156
Biomacromolecules 2011, 12, 156–166
Succinimidyl Residue Formation in Hen Egg-White Lysozyme Favors the Formation of Intermolecular Covalent Bonds without Affecting Its Tertiary Structure Yann Desfougères,*,†,‡,§ Julien Jardin,†,‡ Vale´rie Lechevalier,†,‡ Ste´phane Pezennec,†,‡ and Franc¸oise Nau†,‡ Agrocampus Ouest and INRA, UMR1253 STLO, F-35042 Rennes, France and INRA UMR1253 STLO, F-35042 Rennes, France Received September 13, 2010; Revised Manuscript Received November 12, 2010
Protein chemical degradations occur naturally into living cells as soon as proteins have been synthesized. Among these modifications, deamidation of asparagine or glutamine residues has been extensively studied, whereas the intermediate state, a succinimide derivative, was poorly investigated because of the difficulty of isolating those transient species. We used an indirect method, a limited thermal treatment in the dry state at acidic pH, to produce stable cyclic imide residues in hen lysozyme molecules, enabling us to examine the structural and functional properties of so modified proteins. Five cyclic imide rings have been located at sites directly accessible to solvent and did not lead to any changes in secondary or tertiary structures. However, they altered the catalytic properties of lysozyme and significantly decreased the intrinsic stability of the molecules. Moreover, dimerization occurred during the treatment, and this phenomenon was proportional to the extent of chemical degradation. We propose that succinimide formation could be responsible for covalent bond formation under specific physicochemical conditions that could be found in vivo.
Introduction Among spontaneous chemical degradations that occur during in vitro and in vivo aging of proteins, deamidation and isomerization are widespread. Both reactions proceed via a cyclic imide intermediate that results from the attack of the R-amino group of the n+1 residue on the side-chain carbonyl group of Asp or Glu residue in the isomerization reaction and Asn or Gln residue in the case of deamidation (Figure 1). The resulting five-membered ring (succinimide intermediate) is then hydrolyzed to give an aspartyl or an isoaspartyl residue (IsoAsp) in the ratio 1:3 to 1:2;1,2 hydrolysis is favored at pH >5.0.3 These reactions have been shown to be implicated in a number of biological events such as in paired helical filament formation,4,5 in amyloid fibril formation and deposition,6-10 in the autocatalytic cleavage of polypeptides,11,12 in the cellular distribution of the catalytic subunit of protein kinase A,13 in the modulation of protein activity,14-17 and in the binding affinity of antibodies.15,18-20 In particular, succinimides have been hypothesized to be implicated in autophosphatase activity21 and in protein cross-linking22 and are also suspected to be nucleation points for fibril formation during Alzheimer’s disease.23 The biological importance of these degradations is highlighted by the existence of a repair mechanism based on protein isoaspartyl methyl transferase activity.24 Moreover, it was proposed that deamidation could be considered to be a molecular clock event.25 Deamidation and isomerization depend on protein structure and physicochemical environment. Many studies have shown that the nature of the residue in position n+1 is crucial.1,25,26 * To whom correspondence should be addressed. E-mail: yann.desfougeres@ unil.ch. † Agrocampus Ouest. ‡ INRA. § Current address: Department of Biochemistry, University of Lausanne, 1066 Epalinges, Switzerland.
Glycyl residue is the most favorable for cyclic imide formation and hence for deamidation and isomerization.25 Protein flexibility is necessary for cyclization. Therefore, secondary structures such as helices and sheets inhibit or dramatically decrease the rate of these spontaneous degradations.27 Deamidation is base-catalyzed between pH 5.0 and 8.0, and protein stability against deamidation is maximal between pH 3.0 and 5.0.28 In contrast, isomerization of aspartyl residues is acid-catalyzed between pH 4.0 and 6.0 and does not occur at pH >7.0.29 Temperature increase promotes deamidation and isomerization. The presence of ions (Tris, phosphate, acetate, carbonate) was shown to favor deamidation from Asn residues.30 Although deamidated and isomerized polypeptides are the subject of many investigations, succinimide derivatives are less studied. This is mainly due to the difficulty to isolate cyclic imide intermediates that are rapidly hydrolyzed under physiological conditions.
