Tài liệu Báo cáo khóa học: Quantitative analysis, using MALDI-TOF mass spectrometry, of the N-terminal hydrolysis and cyclization reactions of the activation 2 process of onconase pdf

Quantitative analysis, using MALDI-TOF mass spectrometry, of the N-terminal hydrolysis and cyclization reactions of the activation process of onconase 2 Marc Ribo ´ 1 , Montserrat Bosch 1 , Gerard Torrent 1 , Antoni Benito 1 , Bruno Beaumelle 2 and Maria Vilanova 1 1 Laboratori d’Enginyeria de Proteı ¨ nes, Departament de Biologia, Facultat de Cie ` ncies, Universitat de Girona, Girona, Spain; 2 UMR 5539 CNRS, Department Biologie-Sante ´ , Universite ´ Montpellier II, Montpellier, France Onconase, a member of the ribonuclease superfamily, is a potent cytotoxic agent that is undergoing phase II/III human clinical trials as an antitumor drug. Native onconase from Rana pipiens and its amphibian homologs have an N-ter- minal pyroglutamyl residue that is essential for obtaining fully active enzymes with their full potential as cytotoxins. When expressed cytosolically in bacteria, Onconase is isolated with an additional methionyl (Met1) residue and glutaminyl instead of a pyroglutamyl residue at position 1 of the N-terminus and is consequently inactivated. The two reactions necessary for generating the pyroglutamyl residue have been monitored by MALDI-TOF MS. Results show that hydrolysis of Met()1), catalyzed by Aeromonas aminopeptidase, is optimal 3 at a concentration of ‡ 3 M guanidinium-chloride, and at pH 8.0. The intramolecular cyclization of glutaminyl that renders the pyroglutamyl residue is not accelerated by increasing the concentration of denaturing agent or by strong acid or basic conditions. However, temperature clearly accelerates the formation of pyroglutamyl. Taken together, these results have allowed the characterization and optimization of the onconase activa- tion process. This procedure may have more general appli- cability in optimizing the removal of undesirable N-terminal methionyl residues from recombinant proteins overexpres- sed in bacteria and providing them with biological and catalytic properties identical to those of the natural enzyme. Keywords: onconase; cytotoxicity; recombinant protein activation; MALDI-TOF mass spectrometry. 4 N- or C-terminal modifications constitute post-translational modifications that can modulate a peptide activity and/or resistance to degradation, as is the case with acetylation, pyroglutamyl formation or C-terminal amidation. Many proteins and bioactive peptides exhibit an N-terminal pyroglutamyl, which subsequently minimizes their suscep- tibility to degradation by aminopeptidases, although it may also play a crucial role at the functional level [1]. This residue is also a frequent determinant of overall peptide function, as has been shown by the hypothalamic releasing factor binding to its receptor [2], or by the amyloid b-peptide and the implications in senile plaque formation and pathogenesis in Alzheimer’s disease [3]. Onconase (ONC) is a ribonuclease that is present in the oocytes and early embryos of the frog, Rana pipiens [4]. ONC, discovered as a result of its potent anticancer activity [5], is now in Phase III human clinical trials for the treatment of several types of cancer [6]. The enzyme, isolated from frog oocytes, has an N-terminal pyroglutamyl residue that contributes to the structure of its active site [7] and also to its stability [8]. This N-terminal pyroglutamyl residue is produced in vivo by the cyclization of the N-terminal glutamyl residue. Pyroglutamyl N-termini have been found in other frog ribonucleases that also display interesting cytotoxic and antitumoral properties [9]. It has been reported that non-natural N-terminal residues correlate with a decrease in the catalytic activity and cytotoxicity of these enzymes [10]. The interest in ONC as a therapeutic agent has led to the expression of ONC recombinants, created using site-specific mutagenesis, in order to study the molecular determinants of its biological action. 5 The production of unfused proteins from recombinant vectors generates a cytosolic protein with an additional methionyl residue at the N-terminus. When ONC is produced with an N-terminal methionyl residue, Met1, it retains only 2% of the ribonucleolytic activity of the native enzyme and is an ineffective cytotoxin, despite being folded properly [11]. However, cytosolic expression of the ONC in Escheri- chia coli is interesting because it produces higher yields than the alternative, secretory approach [12]. Initially, an ONC Correspondence to M. Vilanova, Laboratori d’Enginyeria de Proteı ¨ nes, Departament de Biologia, Facultat de Cie ` ncies, Universitat de Girona, Campus de Montilivi, 17071 Girona, Spain. Fax: + 34 972 418150, Tel.: + 34 972 418173, E-mail: maria.vilanova@udg.es Abbreviations: AAP, aminopeptidase of Aeromonas proteolytica; CNBr, cyanogen bromide; DMEM, Dulbecco’s modified Eagle’s medium; (Gln1)-ONC (M23L), onconase variant with a Gln1 and leucine replacing methionine at position 23; GSH, reduced gluta- thione; GSSG, oxidized glutathione; IC 50 , 50% inhibitory concentration; IPTG, isopropyl thio-b- D -galactoside; (Met1)-ONC (M23L), onconase variant with a methionine preceding Gln1; ONC, Onconase; (Pyr)-ONC (M23L), onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23; rONC, wild-type recombinant onconase. (Received 10 December 2003, revised 28 January 2004, accepted 3 February 2004) Eur. J. Biochem. 271, 1163–1171 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04020.x variant, with the change M23L, was designed as a strategy for obtaining recombinant ONC upon removal of (Met1) by treatment with cyanogen bromide (CNBr) [13]. Notomista and co-workers [12] showed that M23L substitution increased the catalytic activity of this variant relative to the native enzyme, but did not modify its potential as a cytotoxin. Alternatively, the exopeptidase Aeromonas amino- proteolytica aminopeptidase (AAP) [14] has proved to be useful in removing Met1 from other ribonucleases, such as angiogenin [15] and bovine seminal ribonuclease [16]. When using (Met1)-ONC (M23L) as a substrate, it was found necessary to completely denature and reduce the protein in order to efficiently remove the methionyl residue [12]. Although, unlike CNBr treatment, the procedure did not generate secondary products, it was very time consuming. 