Enzymatic Properties of the Sweet-Tasting Proteins Thaumatin and Monellin after Partial Reduction

Transcription

1 Eur. J. Biochem. 104, (1980) Enzymatic Properties of the Sweet-Tasting Proteins Thaumatin and Monellin after Partial Reduction Hendrik VAN DER WEL and Wim J. BEL Unilever Research Vlaardingen, The Netherlands (Received July 13, 1979) After partial reduction of disulfide bonds in the thaumatins, the sweet-tasting proteins from the fruits of Thaumatococcus danielii Benth, a rapid autodigestion was demonstrated. In the presence of suitable substrates, the reduced thaumatins showed protease, amidase and esterase activity. Thiol-blocking reagents like mercury(i1) chloride inhibited the enzymatic activity. Of the thaumatins b, c, I, I1 and 111 (with increasing isoelectric points), thaumatin I showed the lowest enzymatic activity. In this series, the enzymatic activity increased from thaumatin I to thaumatin I11 as well as from thaumatin I to thaumatin b. Acetylation of the E-amino group of lysine residues in the thaumatins by acetic anhydride, causing a decrease in basicity, led to an increase in enzymatic activity, which is correlated with the number of acetyl groups introduced. Comparison of the amino acid sequence of thaumatin I with that of cysteine proteases of plant origin showed no similarities. Moreover, the thaumatins lack histidine, one of the amino acids in the active site of the cysteine proteases. Monellin, the sweet-tasting protein from the fruits of Dioscoreopliyllum cumminsii Diels. is not enzymatically active. However, when monellin with acetylated s-amino groups of lysine residues was brought into a reducing environment it appeared to be enzymatically active. The similarities in properties of the thaumatins and monellin suggest a structural relationship between these proteins. The protein sweeteners thaumatin I, 11, 111 [I], b and c [2] from Thaumatococcus daniellii Benth are nearly times sweeter than sucrose on a molar basis. These proteins have very similar relative molecular masses (22 000) and amino acid compositions. However, there are small differences in the number of carboxylic amide groups. They are very basic proteins with isoelectric points around ph 12. Previous data point to the importance of the tertiary structure for the sweet taste sensation [l, 31 and knowledge of this structure can show us the active site in the molecule responsible for the sweet taste sensation. Thaumatin I consists of a single polypeptide chain [4] with 207 amino acid residues, eight disulfide bridges and no free SH groups. To locate the disulfide bonds in the molecule, partial reduction with reducing reagents like dithiothreitol can be applied. When applying this selective reduction we observed an unexpected biological activity of these sweet-tasting proteins : the Abbreviations. Bz-Arg-Nan, N-benzoyl-L-arginine-p-nitroanilide; Cbz-Gly-ONp, N-benzoyloxycarbonyl-glycine p-nitrophenyl ester; dansyl chloride, l-dimethylaminonaphthalene-5- sulfonyl chloride. occurrence of protease activity after reduction of one disulfide bridge. This enzymatic activity is not present in the native thaumatin. In this paper we describe the autodigestive behaviour, the proteolytic activity towards casein and the enzymatically catalysed hydrolysis of low-molecular-weight chromogenic substrates by partially reduced thaumatins. Monellin [5,6], the sweet-tasting protein from Dioscoreophyllum cumminsii Diels with a sweetness intensity on a molar basis equal to that of thaumatin, has been included in these studies. This protein (relative molecular mass 10700) consists of two non-identical polypeptide chains and contains 94 amino acid residues. Only one cysteine residue is present [7]. In addition to this the effect of acetylation on the hydrolytic activity has been studied. MATERIALS AND METHODS Materials Thaumatins I, 11, 111, b and c and monellin were isolated as described previously [l, 51. Acetylated thau-