Figure 1. General scheme of deamidation and isomerization of Asn and Asp residues. Ct and Nt are the carboxy-terminus and the aminoterminus parts of the protein, respectively.
10.1021/bm101089g 2011 American Chemical Society Published on Web 12/17/2010
Succinimidyl Formation in Hen Egg-White Lysozyme
Hen egg-white lysozyme is a model protein that is prone to succinimide formation from Asp and Asn residues.2,14,31,32 Under mildly acidic conditions, isomerization of aspartic acid residues via a cyclic imide have been observed, whereas deamidation is only marginal.2,32 When the protein is kept in solution under nondenaturing conditions (40 °C at pH 4.0 during 10 days), only Asp101 has been shown to cyclize and gives the residue Suc101 (Succinimide 101).2 When lysozyme is hydrolyzed or heated in aqueous solution at 100 °C for more than 1 h, Asp48 and Asp66 have also been found to form succinimide derivatives.2,32 These three residues are located in loops and are accessible to solvent. Because succinimide derivatives are prone to hydrolysis, the isolation of these compounds is challenging. That is the reason why we decided to study the chemical degradation of lysozyme in the solid state (as a powder), to minimize the effect of the solvent in the hydrolysis of succinimide residues. In this study, we characterized the chemical modifications occurring during the thermal treatment of a model protein (the hen egg-white lysozyme). We demonstrated that lysozyme succinimide derivatives could be obtained in large amount by thermal treatment in the dry state (80 °C for up to 7 days), without any secondary or tertiary structure changes. The stability of the resulting molecules was decreased, as highlighted by equilibrium unfolding experiments and ANS fluorescence increase. Finally, covalently bound dimer formation was observed and was correlated to the number of succinimide intermediates produced, suggesting the role of these intermediates in the dimerization process.
Experimental Procedures Materials. Lysozyme hydrochloride powder (protein purity ≈ 98%, water activity of the powder 0.31, 7.5% water content) obtained by freeze-drying a protein solution of pH 3.5 and lysozyme concentrated solutions was a gift from LIOT (Annezin-les-Be´thune, France). When precise lysozyme concentration was needed, it was determined by absorbance at 280 nm using ε ) 37 750 cm-1 M-1.33 Preliminary experiments demonstrated that imide cycles did not modify the molecular absorption coefficient. Porcin pepsin, tris-(2-carboxyethyl)phosphine (TCEP), 8-anilino-1-naphtalenesulfonic acid (ANS), and Micrococcus lysodeikticus were purchased from Sigma-Aldrich (L‘Isle d’Abeau Chesnes, France). All other reagents were of analytical grade and were obtained from Sigma. Thermal Treatment in the Dry State. Lysozyme samples (powder) were placed in hermetically capped glass tubes. Proteins were then heated for up to 7 days in an oven equilibrated at 80 °C. Comparison of water content in the powders resulting from dry heating revealed that no evaporation occurred during the thermal treatment (data not shown). Dry matter content was determined by dehydration at 106 °C for 8 h. Cation Exchange High-Performance Liquid Chromatography (CE-HPLC). CE-HPLC separations were carried out on a S-HYPERD 10 column (Biosepra) connected to a Waters HPLC system consisting of a Waters 2695 separation module and a Waters 2487 dual absorbance detector. Proteins were eluted in 20 mM sodium acetate buffer pH 5.0 at 1 mL min-1 using a NaCl gradient from 0 to 1 M in 44 min. Elution was followed by monitoring absorbance simultaneously at 214 and 280 nm. We estimated the relative proportion of protein species by calculating the area corresponding to each peak using Empower 2 Software. Samples were usually prepared in 20 mM sodium acetate buffer pH 5.0. To evaluate stability of succinimide derivatives as a function of pH, we dissolved dry-heated lysozyme in 20 mM sodium acetate buffer pH 4.0 or 5.0, in 10 mM sodium phosphate buffer pH 6.5 or 7.0, or in 20 mM Tris-HCl buffer pH 8.0. At time point intervals, aliquots were diluted in 20 mM sodium acetate buffer pH 4.0 before
Biomacromolecules, Vol. 12, No. 1, 2011
157