6 In this work we have characterized, using MALDI-TOF MS, the two reactions involved in the activation process of (Met1)-ONC (M23L) and established a rapid procedure to activate the enzyme, without the need for reducing it once purified. By using the purified recombinant protein, the conversion of reagents and products of the exopeptidase- catalyzed hydrolysis of Met1 and cyclization of Gln1 to pyroglutamyl were monitored at different time-points during the reaction by the acquisition of mass spectra. This enabled us to calculate the fraction of each species present at a particular instant and under particular conditions. This strategy was used to characterize the factors that affect the reactions, as well as those responsible for optimizing the process to produce high yields of fully active ONC. This experimental design may also be extended to other protein systems where the undesired presence of additional residues has a negative effect on their activity, modifies their antigenicity or affects their conformational stability 7 . Materials and methods Materials Recombinant (Met1)-ONC (M23L) was expressed in E. coli BL21(DE3) cells harboring pET11d-(Met1)-ONC (M23L), and rONC was produced from E. coli BL21(DE3) cells transformed with pONC. Both vectors were kindly provi- ded by Dr R. T. Raines (University of Madison, Madison, WI, USA), and were as described previously [13]. RNase A was produced as described previously [17]. The E. coli strain, BL21(DE3), was obtained from Novagen (Madison, WI, USA), the Mono-S HR 5/5 column from Amersham Biosciences (Piscataway, NJ, USA), and [ 35 S]methionine from ICN-Biomedicals (Irvine, CA, USA). Cell lines were provided by American Type Culture Collection (Manassas, VA, USA). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum were provided by Invitrogen (Carlsbad, CA, USA). Protein expression BL21(DE3) cells, transformed either with plasmid pET11d- (Met1)-ONC (M23L) encoding (Met1)-Gln1-ONC with a Met in position 23 substituted by Leu, or with plasmid pONC encoding Gln1-ONC, were grown to an attenuance at 550 nm (D 550 )of 1.0–1.5 in LB (Luria–Bertani) medium containing 100 lgÆmL )1 ampicillin. Protein expression was induced by the addition of isopropyl thio-b- D -galactoside (IPTG) to a final concentration of 1.0 m M . After 3 h, the cells were collected by centrifugation, and the cell pellets were stored at )20 °C. Protein purification Both (Met1)-ONC (M23L) and rONC were purified according to a previously described procedure [18]. Briefly, frozen pellets from 1 L of induced culture were resuspended in 30 mL of 50 m M Tris/HCl, 10 m M EDTA, pH 8.0; the cells were lysed by French Press and inclusion bodies were harvested by centrifugation (12 000 g,45min,4°C). The pellets were then resuspended in 10 mL of 6 M guanidinium chloride, 2 m M EDTA, 100 m M Tris-acetate, pH 8.5, to permit the solubilization of proteins from inclusion bodies. The samples were reduced by the addition of glutathione (GSH) to a final concentration of 0.1 M ,thepHwas adjusted to 8.5 with solid Tris, and the samples were incubated at room temperature for 2 h under a nitrogen atmosphere. Insoluble material was then removed by centrifugation (12 000 g,30min,4°C). Solubilized protein was diluted dropwise ( 100-fold), to a final concentration of 50–100 lgÆmL )1 ,into0.5 ML -arginine, 1 m M oxidized glutathione (GSSG), 2 m M EDTA, 0.1 M Tris-acetate, pH 8.5, and then incubated at 10 °Cforatleast24h.To prevent refolding, the pH was adjusted to 5.0 with acetic acid. The sample was then concentrated by ultrafiltration and dialyzed against 50 m M sodium acetate, pH 5.0. Preci- pitated or insoluble material was eliminated by centrifu- gation (12 000 g,10min,4°C). The refolded sample was then loaded onto a Mono-S HR 5/5 FPLC column, and recombinant onconases were eluted with a linear gradient of 0–600 m M NaCl in 30 min. Fractions containing the pure (Met1)-ONC (M23L) or rONC were dialyzed against ultrapure water, lyophilized and stored at )20 °C. Hydrolysis and cyclization as a function of guanidinium chloride concentration and pH (Met1)-ONC (M23L) (0.1 m M ) was prepared in 0.2 M potassium phosphate buffer containing 0.5 m M ZnSO 4 with varying guanidinium chloride concentrations (0, 2, 3, 4.5 and 6 M ) and at different pH values (6.2, 7.2 or 8.0). The samples were equilibrated for 30 min at 37 °C. The hydrolysis reaction was started by adding 1 lLof 4 · 10 )6 M AAP in the appropriate buffer to obtain a 1 : 2500 molar ratio (AAP : ONC) 8 , and incubated at 37 °C. Aliquots of 2 lL were taken at 0, 1, 2, 4, 6, 8, 24 and 30 h, diluted 10-fold in 0.1% TFA-CH 3 CN (2 : 1), mixed with an equal volume of saturated sinapinic acid in the same solvent of the matrix, and 0.5 lL of the mixture was immediately applied to a sample probe spot. Cyclization reaction as a function of pH and temperature In order to determine the effect of pH and temperature on the cyclization of the N-terminal glutaminyl to pyrogluta- myl, 0.1 m M (Met1)-ONC (M23L), in 0.2 M potassium phosphate buffer containing 0.5 m M ZnSO 4 ,3 M guanidi- nium chloride, pH 8.0, was equilibrated for 30 min at 37 °C. The hydrolysis reaction was started, as described 1164 M. Ribo ´ et al. (Eur. J. Biochem. 