2 41 4 Enzymatic Properties of Thaumatin and Monellin matins were prepared [8] using a 2 : 1 molar ratio acetic anhydride : protein. Acetylated monellins were prepared in a similar way. Hammarsten casein, cysteine hydrochloride, iodoacetamide, N-benzoyl-L-arginine-p-nitroanilide hydrochloride and dansyl chloride were purchased from E. Merck (Darmstadt, F.R.G.). N-Benzoyloxycarbonyl-glycine p-nitrophenyl ester was obtained from B.D.H. Chemicals Ltd (Poole, England). Dithiothreitol was bought from Calbiochem (Los Angeles, CA, U.S.A.) and i~do['~c]acetic acid from Amersham, England. Diaflo ultrafiltration membranes were obtained from Amicon Corporation (Lexington, MA, U.S.A.). All other chemicals were of analytical grade. Autodigestion Thaumatin I(0.5 pmol) was incubated with 8 ymol dithiothreitol in 4ml KH2P04/NaOH (0.1 moljl) buffer at ph 8.0. After 0.5, 1.0, 2 and 3 h l-ml samples were taken and subjected to gel filtration on Sephadex G-25. The N-terminal amino acid residues in each fraction were determined by the method according to Gray [9]. A 2-h digest was ultrafiltered on a Diaflo UM-05 filter membrane and subsequently subjected to peptide mapping using the method of van der Ouderaa [lo]. The ph dependence of the autodigestion was determined using the classical method of Kunitz [ll]. Proteolytic Activity The ph dependence and the time dependence of the proteolytic activity were measured according to the method of Kunitz as adapted for papain [12] with casein as substrate. The time dependence of the proteolytic activity was measured by taking samples every 10 min. Amidase Activity Amidase activity towards Bz-Arg-Nan was measured by the spectrophotometric method of Erlanger et al. [13]. The dependence of the hydrolysis on the concentration of the activator was determined by incubation of Bz-Arg-Nan for 1 h with the thaumatins activated by different concentrations of dithiothreitol. The ph dependence was determined by reacting the thaumatins activated by dithiothreitol with Bz-Arg- Nan for 1 h, using acetic acid/naoh buffers at ph and Tris/HCl buffers at ph , all at equal ionic strength (I = moljl). Time-dependent hydrolysis of Bz-Arg-Nan by thaumatin I11 (0.498 pmol/l) in Tris/HCl buffer at ph 7.4 (I = mol/l) and dithiothreitol (0.322 mmol/l) as activator was followed during 40 min. Readings were taken at 5-min intervals. Concentrations of Bz-Arg-Nan were varied between mmol/l. Amidase activity of monellin was tested at different ph values in the same way as described above using a monellin concentration of 21.6 ymol/l. Inhibition of Amidase Activity After 20min of hydrolysis of Bz-Arg-Nan, mercury(i1) chloride or iodoacetic acid was added in 10 excess over the thiol groups of the added activator. Esterase Activity The enzymatic hydrolysis of Cbz-Gly-ONp was followed spectrophotometrically at 400 nm according to Martin et al. [14]. The standard reaction mixture was prepared as follows: to 3.0 ml dithiothreitol (0.187 mmol/l) in KH2P04/NaOH (0.02 mol/l) buffer ph 6.8, small volumes (up to 80 pl) of thaumatin solutions (0.1-5 pmol/l) were added. Immediately afterthat 0.2ml Cbz-Gly-ONp in acetonitrile (2mmol/l) was added to start the hydrolysis. Ion-Exchange Chromatography Acetylated thaumatins and monellin were separated on an SP-Sephadex C-25 ion exchanger equilibrated with Tris/HCl buffer ph 7.65 (0.02 mol/l) using a linear gradient of sodium chloride from 0 to 0.25 mol/l in the same buffer [8]. After pooling and desalting of the fractions amidase and esterase activities were determined by incubation with substrate for 1 h and 5 min respectively. Determination of the Number of Split S-S Bonds in Enzymatically Active Thaumatin I Thaumatin I (0.5 ymol) was dissolved in 47.5 ml Tris/HCl buffer at ph 7.3 (I = 0.2 mol/l). Subsequently ml Bz-Arg-Nan solution (2.99 mmol/l) was added and thaumatin was activated by addition of 4.61 ml dithiothreitol(l0 mmol/l). After 10 min, 1CHzCOOH was added in 10% excess over thiol groups. Ultrafiltration in the dark on a Diaflo UM-2 membrane of the reaction mixture and freeze-drying of the concentrate preceded the separation of blocked enzymatically active thaumatin from native thaumatin using the same column and gradient as described above. Amino acid analysis was carried out on the desalted peak materials obtained by ultrafiltration. Another experiment was carried out with i~do['~c]- acetic acid. In a third experiment, thaumatin was activated in the absence of Bz-Arg-Nan ; blocking was also carried out by the addition of i~do['~c]acetic acid.