chromatographic analysis. When lysozyme molecules were purified, fractions were collected and dialyzed (48 h, 4 °C) against 250 volumes of distilled water (pH 3.5) using a dialysis membrane (Spectrum Laboratories) with a nominal cutoff of 6000-8000 Da before freezedrying. Proteolysis of Lysozyme. Lysozyme (560 µM) was first reduced by TCEP (50 µM) in 20 mM sodium phosphate buffer pH 4.5 containing 8 M urea. After 1 h of incubation at room temperature, samples were diluted in 100 mM HCl/KCl buffer pH 1.0 to reach a final protein concentration of 140 µM. Completeness of lysozyme reduction was checked by mass spectrometry analysis (data not shown). We then added 50 µL of pepsin (170 µM in 100 mM HCl/KCl buffer pH 1.0) to 200 µL of reduced lysozyme solutions. A second addition of the enzyme was then performed after 5 h. Hydrolysis was achieved in a total of 24 h at room temperature. Digests were then stored at -20 °C before analysis. Mass Spectrometry Analysis. Mass spectrometry analyses were carried out on a hybrid quadrupole time-of-flight (Q-TOF) mass spectrometer QStar XL (MDS Sciex, Toronto, Canada). When the entire mass of polypeptides was investigated, experiments were performed by direct infusion electrospray ionization mass spectrometry (ESI-MS) using a nano ESI sample source. Samples were prepared in distilled water (pH 3.5) and diluted in acetonitrile (ACN) and formic acid to reach final concentrations of 50% ACN and 0.1% formic acid. Data were collected and processed using Analyst QS 1.1 Sciex software, and the deconvolution of spectra was carried out using Bioanalyst 1.1.5. Standard deviation (0.2 Da) was estimated by repeated calculations of the average molecular mass of the protein for all ions (m/z) observed in the protein envelope. Analysis of the hydrolytic peptides was performed by reverse-phase HPLC separation using a nano-LC system equipped with a PepMap C18 column (3 µm, 100 Å, 75 µm i.d. × 150 mm) and coupled to the mass spectrometer. Proteins were separated using a linear gradient of solution A (2% ACN, 0.08% formic acid, 0.01% trifluoroacetic acid (TFA)) and B (95% ACN, 0.08% formic acid, 0.01% TFA) from 10 to 70% of solution B in 65 min. A voltage of 3300 V was applied to the nanoelectrospray ion source (Proxeon Biosystems A/S, Odense, Denmark). A full continuous MS scan was carried out, followed by three data-dependent MS/MS. The first, second, and third most intense ions from the MS scan were selected individually for collision-induced dissociation. The mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/ MS acquisition using Analyst QS 1.1 software. To identify peptides, all data (MS and MS/MS) were submitted to MASCOT (v.2.2). The search was performed against a homemade database dealing with major milk and egg proteins, which represents a portion of the Swissprot database (http://www.expasy.org). No specific enzyme cleavage was used, and the peptide mass tolerance was set to 0.2 Da for MS and 0.2 Da for MS/MS. Three variable modifications (oxidation of methionine, deamidation of asparagine or glutamine residues, and succinimide formation on asparagine or aspartic acid residues) were selected. For each peptide identified, a minimum MASCOT score corresponding to a p value below 0.05 was considered to be a prerequisite for peptide validation with a high degree of confidence. Fourier Transformed Infrared Spectroscopy. Attenuated total reflectance FTIR spectra were measured at 4 cm-1 resolution with a Bruker Tensor 27 spectophotometer equipped with a monoreflectance Germanium ATR plate and a liquid-nitrogen-cooled mercury-cadmium Telluride detector from Bruker. Powder or liquid samples were placed on the ATR plate, and raw data were corrected for background or buffer, respectively. Using Opus software, 256 scans were averaged and corrected for water vapor contribution and CO2. Circular Dichroism. Circular dichroism (CD) spectra were recorded at 25 °C on a Jasco 810 (Jasco, Bouguenais, France) spectropolarimeter equipped with a thermostatted cell holder. Three spectra recorded at 50 nm min-1 were averaged for each sample. Quartz cells with 2 mm light path length were used for measurements. The mean residue ellipticity ([θ]MRW) was calculated using the formula: [θ]MRW )
158
Biomacromolecules, Vol. 12, No. 1, 2011
Desfougères et al.