271) Ó FEBS 2004 above, and allowed to proceed for 1 h. Once completed, the sample was split into different aliquots, the pH was adjusted to 2.5, 8.0 and 11.5, and aliquots of each pH value were incubated at 37 °C, 50 °Cand70°C, respectively. From these, aliquots of 2 lL were taken at 0, 1, 2, 4, 6 and 24 h, and prepared for MS analysis as described above 9 . MALDI-TOF MS All mass spectra were acquired on a Bruker ULTRAFLEX TOF mass spectrometer, at the Servei de Proteo ` mica i Bioinforma ` tica (Universitat Auto ` noma de Barcelona, Barcelona, Spain), equipped with a nitrogen laser with an emission wavelength of 337 nm. Spectra were obtained in the linear positive mode at an accelerating voltage of 25 kV, and deflection of the low mass ions m/z<5000 U was used to enhance the target protein signal. Fifty spectra were accumulated from 10-shot runs using the automatic mode in order to minimize subjectivity. The criteria for acceptance were a resolution greater than 1000 (m/FWHM) and a signal-to-noise ratio greater than 10. Each data point was averaged from two independent experiments. For each experiment, mass spectra were acquired from duplicates. The accuracy in mass determination obtained for the different species was ± 2 U. Mass values given through- out the text correspond to protonated molecular ions [M+H] + . Activation of onconase A 10 mg sample of (Met1)-ONC (M23L) was dissolved at a concentration of 1 m M in 0.2 M potassium phosphate, 0.5 m M ZnSO 4 ,3.0 M guanidinium chloride, pH 8.0. AAP was added to the solution at a molar ratio of 1 : 2500 (AAP : ONC) and the mixture was incubated for 1 h at 37 °C followed by a 3 h incubation at 70 °C. The protein was dialyzed overnight at 10 °Cagainst20m M potassium phosphate, pH 7.2. Finally, the protein was purified by ion- exchange chromatography in a Mono S HR 5/5 FPLC column, using a linear gradient of 0–200 m M NaCl in 20 m M potassium phosphate, pH 7.2, for 25 min. Pyr-ONC (M23L), eluted as a single peak, was dialyzed against ultra pure water to remove salts, then lyophilized. Determination of conformational stability The conformational stability of (Met1)-ONC (M23L) and Pyr-ONC (M23L) was determined using UV spectroscopy to measure the change in environment of the aromatic residues during protein thermal unfolding. The proteins were dissolved at 0.5 mgÆmL )1 in a buffer of 50 m M glycine/ HCl, pH 2.0, and the UV absorbance was monitored at 278 nm. The temperature was raised from 35 to 80 °C in 2 °C increments. The decrease in UV absorbance was registered after a 5 min equilibration at each temperature. Reversibility was verified by decreasing the temperature in 10 °C decrements, equilibrating the sample for 10 min and monitoring the absorbance at 278 nm. Unfolding transi- tions curves induced by temperature were fitted to a two- state thermodynamic model combined with sloping linear functions for the native and denatured states, as described previously [19]. The free energy of thermal unfolding (DG U ) was calculated as described previously [17]. The DCp value for (Met1)-ONC (M23L) and rONC was taken from a previous publication [20]. Curves were normalized to facilitate comparison of the unfolding transitions induced by temperature. Cytotoxicity assays Cytotoxicity was evaluated by measuring the incorporation of [ 35 S]methionine into newly synthesized cell proteins. Human chronic myelogenous leukemia cells, K562, and human epidermoid carcinoma cells, A431, were grown at 37 °C in DMEM, supplemented with 2 m ML -glutamine, 50 UÆmL )1 penicillin, 50 UÆmL )1 streptomycin and 10% fetal bovine serum, in a humidified incubator containing 5% CO 2 (v/v). Both cell types were cultured in 96-well plates, at adensityof2.5· 10 4 cellsÆmL )1 . A431 cells were allowed to attach for 2 h before the addition of ONC. After 3 days, [ 35 S]methionine was added (500 000 c.p.m. per well). Twenty-four hours later, the cells were lysed with NaOH and the protein precipitated with 15% trichloroacetic acid (w/v). These proteins were collected onto glass fiber filters using a cell harvester, washed with 5% trichloroacetic acid (w/v), and radioactivity was counted using a liquid scintil- lation counter. Background levels were obtained from cells treatedwith1m M cycloheximide [21]. The results are expressed as a percentage of control values obtained from samples without ONC. The data for a single experiment was the average of four determinations and experiments were repeated three times. The IC 50 values represent the ONC concentration that inhibited the cell protein synthesis by 50%. Other methods ONC concentration was determined by UV spectroscopy, using an extinction coefficient of 10 280 M )1 Æcm )1 [22]. AAP concentrations were also determined spectrophotometri- cally using an extinction coefficient of E 1% 1cm ¼ 14.4, at 278.5 nm [23]. Protein sequence determinations were per- formed on a LF-3000 Beckman-Coulter sequencer, connec- ted online to an HPLC apparatus for the identification of the phenylthiohydantoins, at the Servei de Proteo ` mica i Bioinforma ` tica (Universitat Auto ` noma de Barcelona). Electrophoresis on polyacrylamide gels under denaturing conditions (SDS/PAGE) was carried out as described previously [24]. Results Expression and purification of recombinant onconases Recombinant (Met1)-ONC (M23L) was expressed in the cytosol of E. coli, whilst rONC was targeted to the periplasmic space using a signal sequence. Both proteins were purified according to the procedure described previ- ously [18]. After purification, both recombinant products were chromatographically homogeneous and migrated as single species of identical relative mass when assayed by SDS/PAGE (data not shown). For recombinant (Met1)- ONC (M23L), the yield was typically 15–20 mgÆL )1 of culture, which was, on average, double the yield observed Ó FEBS 2004 MS characterization of onconase activation 1 (Eur. J. Biochem. 