3 H. van der We1 and W. J. Be F m 1 ' ' Fraction number Fig. 1. Gelfiltration pattern on Sephadex G-25 of thaumatin I. Autodigests of thaumatin I (0.5 pmol in KH2P04/NaOH, 0.1 mol/l, ph 8.0) after incubation with 8 pmol dithiothreitol. Bed: 1.5 x46cm; flow rate: 10 ml/h; fraction volume: 2.5 ml; eluant: KHzPO4/ NaOH (0.1 mol/l) ph 8.0. Incubation time: (-) 30 min; (-.-.-) 1 h; (----) 2 h; (...) 3 h /' O I / ' Timelrnin Fig. 3. Time-dependent proteol~sis of casein by thaumatin I, II and III. Hammarsten casein 8.33 mg/l. Thaumatin I, 11, I pmol/l, 1.8 pmol/l and 0.9 pmol/l respectively, activated by dithiothreitol (1.25 mmol/l) or cysteine (62.5 mmol/l) in Tris/HCl buffer ph 8.0 (I = molil). (0) Thaumatin I; (0) thaumatin 11; (+) thaumatin I11 activated by dithiothreitol; (A) thaumatin I activated by cysteine " " " ' 0 i Fig. 2. Peptide-map of thaumatin I after autodigestion for 2 h. High voltage electrophoresis in pyridine/acetic acid/water (25 : 1 : 225, by vol; ph 6.5) at 30 V cm-' for 2 h and descending paper chromatography in 1-butanol/acetic acid/water/pyridine (15 : 3 : 12: 10, by vol). Peptides were located with ninhydrine in acetone (1 g/kg) RESULTS Autodigestion The gel filtration patterns of the autodigests of thaumatin I are shown in Fig. 1. The front peak appeared to be unchanged thaumatin while the other fractions are peptides with molecular masses lower than those of thaumatin and other N-terminal amino acids such as valine, leucine, isoleucine, phenylalanine, lysine and tyrosine. After a digestion time of 3 h, most of the thaumatin had been split up into smaller peptides. As peptide mapping (Fig.2) revealed at least 20 well separated spots, it is clear that autodigestion has taken place. The autodigestion as a function of ph for thaumatin I showed optimum activity at ph 8.0, and for thaumatin I1 at ph 8.2. Proteolytic Activity Thaumatin activated by dithiothreitol shows proteolytic activity towards casein. Optimum caseinolytic activities of thaumatins I, I1 and I11 were found at ph 9.0 with a rapid decline below and above this value. Due to the rapid autodigestion it is necessary to add the substrate to the thaumatin solution before the addition of the activator. The time-dependent hydrolysis of cysteine and of dithiothreitol-activated thaumatins is linear as demonstrated in Fig.3. This indicates that within 10 min the maximum amount of activated enzyme is present. Thaumatin I11 appears to be the most active enzyme. The concentrations of thaumatins I and I1 required to reach the same proteolytic activity are 9.8 times and twice that of thaumatin 111 respectively. Amidase and Esterase Activity The amidase and esterolytic properties of the thaumatins were shown by hydrolysis of the synthetic substrates Bz-Arg-Nan and Cbz-Gly-ONp. The amidase activity towards Bz-Arg-Nan in Tris/HCl buffers ph 7.3 with increasing ionic strengths showed a relative maximum at I = mol/l for thaumatins I1 and 111. For thaumatin I the hydrolytic activity increases to Z = mol/l and remains constant with increasing ionic strength. The dependence on the dithiothreitol concentration is illustrated in Fig. 4. To reach a maximum amidase activity at a thaumatin concentration of 1 mol/l, the addition of at least 111, 320 and 360 mol dithiothreitol/l to thaumatins I, I1 or I11 respectively is needed, as calculated from the data of Fig. 4. Addition of excess mercury(i1) chloride or iodoacetic acid inhibited the hydrolysis of Bz-Arg- Nan by the thaumatins. Monellin did not show amidase activity in the ph range tested. When using dithiothreitol concentrations occurring in the plateau of the curve, the hydrolysis of Bz-Arg-Nan against time was linear. The Km(app) for the hydrolysis of

