in the vicinity of the active site of RTA. In order to establish the role of arginyl residues around the active site of RTA in the binding of rRNA, a number of RTA variants have been constructed. Four such arginyl residues were selected as targets for site-directed mutagenesis. We appreciate that the mutagenic approach we have taken does not distin- guish between a role for particular residues in substrate binding or the catalytic reaction itself. Both roles are required for RTA catalysis and hence for the cytoxicity of ricin. However, it does not seem unreasonable to assume, at least as a broad generalization, that residues lying within the active site cleft are involved in the reaction catalysed, while those positively charged residues surrounding but outside the active site are probably involved in binding the negatively charged RNA substrate. Arg48 is a variable residue that lies in a loop on the edge of the active site cleft of RTA that has been implicated in substrate binding by modelling studies [28]. Substitution of this RTA residue with alanine reduced catalytic activity Fig. 5. Assessment of the N-glycosidase activity and cytotoxicity of RTA N78S and ricin N78S. (A) N-glycosidase activity of RTA N78S. Isolated yeast ribosomes (20 lg) were incubated with RTA (d)or RTA N78S (j)for60minat30 °C. rRNA was isolated and 4 lgwas aniline-treated and electrophoresed on an agarose/formamide gel. rRNAs were quantified from digital images using IMAGEQUANT soft- ware. The depurination was calculated by relating the amounts of small aniline-fragment and 5.8S rRNA and expressing values as a percentage. Symbols indicate the experimental data, error bars repre- sent the SD and solid lines represent best-fitted curves. (B) Dose dependent cytotoxicity assay of ricin N78S. Vero cells were challenged with increasing concentrations of ricin (d),orricinN78S(j)at37°C for 4 h. Remaining protein synthesis after this time was measured by the incorporation of ( 35 S)methionine. Symbols indicate the mean of four replicate samples, error bars represent the SD. Fig. 6. Assessment of the N-glycosidase activity and cytotoxicity of RTA N122A and ricin N122A. (A) N-glycosidase activity of RTA N122A. Isolated yeast ribosomes (20 lg) were incubated with RTA (d)orRTAN122A(j)for60minat30°C. rRNA was isolated and 4 lg was aniline-treated and electrophoresed on an agarose/forma- mide gel. rRNAs were quantified from digital images using IMAGE- QUANT software. The depurination was calculated by relating the amounts of small aniline-fragment and 5.8S rRNA and expressing values as a percentage. Symbols indicate the experimental data, error bars represent the SD and solid lines represent best-fitted curves. (B) Dose dependent cytotoxicity assay of ricin N122A. Vero cells were challenged with increasing concentrations of ricin (d) or ricin N122A (j)at37°C for 4 h. Remaining protein synthesis after this time was measured by the incorporation of ( 35 S)methionine. Symbols indicate the mean of four replicate samples, error bars represent the SD. Ó FEBS 2003 Mutations affecting the activity of ricin A chain (Eur. J. Biochem. 271) 159 10-fold, consistent with the role predicted for this residue by modelling, but in contrast to an earlier study [29]. The apparent discrepancy with this earlier study is probably due to the type of the assays used. The earlier study [29] used single point assays which may not have been sufficiently sensitive to observe a 10-fold decrease in activity that is readily observed in the dose–response assay used here. Arg258 is another variable residue that Monzingo and Robertus [10] suggested might form an ion-pair with the phosphodiester backbone of rRNA, specifically with that of the second guanine in the GAGA motif. Olson and Cuff [28] also proposed that this residue was involved in substrate binding, interacting with the loop-closing guanine base (the first G after the GAGA) and, due to it’s variable nature, could perhaps be responsible for the differing binding affinities of RIPs for their substrates. Substitution of Arg258 with Ala had no effect on activity, implying that this residue is not necessary for the activity of RTA. This is supported by the earlier observation showing that deletion of residues 258–262 did not inactivate RTA [30]. However, although this residue is not essential for activity, reversing