271) 1165 for rONC. This correlates well with the observation of an induction band only in the SDS/PAGE analysis of crude extracts of BL21(DE3)-pET11d-(Met1)-ONC (M23L) induced cells. These yields were comparable to those obtained by using other methods for the production and purification of the enzyme [13]. The molecular mass of each purified product, as measured by MALDI-TOF MS, was 11 947.87±2 for (Met1)-ONC (M23L) and 11 838.21±2 for rONC, which conforms well to the expected values: 11 949.90 for (Met1)-ONC (M23L) and 11 836.82 for rONC with a glutaminyl residue at the N-terminus. Reaction monitoring by MALDI-TOF MS To determine the optimal technical conditions for acquiring the spectra, we ran experiments in which samples were taken at a specific time and placed on a probe spot. As soon as the samples were dry, the probe was transferred into the mass spectrometer so that spectra could be acquired immediately. The mass spectra for the same samples were acquired again, 24 h and 72 h later. Comparing the peaks of the different spectra revealed that they were indistinguishable in relative intensity. Probably dehydration, together with acidification and dilution of the samples, stopped the reactions. This finding allowed the acquisition of spectra in the automatic mode, once all the samples had been taken and placed on the sample probe spot. A second relevant technical aspect was that the three species under study (Met1)-ONC (M23L), Gln1-ONC (M23L) and Pyr-ONC (M23L) were resolved in the same spectrum. In Fig. 1, a time-course graph showing the conversion from the reactive to the final product is presented. It is clear, from the figure, how the three species are clearly identifiable and how their intensities vary with time, allowing the fraction of each species to be determined at any time. Characterization of the hydrolysis reaction Met1 removal. The election of the buffer system was based on previous studies on AAP activity [23] and on the influence of ions in the cyclization reaction [25]. Both studies coincided in that phosphate was the system of choice – while it was recommended for the exopeptidase activity, it also accelerated the conversion of glutaminyl to pyroglutamyl residues. The hydrolysis reaction was monitored by calculating the sum of the relative intensities of the peaks corresponding to Gln-ONC (M23L) (theoretical molecular mass 11 818.87) and Pyr-ONC (M23L) (theoretical molecular mass 11 801.60), and expressing it as the fraction of hydrolyzed (Met1)-ONC (M23L). Figure 2 shows the results of the characterization of the hydrolysis reaction as a function of guanidinium chloride concentration and pH. In order to make the differences between the analyzed conditions clearer, only the values for the first 8 h are shown. Within this period of time, and under specific conditions, the hydrolysis reaction is com- plete. When analyzing at the effect of guanidinium chloride, it is notable that hydrolysis is favored by the increase in guanidinium chloride concentration, being completed in as little as 1–2 h at pH 7.2 or pH 8.0, when the guanidinium chloride concentration is ‡ 3 M . At pH 6.2, significant differences can be observed in terms of guanidinium chloride concentrations and the extent of the hydrolysis reaction, which does not reach completion, in any case, before the first 8 h. At 3 M guanidinium chloride, the reaction is complete within 24 h; however, it is still only partially complete after 30 h at 0 M (60%) and 2 M (80%) guanidinium chloride. At pH 7.2, the differences in terms of guanidinium chloride concentrations and Met1 removal are also clear. In this case, the reaction is favored by higher guanidinium chloride concentrations, being completed in 1 h at 4.5 M and 6 M guanidinium chloride (data not shown), in 2 h at 3 M guanidinium chloride, whereas it does not reach completion (90%) after 30 h of digestion at lower guani- dinium chloride concentrations. Finally, hydrolysis at pH 8.0 was completed at all the guanidinium chloride concentrations assayed: it takes 1, 8 and 24 h at 3, 2 and 0 M guanidinium chloride, respectively. In short, the hydrolysis reaction was clearly slowed down by decreasing the concentration of denaturing agent or lowering the pH. At ‡ 3 M guanidinium chloride, and at pH 7.2 or pH 8.0, the reaction was complete within 1–2 h. Characterization of the intramolecular cyclization reaction Pyroglutamyl formation. From the set of data described in the preceding section, it was also possible to monitor the cyclization reaction as a function of guanidinium chloride concentration and pH, by calculating the fraction of Pyr- ONC (M23L) or cyclization. This is defined as the ratio Fig. 1. Mass spectra [(m/z) range 9000–13 000 U] showing the con- version of (Met1)-ONC (M23L) (an onconase variant with a methionine preceding Gln1) to (Gln1)-ONC (M23L) (an onconase variant with a Gln1 and leucine replacing methionine at position 23) and afterwards to (Pyr)-ONC (M23L) (an onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23). Samples were takenat0,1,2,6and24handcorrespondtoanexperimentcarried outat37°C and pH 8.0 in the presence of 3 M guanidinium chloride. 1166 M. Ribo ´ et al. (Eur. J. Biochem. 271) Ó FEBS 2004 between the relative intensities [Pyr-ONC (M23L)/Gln- ONC (M23L) + Pyr-ONC (M23L)]. This value is prefer- able to measuring only the appearance of Pyr-ONC (M23L), as this would not be a true estimate of the cyclization because this second reaction depends on the hydrolysis reaction. When analyzing this set of data, we did not find such a strong dependence of cyclization on pH or guanidinium chloride, as was the case for hydrolysis. However, when considering only the first 6 h of the reaction, it was found that the rate of cyclization is almost linear and can be obtained by calculating the slope of the curves of Pyr-ONC (M23L) vs. time. From the rates of cyclizationshowninTable1,itcanbeinferredthatthe formation of the N-terminal pyroglutamyl is not dependent on pH in the range 6.2–8.0. Interestingly, however, the reaction is slowed down when the guanidinium chloride concentration increases. These results concur with the observation that the rate of conversion of L-Gln to Pyr in aqueous solution is minimum at near neutral pH values [26]. To further characterize and establish the optimum productive procedure for obtaining Pyr-ONC (M23L), once the optimal hydrolysis conditions had been estab- lished, we focused our attention on the cyclization reaction. To evaluate the contribution of both pH and temperature to the cyclization of the N-terminal glutaminyl to pyroglu- tamyl, Met1 was completely hydrolyzed from (Met1)-ONC (M23L) by the addition of AAP, as described in the Materials and methods. Completion of the hydrolysis was confirmed by the