4 416 Enzymatic Properties of Thaumatin and Monellin 0.4 i 1.8 r z Id, 1.5 m :. - L, r &.I--- -A.~ OO Oithiothreitol /rnrnol.c Fig. 4. Hydrolysis of Bz-Arg-Nan by thuuniutins. Dependence on dithiothreitol concentration intris/hcl buffer pli7.3 (I = mol/l). Bz-Arg-Nan 0.89 mmol/l and thaumatin pmol/l; Bz-Arg-Nan 0.93 mmol/l and thaumatin I pmol/l; Bz-Arg-Nan 0.95 mmol/l and thaumatin pmol/l Thaurnatin I/qrnol.l- Fig. 6. Relation between h}jdrolysis OJ Br-Arg-Nan and the thaumatin concentration. Substrate and activator concentrations in Tris/HCl buffer ph 7.3 (I = mol/l): thaumatin 111: Bz-Arg-Nan 0.95 mmol/l, dithiothreitolo.24 mmol/l; thaumatin 11: Bz-Arg-Nan 0.95 mmol/l, dithiothreitol 0.32 mmol/l; thaumatin I: Bz-Arg-Nan 0.87 mmol/l, dithiothreitol 0.58 mmol/l 50 I A [s]- / rnrnol-.~ Fig. 5. Lmew,eaver-Bud plot for the hydrolysis of Bz-Arg-Nan by thuuwlatin 111 Fraction number Bz-Arg-Nan by thaumatin 111 as extrapolated from Fig. 5 was 1.15 mmol/l (compare trypsin with a Km(app) of 0.94 mmol/l). We found a linear relationship between the amidase activity and the enzyme concentration (Fig. 6). Optimum hydrolytic activity towards Bz-Arg-Nan occurred between ph 7.2 and 7.4. The esterolytic activity of the thaumatins was investigated by hydrolysis of Cbz-Gly-ONp. This hydrolysis was studied as a function of time and of thaumatin concentration and both appeared to be linear. Hydrolytic Activity of Acetjlated Thaurnatins and Monellin The ion-exchange patterns of acetylated thaumatin 1 and I1 and monellin are given in Fig. 7 A. The pattern of an aqueous extract of the entire thaumatococcus aril (see [1,2]) is given in Fig.7B which shows the elution volumes of the various thaumatins under exactly the same conditions as for the acetylated ones. Peaks IA and IIA are the native thaumatins I and I1 Fraction wmber Fig. 7. Ion-exchange chromutogriip/ij put/iw. SP-Sephadex C-25 column (37 x 1.5 cm) using a linear sodium chloride gradient from 0 to 0.25 mol/l in 0.02 mol/l Tris/HCl buffer ph 7.65 with a flowrate of 14.8 ml/h and fraction volume of 3.7 ml. (A) Acetylated thaumatin I and I1 and monellin; (B) aqueous extract of entire thaumatococcus aril as shown by their elution volume and number of primary amino groups. Peaks IE and IIE contained insoluble material. Peaks IB, IC and ID are thaumatins I with one, two and three acetylated E-aminolysine residues respectively, determined as described previously [S]. Thaumatins I1 with similar numbers of acetylated lysine residues are present in peak IlB, B

5 H. van der We1 and W. J. Be1 417 Table 1. Hydrolytic activities of acetylated thaumatins and monellins Concentrations for determination of amidase activity in Tris/HCl buffer ph 7.4 (I = molil): Bz-Arg-Nan, 0.93 mmol/l; dithiothreitol, mmol/l; thaumatins, 3.41 pmol/l; monellins, 22.9 pmol/l. Incubation time 1 h. Solutions for determination of esterase activity: 3.0 ml KH2P04/NaOH, 0.02 mol/l, buffer ph 6.7; dithiothreitol, 0.2 mmol/l; 50 pl thaumatins (218 pmol/l) and 0.2 ml Cbz-Gly-ONp in acetonitrile (2 mmol/l). Incubation time 5 min. = not determined Protein Amidase Esterase activity activity Thaumatin IA IB IC ID Monellin IIA IIB IIC IID MA MB MC native A A Number of Split S-S Bonds in Activated Thaumatin I After ion-exchange chromatography of the material obtained from the experiments in the presence of Bz-Arg-Nan, two peaks were present. One peak was intensely sweet and the elution volume and amino acid composition were equal to those