the charge of this sidechain resulted in a 10-fold reduction in activity, consistent with this residue being close to the phosphodi- ester backbone of the substrate but not making an essential interaction with it. Arg134 is highly conserved in RIPs, and it has been proposed that it could make a forked interaction with both of the phosphodiester backbones of the target adenine and of the preceding guanine residue in the GAGA motif [10]. Substitution of Arg134 with Ala or Gln abolished expres- sion of folded protein, and consequently it was concluded that Arg134 played a critical structural role. Although Arg134 is predicted to be able form an ionic interaction with the phosphodiester backbone of the substrate, the structure of RTA also shows that this residue is at the centre of a hydrogen bonding network, forming H-bonds with Glu127, Asn209 and Glu208. Arg134 also forms a p- stacking interaction with Tyr123, a residue directly involved in substrate binding. Substitution of residues Asn209, Glu208 and Tyr123 permitted RTA folding and reduced activity between threefold and 10-fold, suggesting that they form only modest interactions with the substrate [14,18]. Thus, while it remains possible that Arg134 makes a significant contribution to substrate binding, this can not be readily tested by site-directed mutagenesis due to its structural role. Arg213 is a weakly conserved residue, the amino acyl residue at this position being positively charged in nearly 60% of ribosome-inactivating proteins [28]. The Monzingo and Robertus models [10], and the 29mer oligonucleotide-binding model of Olson and Cuff [28], show this residue forming an ion-pair with the phosphodiester backbone of the first cytosine residue in the CGAGAG tetraloop motif. Furthermore, an RTA mutant in which Arg213 had been deleted was found to be inactive [31]. Substitution of Arg213 with Ala reduced activity 10-fold against both purified ribosomes and whole cells, while reversal of the charge at this site reduced activity a further 10-fold. This additive effect of changing the charge of residue 213 strongly suggests that it is the charged nature of Arg213 that is responsible for effects on enzyme activity rather than any structural changes induce by the substitutions. This is confirmed by the finding that the structure of RTA R213D is identical to that of the wild-type enzyme except for the sidechain of residue 213. Thus, Arg213 forms a significant electrostatic interaction with the substrate ribosome, prob- ably via interaction with the phosphodiester backbone. Individual substitution of arginyl residues around the active site cleft of RTA resulted in different effects on the activity of RTA. Whereas, one had no effect on the activity of the enzyme (R258A), others reduced activity by one (R213A, R258D, R48A) or by two (R213D) orders of N122A Y80 N122A Y80 Fig. 7. Electron density of RTA N122A in the vicinity of residue 122. The backbone and sidechains of the N122A substitution are shown as stereo images in thick ball and stick and the position of the Asn122 side-chain of the wild-type enzyme is overlayed and shown in thin ball and stick. The SIGMAA [33] weighted 2mF o -DF c electron density using phases from the final model is contoured at 1 r level, where r represents the rms electron density for the unit cell. Contours more than 1.4 A ˚ from any of the displayed atoms have been removed for clarity. Drawn with MOLSCRIPT [34,35]. 160 C. J. Marsden et al. (Eur. J. Biochem. 271) Ó FEBS 2003 magnitude. It then appears that, in addition to catalytic site residues, a number of the Arg residues located around the active site cleft play a significant role in the activity of RTA. As none of these arginyl residues have been implicated in catalysis, the reduction in activity of each of these RTA variants is probably due to a loss of electrostatic interactions that are critical for the stable binding of the substrate RNA molecule in the active site pocket. Although the contribution of each individual residue is relatively small, the cumulative effect of these residues would make a substantial contribu- tion to substrate binding. This contribution is probably relatively nonspecific in nature such that these residues probably make a major contribution to binding, but do not greatly contribute to specificity. Further interactions of the substrate RNA with