disappearance of the (Met1)-ONC (M23L) signal in the MALDI-TOF mass spectra 1 h after the addition of AAP. The pH was adjusted to 2.5, 8.0 and 11.5 and the samples were incubated at 37 °C, 50 °Cand 70 °C. Aliquots were taken at different time-points from each sample and analyzed by MALDI-TOF MS. The addition of the protease is considered the zero time point. The relative intensities of the peaks from the mass spectra were used to calculate the fraction of Pyr-ONC (M23L), as described above. The results are shown in Fig. 3. Because the cyclization reaction is readily accomplished at very short times under specific conditions, and to make comparison easier, only data corresponding to the first 6 h are shown. Comparing the data obtained for particular temperatures (Fig. 3), it can be seen that at 37 °C, the reaction was as fast at pH 2.5 as it was at pH 8.0 for the first 6 h, although it was not completed before 24 h. At pH 11.5, degradation by nonspecific hydrolysis was observed after 6 h. At 50 °C, the protein was degraded after 2 h under basic conditions and after 6 h under acidic conditions. At this temperature, the reaction was slightly faster at pH 8.0 than at pH 2.5, and 95% of Pyr-ONC (M23L) was observed after 6 h. Finally, data obtained at 70 °C showed that the formation ofpyroglutamylwascompletedin4hatpH8.0,withno degradation of the protein. At pH 2.5, although cyclization was faster at 70 °C when compared with lower tempera- tures, it was not completed because degradation of the protein was also observed after 6 h of reaction. Analysis of the data as a function of pH (Fig. 3), showed that cyclization at pH 2.5 is faster when the temperature is increased, although for all the temperatures used in the assay, the reactions were not complete within the first 6 h. Fig. 2. Comparison of the hydrolysis reaction as a function of denaturing agent and pH. (A), (B) and (C): (d), (s)and(.) correspond to guanidinium chloride concentrations of 0 , 2 and 3 M , respectively. Hydrolysis was calculated as described in the Results. Table 1. Rate of (Pyr1)-ONC (M23L) (an onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23) formation as a function of guanidinium chloride concentra- tion and pH. The rate of (Pyr1)-ONC (M23L) formation was calculated as the slope of the curves generated by representing the cyclized fraction, defined as [(Pyr1)-ONC (M23L)/(Gln1)-ONC (M23L) + (Pyr1)-ONC (M23L)] during the first 6 h of the reaction. pH Guanidinium chloride concentration 0 M 2 M 3 M 6.2 0.1155 0.0728 0.067 7.2 0.1269 0.0734 0.066 8.0 0.1133 0.0834 0.067 Ó FEBS 2004 MS characterization of onconase activation 1 (Eur. J. Biochem. 271) 1167 At pH 8.0, cyclization also increased as a function of the temperature and, at 70 °C, only 4 h of incubation was required to reach 100% completion. When the temperature was lowered to 50 °C, 6 h of incubation was required to obtain 90% Pyr-ONC (M23L), while, at 37 °C, the reaction was not complete until 24 h after initiation. At pH 11.5, it was only possible to quantify the sample incubated at 37 °C up to 6 h because, after this time, extensive degradation of the sample was observed. At 50 °Cand70°C, the same phenomenon was observed after just 2 h. To sum up, under the conditions of the assay, the intramolecular cyclization of glutaminyl to pyroglutamyl at the N-terminus of ONC (M23L) is not accelerated by lowering the pH, as observed for free L-Gln in aqueous solutions [26]. Strong basic conditions are not recommen- ded, because protein degradation is observed even after short reaction times and at 37 °C. The cyclization reaction is clearly accelerated by temperature and can be completed in as little as 3 h at 70 °C (4 h after the addition of the exopeptidase). Preparative activation Activation of (Pyr)-ONC (M23L) was performed as described in the Materials and methods. The presence of pyroglutamyl at the N-terminus of the protein was confirmed by protein sequence determination, which resul- ted in no phenylthiohydantoin derivatives, indicating a blocked N-terminus. Protein identity and integrity was also confirmed by MALDI-TOF MS. The measured molecular mass of the purified product was 11 799.70 ± 2, which concurs closely with activated (Pyr)-ONC (M23L), which has a theoretical molecular mass of 11 801.60. Contribution of Met1 hydrolysis and pyroglutamyl formation to the conformational stability of active onconase In order to evaluate the contribution of the (Met1) hydrolysis and pyroglutamyl formation to the global conformational stability of the protein, the thermal dena- turation curves of (Met1)-ONC (M23L), Pyr-ONC (M23L) and rONC were determined at pH 2 in 100 m M glycine/HCl buffer by monitoring the UV absorbance at 278 nm. The temperature unfolding transitions were reversible and fitted, as previously described [19], to a two-state thermodynamic model. In Fig. 4, the normalized curves are shown. The fitted thermodynamic parameters for thermal transitions are listed in Table 2. A comparison of the experimental data collected for the three proteins makes it possible to estimate the contribution of Met1, and Met23 substituted for Leu, Fig. 3. Comparison of the cyclization reaction as a function of tem- perature and pH. (A), (B) and (C): (d), (s)and(.) correspond to pH 2.5, 8.0 and 11.5, respectively. Cyclization was calculated as des- cribed in the Results. Fig. 4. Normalized temperature unfolding curves. Samples at pH 2 in 100 m M glycine/HCl buffer were subjected to temperature increase (2 °C increments) and changes in absorbance at 278 nm, monitored at each temperature. (m), (s)and(j) correspond to (Met1)-ONC (M23L) (an onconase variant with a methionine preceding Gln1), (Pyr)-ONC (M23L) (an onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23) and rONC (wild-type recombinant onconase), respectively. 1168 M. Ribo ´ et al. (Eur. J. Biochem. 