of thaumatin I. The other peak was not sweet and was eluted ahead of thaumatini. The only difference in amino acid analysis between this peak and the native thaumatin was the presence of two carboxymethylcysteine residues and 14 cysteine residues instead of 16, so we conclude that in activated thaumatin one S-S bond has been split. In the experiment with iodo[14c]acetic acid the radioactivity was demonstrated in the peak eluted ahead of thaumatin and quantitative radiochemical analysis revealed two [14C Icarboxymethylated cysteines. In the third experiment, the radioactivity was present in a peak with the same elution volume as in the other experiments. In a tryptic digest of the radioactively labelled protein of both I4C experiments, the radioactivity was found in only one tryptic peptide (T15) according to the method of Iyengar et al. [4]. Table 2. Hydrolytic activity of thaumatin b, c, I, II and III Concentrations for the determination of the amidase activity in Tris/HCl buffer ph 7.4 (I = 0.064mol/l) : Bz-Arg-Nan, 0.93 mmol/l; dithiothreitol, mmol/; thaumatins, 3.41 pmol/l. Solutions for the determination of esterase activity : 3.0 ml KH2PO4/Na0H (0.02 mol/l) buffer ph 6.8; dithiothreitol, 0.2 mmol/l; 0.2 ml Cbz- Gly-ONp in acetonitrile (2 mmol/l); 50 p1 thaumatins (218 pmol/l). = not determined Thaumatin Amidase Esterase activity activity A410 A400 b C I I IIC and IID. Peaks MA, MB and MC are acetylated monellins. The amidasic and esterolytic activities have been listed in Table 1. Comparison of Amidase and Esterase Activity of All Thaumatins The isoelectric points of thaumatins b, c, I, I1 and 111 increase as indicated by the increasing elution volume on ion-exchange chromatography (Fig. 7 B). The amidase activity of all thaumatins and the esterase activities of most of them are demonstrated in Table 2. DISCUSSION The disappearance of the sweet taste sensation associated with the thaumatins after incubation with 0.5 mol cysteine/l water was attributed to the disruption of the tertiary structure of the protein [l]. We have now shown that incubation with a reducing reagent like dithiothreitol or cysteine causes autodigestion, leading to loss of sweetness. These sweet proteins might be classified as cysteine proteases as they are activated by reducing reagents. However, the thaumatins as well as monellin lack histidine [l] which is present in the active site of cysteine proteases. Also the amino acid sequence of thaumatin I [4] and that of cysteine proteases from plant origin like papain do not show any similarity. The remarkable rate of autodigestion of thaumatin is another difference between thaumatin and, for example, papain. Unlike thaumatin, papain can be activated and kept fully enzymatically active for 4 h in cysteine (0.05 mol/l) [15]. Since we have demonstrated the enzymatic activity of thaumatin, there are two measurable biological activities : the sweetness and the hydrolytic activity. The question which arises is whether and how these two activities are correlated. In the series of thaumatins consisting of sweet-tasting proteins with the same molecular weight, sweetness intensity and amino acid composition, and which differ probably only in their number of carboxylic amide groups, thaumatin I has the lowest enzymatic activity. This activity increases with increasing as well as with decreasing basicity of the other thaumatins (Table 2). So the influence of a