Figure 2. Chemical modifications of lysozyme during a thermal treatment (80 °C) in the dry state. (A) Cation exchange liquid chromatography (pH 5.0) of native lysozyme (solid line) and lysozyme heated in the dry state for 7 days (dotted line). (B) Rate of formation of basic (0) and acidic (4) species and rate of loss of native species (O) as a function of heating duration (80 °C). Percentage of each species was assessed by cation exchange liquid chromatography. (C,D) Deconvoluted mass spectra obtained for the native and the 7 day dry-heated samples, respectively. amu is atomic mass unit.
(θobsdMRW)/(10lc) where θobsd is the observed ellipticity in degrees, MRW is the mean residue molecular weight (111.8 for lysozyme), l is the path length in centimeters, and c is the protein concentration in grams per milliliter. Samples were prepared by diluting concentrated aqueous lysozyme solution (pH 3.5) in 10 mM phosphate buffer pH 7.0 immediately before analysis. Protein concentrations were 119 and 11.9 µM for near and far UV measurements, respectively. Fluorescence Measurements. Fluorescence measurements were carried out on a Perkin-Elmer LS-50B using a 1 cm path length quartz cell at 100 nm min-1. Concentrated aqueous lysozyme solutions were diluted in 10 mM phosphate buffer pH 7.0 immediately before analysis to reach a final protein concentration of 4 µM. For intrinsic fluorescence experiments, excitation wavelength was 295 nm, and three accumulated spectra were obtained at emission wavelength from 320 to 420 nm. For ANS fluorescence experiments, samples contained 40 µM ANS, and three accumulated spectra were recorded between 420 and 520 nm after excitation at 370 nm. Muramidase Activity Measurements. Lysozyme catalytic activity was measured as described by Ibrahim et al.34 We added 100 µL of aqueous lysozyme solution (7 µM) to 1.9 mL of Micrococcus lysodeikticus suspension (0.17 mg mL-1) prepared in 50 mM phosphate buffer pH 6.2. The absorbance decrease was recorded at 450 nm during 3 min at 25 °C. Activity is expressed as the rate of absorbance decrease per minute. The results presented here are the mean values of at least five measurements of independent experiments. Electrophoresis. Electrophoresis were carried out following standard procedures.35 The gels were stained with Coomassie Brilliant Blue
G-250. When needed, proteins were reduced and alkylated: 350 µM lysozyme were prepared in 10 mM Tris-HCl pH 8.0 containing 8.0 M urea 50 mM DTT and 10 mM EDTA, and kept 1 h at 37 °C. Then, samples were diluted twice in 10 mM Tris-HCl pH 8.0 before alkylation by the addition of iodoacetamide to a final concentration of 100 mM (1 h in the dark).
Results Charge and Mass Heterogeneity As a Result of Succinimide and Isoaspartate Formation during Dry-Heating of Lysozyme. Analysis of native (not dry-heated) sample by cationexchange high-performance liquid chromatography (CE-HPLC) at pH 5.0 revealed three peaks (Figure 2A). We suggest that the major one (85% of total proteins) corresponded to the unmodified lysozyme. The more basic species (8%) might correspond to proteins containing one cyclic imide already described.2 The more acidic species (7%) could result from deamidation of lysozyme.1,2 Indeed, mass spectrometry analysis indicated that the molecular mass of this acidic species was 14 306.1 ( 0.2 Da, that is, 1 Da heavier than native lysozyme, suggesting deamidation (Table 1), whereas the major peak consisted of lysozyme of molecular mass identical to theoretical mass deduced from the amino acid sequence (14 305.2 Da). The third species, of molecular mass 14 287.5 Da, corresponded to a lysozyme that lost 18 Da (Figure 2C). We propose that these
Succinimidyl Formation in Hen Egg-White Lysozyme
Biomacromolecules, Vol. 12, No. 1, 2011
159
Table 1. Molecular Mass of the Lysozyme Species Resulting from a Thermal Treatment in the Dry State (80°C) and Purified by Cation Exchange Chromatographya elution time (min) mass ((0.2 Da)b experimental interpretation 13 15 18 22 24 27
14 306.1 14 305.2 14 287.5 14 269.4 14 251.7 14 233.8
0.9 0.0 –17.7 –35.8 –53.5 –71.4
+H –H2O –2H2O –3H2O –4H2O
a Elution times are those indicated in the chromatograms shown in Figure 2. b Standard deviation.