271) Ó FEBS 2004 to the global stability of ONC. It can be seen that the N-terminal Met diminished the T m 10;11 by 6 °CandtheDG U 10;11 by 3.5 kJÆmol )1 , whereas substitution of methionine with leucine at position 23 diminished the midpoint of the thermal denaturation curve (T m )by7°Candthefree energy of unfolding (DG U )by11kJÆmol )1 . These values are in very close agreement with those obtained using differen- tial scanning calorimetry [20], suggesting that the procedure described in this work for the activation of ONC has no detrimental effect on the conformational stability of the enzyme. Cytotoxic activity The effect of (Met1) removal on the cytotoxic properties of ONC was examined on K562 human erythroleukemic cells and A431 human epidermoid carcinoma cells by measuring the incorporation of [ 35 S]methionine into newly synthesized proteins after 96 h of incubation with an increasing concentration of either (Met1)-ONC (M23L) or the activa- ted Pyr-ONC (M23L). The results are shown in Fig. 5. For both cell lines, activated Pyr-ONC (M23L) inhibited cell proliferation, with an 50% inhibitory concentration (IC 50 ) value of 0.3–1 l M , several orders of magnitude lower than those measured for (Met1)-ONC (M23L). IC 50 values in the l M range for ONC have previously been described [27]. These data confirmed that the procedure, described in this work, used to generate activated Pyr-ONC (M23L), resulted in a protein that retains its full cytotoxic potential. Discussion A great deal of work has been directed towards the optimization of experimental conditions in order to control external factors, such as sample preparation methods, matrix solution conditions and matrix crystal morphology, which have been shown to affect the peak intensity of peptides [28]. Recent results suggest that several intrinsic properties of peptides influence their MALDI behavior. Besides charged side-chains, the presence of aromatic amino acids, peptide hydrophobicity, size and the potential to form stable secondary structures, have been reported as influen- cing ion intensity. Furthermore, in proteomics, when using complex systems, so-called suppressive effects were observed in peptide mixtures, such as those obtained by tryptic digestion [29]. As, in the present work, we are dealing with a simple system constituted by only three molecular species, each very similar in net charge, surface charge distribution, hydrophobicity and molecular mass, it is worthwhile to assume that their behavior, in terms of reactivity to the matrix and their ionization properties, will not differ significantly. Thus, the relative intensities of the different peaks constitute a fair estimate of the fraction of each species in a particular condition or reaction time. This has allowed us to monitor how they interconvert over time under the different experimental conditions and to quantify the process. Hydrolysis reaction The enzymatic hydrolysis of (Met1)-ONC (M23L) to produce (Gln1)-ONC (M23L) was monitored by MALDI- TOF MS under different guanidinium chloride concentra- tions and/or pH. It was found that the optimum pH is pH 8.0. This result is in accordance with the fact that the maximum stability and activity of the AAP occurs in the pH range 8.0–8.5 [23]. When evaluating the influence of Table 2. Thermodynamic parameters of the thermal denaturaturation of the different onconase (ONC) variants at 25 °C and pH 2.0. (C Gdm/HCl ) ½ 13 , midpoint of guanidinium chloride denaturation curve; DG U 14;15;16 , free energy of unfolding; DH Tm 14;15;16 , enthalpy of unfolding calculated at T m ; T m 14;15;16 ,midpoint of the thermal denaturation curve. (Met1)-ONC (M23L), an onconase variant with a methionine preceding Gln1; (Pyr1)-ONC (M23L), an onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23; rONC, wild-type recombinant onconase. Gdm, guanidinium. T m a (°C) DH Tm a (kJÆmol )1 ) DG U (25 °C) (kJÆmol )1 ) (C Gdm/HCl ) ½ b ( M ) rONC (wild-type) 65.1 (0.4) 417 (47) 33.8 4.5 (Pyr1)-ONC (M23L) 58.5 (0.2) 355 (29) 26.7 3.4 (Met1)-ONC (M23L) 52.9 (0.08) 345 (32) 23.2 ND a Values in parentheses are the standard errors of the data. b Values reported previously [20]. Fig. 5. Cytotoxicity of onconase variants to K-562 and A-431 cell lines. Cell viability was determined by measuring the incorporation of [ 35 S]methionine into newly synthesized protein after 96 h of incubation with each onconase: (d) (Met1)-ONC (M23L) (an onconase vari- ant with a methionine preceding Gln1) and (s) (Pyr)-ONC (M23L) (an onconase variant with a pyroglutamyl residue at position 1 and leucine replacing methionine at position 23). The results represent the average of three experiments. Ó FEBS 2004 MS characterization of onconase activation 1 (Eur. J. Biochem. 271) 1169 guanidinium chloride on the reaction, the reaction rate increases with increasing concentrations of guanidinium chloride. Enzymatic removal of the N-terminal methionyl residue from other recombinant ribonucleases, such as angiogenin [15] and BS-RNase [16], was performed using the AAP [14] without the need for denaturing agents. However, when the same approach was attempted with onconase, it was found that it was previously necessary to reduce and denature the protein to permit removal of the (Met1) [12], which resulted in a time-consuming procedure. AAP is a very stable enzyme, as it tolerates exposure to a temperature of 70 °C for several hours and is only partially inactivated in 8 M urea. The active site, located at the surface of the protein, is fairly open to the bulk solvent and is accessible via a solvent channel [30]. Thus, the accessibility of the peptidase active site does not seem to be a limiting factor for catalysis under the conditions of this study. On the other hand, mass peaks corresponding to the removal of additional residues were not detected, indicating that the aminopeptidase action was limited to Met1 removal. It has been described that the presence of an acidic residue in the penultimate position reduces the rate of release of the N-terminal residue [23]. In onconase, there is an aspartyl residue in position 2 that could reduce the rate of hydrolysis to an extent sufficient to allow Gln1 to cyclize to pyroglutamyl, avoiding its removal. Afterwards, cyclization prevents further hydrolysis by the aminopeptidase, and