RTA were studied by modelling a hexanucleotide into the active site [10] whose structure was based upon the crystal structure of a GNRA loop which had been solved [19]. A series of RTA variants were made based on two of the models proposed by Monzingo and Robertus [10]. Site- directed mutagenesis was used to make RTA variants with substitutions at either Asp75 or Asn78, both of which are highly conserved in the RIP family and were proposed to make interactions with the second G 4 in the GAGA tetraloop in the second of the models (Fig. 1B). Whereas none of the substitutions to Ala, Ser, Asp, or Arg were tolerated at Asp75 implying that this residue might play an important structural role, RTA N78S had the same catalytic activity as wild-type RTA suggesting that this residue may not play a significant role in substrate binding, although it remains possible that a substituted serine could still make the hydrogen bond normally made by Asn78. In the first model (Fig. 1A) G 4 is proposed to interact with Asn122 and, in agreement with this, substitution of Ala for Asp at this site was shown to lower the N-glycosidase activity 37.5-fold, and cytotoxicity of the reassociated holotoxin by 30-fold, without affecting the overall structure or the position of either active site residues or residues in the vicinity of the substitution. That the RTA N78S variant had no effect on activity, but the RTA N122A variant reduced catalytic activity over 30-fold compared to wild-type RTA, is consistent with the first of the models proposed by Monzingo and Robertus [10]. In order to understand more fully the precise role of Asn122 in the catalytic activity of RTA, it would be of value to solve the structure of RTA N122A in the presence of small nucleotides such as ApG that can be diffused into RTA crystals. RTA cleaves a single N-glycosidic bond from among over 4000 in 28S rRNA [32]. Ultimately, complete understanding of the specificity of RTA for its ribosomal RNA substrate, and the role that recognition and binding of the RNA stem and loop might play in this, may only be fully achieved when crystal structures of complexes of RTA with much larger RNA fragments become available. Acknowledgements This work was supported by the UK Biotechnology and Biological Sciences Research Council grant 88/B16355 and Wellcome Trust Programme Grant 063058/Z/00/Z. VF is a Royal Society University Fellow. References 1. Van Damme, E.J.M., Hao, Q., Chen, Y., Barre, A., Vandenbus- sche, F., Desmyter, S., Rouge ´ , P. & Peumans, W.J. (2001) Ribo- some-inactivating proteins: a family of plant proteins that do more than inactivating ribosomes. Crit. Rev. Plant Sci. 20, 395–465. 2. Sandvig, K. & van Deurs, B. (2002) Transport of protein toxins into cells: pathways used by ricin, cholera toxin and Shiga toxin. FEBS Lett. 529, 49–53. 3. Endo, Y. & Tsurugi, K. (1987) RNA N-glycosidase activity of ricin A chain. Mechanism of action of the toxic lectin ricin on eukaryotic ribosomes. J. Biol. Chem. 262, 8128–8130. 4. Orita, M., Nishikawa, F., Shimayame, T., Taira, K., Endo, Y. & Nishikawa, S. (1993) High resolution NMR study of a synthetic oligoribonucleotide with a tetranucleotide GAGA loop that is a substrate for the cytotoxic protein, ricin. Nucleic Acids Res. 21, 5670–5678. 5. Szewczak, A.A. & Moore, Y.L. (1995) The sarcin/ricin loop, a modular RNA. J. Mol. Biol. 247, 81–98. 6. Correll, C.C., Munishkin, A., Chan, Y., Ren, Z., Wool, I.G. & Steitz, T.A. (1998) Crystal structure of the ribosomal RNA domain essential for binding elongation factors. Proc. Natl Acad. Sci. USA 95, 13436–13441. 7. Gluck, A., Endo, Y. & Wool, I.G. (1992) Ribosomal RNA identity elements for ricin A chain recognition and catalysis. Analysis with tetraloop mutants. J. Mol. Biol. 226, 411–424. 8. Wool, I.G., Gluck, A. & Endo, Y. (1992) Ribotoxin recognition of ribosomal RNA and a proposal for the mechanism of transloca- tion. Trends in Biochem. 17, 267–269. 9. Watanabe, K., Dansako, H., Asada, N., Saki, M. & Funatsu, G. (1994) Effects of chemical modification of arginine residues out- side the active site cleft of ricin A chain on its N-glycosidase activity for ribosomes. Biosci. Biotechn. Biochem. 58, 716–721. 10. Monzingo, A.F. & Robertus, J.D. (1992) X-ray analysis of sub- strate analogs in the ricin A chain active site. J. Mol. Biol. 227, 1136–1145. 