more basic species consist of succinimide derivatives, as corroborated by the results described below. It is noticeable that the charge heterogeneity, detected in native lysozyme by CE-HPLC at pH 5.0 (Figure 2A, solid line), was not detected by methods such as acid-urea polyacrylamide gel electrophoresis, reverse-phase HPLC, and size exclusion chromatography (data not shown) and was hardly detectable when CE-HPLC was performed at pH 8.0 (data not shown). Heating lysozyme in the dry state (80 °C) resulted in an increase in the succinimide derivatives content, whereas the proportion of acidic species (deamidated lysozyme) did not significantly increase (Figure 2A,B). During the first 4 days, the rate of conversion from native to succinimidyl residues (Suc) was high and then decreased with increasing heating time up to 7 days (Figure 2B). Remarkably, more than one modification per polypeptide chain occurred during the thermal treatment in the dry state, as shown by the presence of several peaks in the more basic fraction (Figure 2A). This was confirmed by mass spectrometry analysis of the dry-heated lysozyme: up to five species with a molecular mass of n times -18 Da compared with native lysozyme were obtained (Figure 2D). This is consistent with succinimide formation with the loss of a water molecule (18 Da). After 7 days of dry heating, native lysozyme molecules, acidic species, and succinimide derivatives represented about 35, 10, and 55% of initial proteins, respectively (Figure 2B). Native lysozyme molecules that are present at the end of the treatment, meaning those eluted at 15 min in CEHPLC at pH 5.0, will subsequently be called “native-like” species to distinguish them from lysozyme molecules that were not heated in the dry state. Because succinimide formation and stability are pH-dependent,3,28,29 lysozyme samples were prepared at pH 3.5, 7.0, and 8.5 and freeze-dried before dry-heating. Figure 3 shows that thermal treatment in the dry state (80 °C, 7 days) at pH 7.0 and 8.5 resulted in low succinimide derivatives formation. In contrast, the higher the pH value, the higher the amount of acidic species. The lowest amount of native-like residual proteins was observed when heating at pH 3.5, whereas the higher amount was detected at pH 7.0 (Figure 3). Altogether, these results are consistent with succinimide formation at all pH values but hydrolysis only at neutral or basic pH values, hydrolysis mainly resulting in isoaspartate formation or deamidation. That is the reason why pH 3.5 was chosen for this study, to produce high quantities of succinimide derivatives. The stability of the succinimide derivatives, once formed by dry-heating at pH 3.5, was assessed by following the rate of cyclic imide hydrolysis at different pH values (Figure 4). At pH 4.0 and 5.0, the proportion of each species did not vary significantly during 48 h. At pH values g6.5, succinimide derivatives were hydrolyzed to give native-like and acidic species: the higher the pH value, the higher the rate of conversion. Because the whole dry-heated sample exhibits a
Figure 3. Relative abundance of acidic, succinimide, and native-like species after dry-heating of lysozyme (80 °C, 7 days). Sample pH before freeze-drying is 3.5 (white bars), 7.0 (gray bars), or 8.5 (black bars). The percentage of each species was assessed by cation exchange chromatography.
unique molecular mass of 14 305 Da after hydrolysis at pH 8.0 (data not shown) and because previous studies reported different ionization properties for aspartic and isoaspartic residues,1,2 we propose that the more acidic species formed under these conditions are isoaspartic-acid-containing molecules. It is noticeable that at pH 8.0, 35 ( 3% of the succinimide-containing proteins were transformed into the IsoAsp species as a result of cyclic imide hydrolysis, whereas only