a unique final product is formed. ONC has an unusually high resistance against guanidi- nium chloride denaturation and proteases [20]. The closer positioning of the N-terminus to the main protein body is essential for these properties as well as for its activity. Together with a disulfide bridge between the C-terminal Cys104 and Cys87, it contributes to the overall ONC stability [8,31]. The midpoint of the guanidinium chloride denaturation curve of Pyr-ONC (M23L) is 3.4 M [20] (Table 2). At low or zero concentrations of guanidinium chloride, the methionyl residue might be not sufficiently accessible to the AAP and therefore the reaction is slow. At 3 M guanidinium chloride, ONC begins to unfold, making the N-terminal a-helix of ONC more accessible to the active site of AAP and, thus, the hydrolysis reaction is favored. High concentrations of guanidinium chloride do not have a negative effect on the rate of the reaction because, as mentioned previously, AAP remains active, even at high concentrations of the denaturing agent. Cyclization reaction In contrast to other ribonuclease family members, the N-terminal residue of ONC (Pyr1) forms part of the active site [32]. The N-terminal glutamyl residue (Glu1) of ONC cyclizes to form a pyroglutamyl (Pyr) residue, which folds back against the N-terminal a-helix (a1) and forms a hydrogen bond with the CO group of Val96 in the C-terminal b-sheet. Pyr1 is also hydrogen bonded to the e-amino group of the Lys9 side-chain, which simultaneously interacts with the main phosphate of the substrate [32]. The closer positioning of the N-terminus to the main protein body is essential for the activity and, together with the 87–104 disulfide bridge, contributes to the high stability of ONC against temperature [8,31]. Taking advantage of this property, the cyclization reaction from Gln to Pyr was evaluated as a function of temperature, up to 70 °C, by using MALDI-TOF MS. Similarly to many reactions, it was observed that temperature accelerates the reaction rate. However, because the substrate is a polypeptide, care was required with the pH of the reaction. Combining high or low pH with high temperature resulted in the degradation, by partial nonspecific hydrolysis, of the polypeptide chain. When investigating the effect of pH on the cyclization reaction, we expected that an acidic pH might enhance the rate of conversion of (Gln1)-ONC(M23L) to (Pyr1)- ONC(M23L), based on previous studies on the conversion of free L -Gln to Pyr in aqueous solutions which demon- strated that the reaction was accelerated at acidic pH values [26]. It is worth mentioning that cyclization of ONC occurs in the presence of 3 M guanidinium chloride, because this was determined as the minimum concentration of guanid- inium chloride that allowed the maximum rate of hydro- lysis. It was also observed that increasing the guanidinium chloride concentration seemed to slow the cyclization. It is possible, then, that the nucleophilic substitution in the cyclization reaction of Gln to Pyr was not favoured at acidic pH values in the presence of guanidinium chloride, as it is in aqueous solutions. The decrease in the reaction rate, as a function of guanidinium chloride, could be explained because the intramolecular substitution is hampered by a lower local concentration of H + that is necessary for the release of ammonia. Concluding remarks Treatment with AAP was used to remove the undesirable N-terminal methionyl from (Met1)-ONC (M23L). This hydrolysis reaction is essential for rendering an N-terminal Gln1 that cyclizes nonenzymatically to Pyr1. These two reactions are necessary to confer full cytotoxic potential and a higher stability to ONC. Optimal conditions for activation were pH 8.0, 3 M guanidinium chloride and a temperature of 37 °C, to permit the complete release of Met1, followed by a 3 h incubation at 70° to allow the complete cyclization of Gln1 to pyroglutamyl. The activation conditions defined in this work significantly simplify the protocol and shorten the time needed to obtain a fully active ONC compared with other procedures described in the literature. The application of MALDI-TOF MS to the characterization of enzymatic and nonenzymatic reactions could be extended to other simple systems in order to optimize the conditions under which a reaction provides a higher product yield. Of particular interest is the applicability of this strategy to recombinant proteins where the presence of non-wild type, N-terminal residues might have critical consequences on the protein’s activity and its special biological functions, such as cytotoxicity, stability or immunogenicity. Acknowledgements The authors thank Professor Ron Raines for the generous gift of the genes encoding onconase and the variant. We are also indebted to Dr Francesc Canals for his help in the mass spectra acquisition and for the discussion of the results. This work was supported by grants from the Ministerio de Ciencia y Tecnologia (BMC2000-0138-CO2-02, 1170 M. Ribo ´ et al. (Eur. J. Biochem. 271) Ó FEBS 2004 HF2000-0017), and from the Generalitat de Catalunya (SGR2000-64 and SGR2001-00196). M. B. gratefully acknowledges a predoctoral fellowship grant from CIRIT, Generalitat de Catalunya. We are also indebted to Fundacio ´ ÔM. F. de RoviraltaÕ,Barcelona,Spain,for equipment-purchasing grants. References 1. Cummins, P.M. & O’Connor, B. (1998) Pyroglutamyl peptidase: an overview of the three known enzymatic forms. Biochim. Bio- phys. Acta 1429, 1–17. 2. Cockle, S.M., Aitken, A., Beg, F. & Smyth, D.G. (1989) A novel peptide, pyroglutamylglutamylproline amide, in the rabbit pros- tate complex, structurally related to thyrotrophin-releasing hor- mone. J. Biol. Chem. 264, 7788–7791. 3. Saido, T.C., Iwatsubo, T., Mann, D.M., Shimada, H., Ihara, Y. & Kawashima, S. (1995) Dominant and differential deposition of distinct beta-amyloid peptide species, A beta N3 (pE), in senile plaques. Neuron 14, 457–466. 