6 418 H. van der We1 and W. J. Bel: Enzymatic Properties of Thaumatin and Monellin different number of carboxylic amide groups on the sweet-taste intensity is not obvious, while the influence on the enzymatic activity is remarkable. Thus the correlation between the sweetness and hydrolytic activity seems to be questionable. We already know that the introduction of acetyl groups into thaumatin, which decreases the basicity, reduces the sweet-taste intensity in correlation with the number of introduced acetyl groups [S]. However, there was a remarkable increase in hydrolytic activity with an increasing number of acetylated E-amino groups of the lysine residues in acetylated thaumatin I as well as in acetylated thaumatin 11. So the influence of the modified lysine residues on the sweetness intensity and the enzymatic activity is opposite. Monellin, the other sweettasting protein which is equally sweet as thaumatin on a molar basis, showed immunological cross-reactivity with thaumatin [16,17] (compare immunological cross-reactivity between enzymes having the same active centre such as papain and chymopapain [12], bovine trypsin and a-chymotrypsin [18] as well as stem and fruit bromelain [19]). This cross-reactivity indicates a common region in these molecules and it was tentatively concluded that these are the active sites responsible for the sweetness [16]. Nevertheless, neither native monellin nor monellin in a reducing environment showed enzymatic activity. Consequently there seems to be no correlation between the enzymatic activity and the sweet taste. That is why it is even more surprising that in monellin, enzymatic activity arises by acetylation of the &-amino groups of the lysine residues and, just as for thaumatin, a reducing medium is necessary. This again is a similarity between thaumatin and monellin. The similarities such as equal molar sweetness, immunological relationship and hydrolytic activity under certain conditions makes it of interest to in- vestigate whether a structural relationship is responsible for these phenomena. We like to thank Mr P. D. van Wassenaar for carrying out the peptide mapping, Mr G. Richters for carrying out the amino acid analysis and Mr J. N. Brouwer for his contribution to the discussions. REFERENCES 1. Van der Wel, H. & Loeve, K. (1972) Eur. J. Biochem. 31, Higginbotham, J. D. & Hough, C. A. M. (1977) in Sensory Properties of Foods (Birch, G. G., Brennan, J. G. & Parker, K. J.. eds) pp , Applied Sciences Publisher Ltd, London. 3. Korver, O., van Gorkom, M. & van der Wel, H. (1973) Eur. J. Biochem. 35, Iyengar, R. B., Smits, P., van der Wel, H., van der Ouderaa, F., van Brouwershaven, J. H., Ravenstein, P., Richters, G. Sr. van Wassenaar, P. D. (1979) Eur. J. Biochem. 96, Van der Wel, H. (1972) FEBS Lett. 21, Morris, J. A. & Cagan, R. H. (1972) Biochim. Biophys. Acta, 261, Frank G. & Zuber, H. (1976) Hoppe-Seyler s Z. Physiol. Chem. 357, Van der Wel, H. & Bel, W. J. (1976) Chem. Senses Flavor, 2, Gray, W. R. & Hartley, B. S. (1963) Biochem. J. 89, 59p. 10. Van der Ouderaa, F. J., De Jong, W. W. & Bloemendal, H. (1973) Eur. J. Biochem. 39, Kunitz, M. (1947) J. Gen. Physiol. 30, Arnon, R. & Shapka, E. (1967) Biochemistry, 6, Erlanger, B. F., Kokowsky, N. & Cohen, W. (1961) Arch. Biochem. Biophys. 95, Martin, C. J., Golubon, J. & Axelrod, A. E. (1958) Biochim. Biophys. Acta, 27, Kirsch, J. F. & Igelstrom, M. (1966) Biochemistry, 5, Van der Wel, H. & Bel, W. J. (1978) Chem. Senses Flavor, 3, Hough, C. M. & Edwardson, J. A. (1978) Nature (Lond.) 271, Sanders, M. H., Walsh, K. A. & Arnon, R. (1970) Biochemistry, 9, Iida, S., Sasaki, M. & Ota, S. (1973) J. Biochem. (Tokyo) 73, H. van der We1 and W. J. Bel, Unilever Research Vlaardingen, Postbus 114, NL-3130AC Vlaardingen, The Netherlands