11. Weston, S.A., Tucker, A.D., Thatcher, D.R., Derbyshire, D.J. & Pauptit, R.A. (1994) X-ray structure of recombinant ricin A chain at 1.8A ˚ resolution. J. Mol. Biol. 244, 410–422. 12. Yan, X., Hollis, T., Svinth, M., Day, P., Monzingo, A.F., Milne, G.W.A. & Robertus, J.D. (1997) Structure-based identification of a ricin inhibitor. J. Mol. Biol. 266, 1043–1049. 13. Schlossmann, D., Withers, D., Welsh, P., Alexander, A., Rober- tus, J. & Frankel, A. (1989) Role of glutamic acid 177 of the ricin toxin A chain in enzymatic inactivation of ribosomes. Mol. Cell. Biol. 9, 5012–5021. 14. Ready, M.P., Kim, Y. & Robertus, J.D. (1991) Site-directed mutagenesis of ricin A chain and implications for the mechanism of action. Proteins 10, 270–278. 15. Kim, Y. & Robertus, J.D. (1992) Analysis of several active site residues of ricin A chain by mutagenesis and X-ray crystal- lography. Protein Eng. 5, 775–779. 16. Chaddock, J.A. & Roberts, L.M. (1993) Mutagenesis and kinetic analysis of the active site Glu177 of ricin A chain. Protein Engineering 6, 425–431. 17. Day, P.J., Ernst, S.R., Frankel, A.E., Monzingo, A.F., Pascal, J.M., Molina-Svinth, M.C. & Robertus, J.D. (1996) Structure and activity of an active site substitution of ricin A chain. Biochemistry 35, 11098–11103. 18. Frankel,A.,Welsh,P.,Richardson,J.&Robertus,J.D.(1990) role of arginine 180 and glutamic acid 177 of ricin toxin A chainin enzymatic inactivation of ribosomes. Mol. Cell. Biol. 10, 6257– 6263. 19. Heus, H.A. & Pardi, A. (1991) Structural features that give rise to the unusual stability of RNA hairpins containing GNRA tetra- loops. Science 253, 191–194. Ó FEBS 2003 Mutations affecting the activity of ricin A chain (Eur. J. Biochem. 271) 161 20. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray diffrac- tion data collected in oscillation mode. Methods Enzymol. 276, 307–326. 21. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. (1997) Refine- ment of macromolecular structures by the maximum-likelyhood method. Acta Crystallog. Sect. D 53, 240–255. 22. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. (1991) Improved methods for binding protein models in electron density maps and the location of errors in these models. Acta Crystallog. 47, 110–119. 23. Perrakis, A., Sixma, T.K., Wilson, K.S. & Lamzin, V.S. (1997) warp: improvement and extension of crystallographic phases by weighted averaging of multiple refined dummy models. Acta Crystallogr. Sect. D 53, 448–455. 24. Bru ¨ nger, A.T. (1992) Free R. value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature 355, 472–474. 25. Kirby, K.S. (1968) Isolation of nucleic acids with phenolic sol- vents. Methods Enzymol. 12B, 87–100. 26. Katzin, B.J., Collins, E.J. & Robertus, J.D. (1991) Structure of ricinAchainat2.5A ˚ . Proteins 10, 251–259. 27. Olson, M.A. (1997) Ricin A chain structural determinant for binding substrate analogues. Proteins 27, 80–95. 28. Olson, M.A. & Cuff, L. (1999) Free-energy determinants of binding the rRNA substrate and small ligands to ricin A chain. Biophys. J. 76, 28–39. 29. May, M.J., Hartley, M.R., Roberts, L.M., Kreig, P.A., Osborn, R.W. & Lord, J.M. (1989) Ribosome inactivation by ricin A chain: a sensitive method to assess the activity of wild-type and mutant polypeptides. EMBO J. 8, 301–308. 30. Kitaoka, Y. (1988) Involvement of the amino acids outside the active site cleft in the catalysis of ricin A chain. Eur. J. Biochem. 257, 255–262. 31. Munishkin, A. & Wool, I.G. (1995) Systematic deletion analysis of ricinAchainfunction.J. Biol. Chem. 270, 30581–30587. 32. Endo, Y. & Tsurugi, K. (1988) The RNA N-glycosidase activity of ricin A chain. J. Biol. Chem. 263, 8735–8739. 33. Read, R.J. (1986) Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallog. Sect. A 42, 140–149. 34. Kraulis, P.J. (1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crys- tallog. 24, 946–950. 35. Esnouf, R.M. (1997) An extensively modified version of MolScript that includes greatly enhanced coloring capabilities. J. Mol. Graph. Model. 15, 132–134. 162 C. J. Marsden et al. (Eur. J. Biochem. 271) Ó FEBS 2003 . The effect of mutations surrounding and within the active site on the catalytic activity of ricin A chain Catherine J. Marsden, Vilmos Fu¨lo¨. To assess whether the substitution at Asn122 for Ala had caused any change in the catalytic activity of RTA, the N-glycosidase activity of RTA N12 2A against