4. Youle, R.J. & D’Alessio, G. (1997) Ribonucleases: Structure and Function. Academic Press, New York. 5. Darzynkiewicz, Z., Carter, S.P., Mikulski, S.M., Ardelt, W.J. & Shogen, K. (1988) Cytostatic and cytotoxic effects of Pannon (P-30 Protein), a novel anticancer agent. Cell Tissue Kinet. 21, 169–182. 6. Saxena, S.K., Sirdeshmukh, R., Ardelt, W., Mikulski, S.M., Shogen, K. & Youle, R.J. (2002) Entry into cells and selective degradation of tRNAs by a cytotoxic member of the RNase A family. J. Biol. Chem. 277, 15142–15146. 7.Ardelt,W.,Mikulski,S.M.&Shogen,K.(1991)Aminoacid sequence of an anti-tumor protein from Rana pipiens oocytes and early embryos. Homology to pancreatic ribonucleases. J. Biol. Chem. 266, 245–251. 8. Notomista, E., Catanzano, F., Graziano, G., Di Gaetano, S., Barone,G.&DiDonato,A.(2001)Contributionofchaintermini to the conformational stability and biological activity of onconase. Biochemistry 40, 9097–9103. 9. Okabe, Y., Katayama, N., Iwama, M., Watanabe, H., Ohgi, K., Irie, M., Nitta, K., Kawauchi, H., Takayanagi, Y. & Oyama, F. (1991) Comparative base specificity, stability, and lectin activity of two lectins from eggs of Rana catesbeiana and R. japonica and liver ribonuclease from R. catesbeiana. J. Biochem. (Tokyo) 109, 786–790. 10. Leu, Y.J., Chern, S.S., Wang, S.C., Hsiao, Y.Y., Amiraslanov, I., Liaw, Y.C. & Liao, Y.D. (2003) Residues involved in the catalysis, base specificity, and cytotoxicity of ribonuclease from Rana catesbeiana based upon mutagenesis and X-ray crystallography. J. Biol. Chem. 278, 7300–7309. 11. Mikulski, S.M., Grossman, A.M., Carter, P.W., Shogen, K. & Kostanzi, J.J. (1993) 12 Phase I human clinical trial of ONCON- ASEÒ (p-30 protein) administered intravenously on a weekly sch- edule in cancer patients with solid tumors. Int. J. Oncol. 3, 57–64. 12. Notomista, E., Cafaro, V., Fusiello, R., Bracale, A., D’Alessio, G. & Di Donato, A. (1999) Effective expression and purification of recombinant onconase, an antitumor protein. FEBS Lett. 463, 211–215. 13. Boix, E., Wu, Y., Vasandani, V.M., Saxena, S.K., Ardelt, W., Ladner, J. & Youle, R.J. (1996) Role of the N terminus in RNase A homologues: differences in catalytic activity, ribonuclease inhibitor interaction and cytotoxicity. J. Mol. Biol. 257, 992–1007. 14. Wagner, F.W., Wilkes, S.H. & Prescott, J.M. (1972) Specificity of Aeromonas aminopeptidase toward amino acid amides and dipeptides. J. Biol. Chem. 247, 1208–1210. 15. Shapiro, R., Harper, J.W., Fox, E.A., Jansen, H.W., Hein, F. & Uhlmann, E. (1988) Expression of Met-(-1) angiogenin in Escherichia coli: conversion to the authentic less than Glu-1 pro- tein. Anal. Biochem. 175, 450–461. 16. Adinolfi, B.S., Cafaro, V., D’Alessio, G. & Di Donato, A. (1995) Full antitumor action of recombinant seminal ribonuclease depends on the removal of its N-terminal methionine. Biochem. Biophys. Res. Commun. 213, 525–532. 17. Torrent,J.,Connelly,J.P.,Coll,M.G.,Ribo,M.,Lange,R.& Vilanova, M. (1999) Pressure versus heat-induced unfolding of ribonuclease A: the case of hydrophobic interactions within a chain-folding initiation site. Biochemistry 38, 15952–15961. 18. Canals, A., Ribo, M., Benito, A., Bosch, M., Mombelli, E. & Vilanova, M. (1999) Production of engineered human pancreatic ribonucleases, solving expression and purification problems, and enhancing thermostability. Protein Expr. Purif. 17, 169–181. 19. Mozhaev, V.V., Heremans, K., Frank, J., Masson, P. & Balny, C. (1996) High pressure effects on protein structure and function. Proteins 24, 81–91. 20. Notomista, E., Catanzano, F., Graziano, G., Dal Piaz, F., Barone, G.,D’Alessio,G.&DiDonato,A.(2000)Onconase:anunusually stable protein. Biochemistry 39, 8711–8718. 21. Beaumelle, B., Taupiac, M.P., Lord, J.M. & Roberts, L.M. (1997) Ricin A chain can transport unfolded dihydrofolate reductase into the cytosol. J. Biol. Chem. 272, 22097–22102. 22. Leland,P.A.,Schultz,L.W.,Kim,B.M.&Raines,R.T.(1998) Ribonuclease A variants with potent cytotoxic activity. Proc. Natl Acad. Sci. USA 95, 10407–10412. 23. Prescott, J.M. & Wilkes, S.H. (1976) Aeromonas aminopeptidase. Methods Enzymol. 45, 530–543. 24. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 25. Khandke, K.M., Fairwell, T., Chait, B.T. & Manjula, B.N. (1989) Influence of ions on cyclization of the amino terminal glutamine residues of tryptic peptides of streptococcal PepM49 protein. Resolution of cyclized peptides by HPLC and characterization by mass spectrometry. Int. J. Pept. Protein Res. 34, 118–123. 26. Arii, K., Kobayashi, H., Kai, T. & Kokuba, Y. (1999) Degrada- tion kinetics of L -glutamine in aqueous solution. Eur. J. Pharm. Sci. 9, 75–78. 27. Leland, P.A., Staniszewski, K.E., Kim, B.M. & Raines, R.T. (2001) Endowing human pancreatic ribonuclease with toxicity for cancer cells. J. Biol. Chem. 276, 43095–43102. 28. Cohen, S.L. & Chait, B.T. (1996) Influence of matrix solution conditions on the MALDI-MS analysis of peptides and proteins. Anal. Chem. 68, 31–37. 29. Krause, E., Wenschuh, H. & Jungblut, P.R. (1999) The domi- nance of arginine-containing peptides in MALDI-derived tryptic mass fingerprints of proteins. Anal. Chem. 71, 4160–4165. 30.Chevrier,B.,Schalk,C.,D’Orchymont,H.,Rondeau,J.M., Moras, D. & Tarnus, C. (1994) Crystal structure of Aeromonas proteolytica aminopeptidase: a prototypical member of the co-catalytic zinc enzyme family. Structure 2, 283–291. 31. Leland, P.A., Staniszewski, K.E., Kim, B. & Raines, R.T. (2000) A synapomorphic disulfide bond is critical for the conformational stability and cytotoxicity of an amphibian ribonuclease. FEBS Lett. 477, 203–207. 32. Mosimann, S.C., Ardelt, W. & James, M.N. (1994) Refined 1.7 A X-ray crystallographic structure of P-30 protein, an amphibian ribonuclease with anti-tumor activity. J. Mol. Biol. 236, 1141– 1153. Ó FEBS 2004 MS characterization of onconase activation 1 (Eur. J. Biochem. 271) 1171 . Quantitative analysis, using MALDI-TOF mass spectrometry, of the N-terminal hydrolysis and cyclization reactions of the activation process of onconase 2 Marc. 4.5 (Pyr1)-ONC (M23L) 58.5 (0 .2) 355 (29 ) 26 .7 3.4 (Met1)-ONC (M23L) 52. 9 (0.08) 345 ( 32) 23 .2 ND a Values in parentheses are the standard errors of the data. b Values