45 Acta Chim. Slov. 1998, 45(1), pp. 45-57 (Received 28.1.1998) ENZYMES WITH A LOW MOLECULAR WEIGHT Michael Mattey1, Davina Simoes2, Angela Brown1 and Xiaolian Fan1 1Department of Bioscience and Biotechnology, University of Strathclyde, Todd Centre, 31, Taylor Street, Glasgow, G4 0NR, Scotland. 2University of Sunderland, School of Health Sciences, Chester Road, Sunderland SR1 35D, UK. Abstract. Two esterase enzymes have been isolated, one from Candida lipolytica and one from Bacillus stearothermophilus, which are characterised by an unusually small molecular weight. The Candida enzyme is 5.7 kDa, with 56 amino acid residues and the Bacillus enzyme is 1.57 kDa, with only 17 residues (1). In both cases the catalytic activity appears to depend on a bound metal ion, as shown by dialysis against chelating agents, ion replacement and inhibition by metal complexing agents. Specific activities are similar to reported esterase activities. The Candida esterase has a temperature optimum of 280 , as might be expected from a mesophilic organism, but it has a half-life of 2 hours at 500 . The esterase from B.stearothermophilus is thermophilic, but whereas the optimum growth temperature is 550 the enzyme optimum is about 1200. Both enzymes exhibit some substrate specificity. The Bacillus enzyme has no particular specificity towards the chain length of the substrate, but shows a marked activity towards the 2-position of triglycerides. The Candida enzyme shows both chain length specificity (optimum at butyl esters), as well as specificity towards the 1-position. Introduction The size of enzymes has been the subject of debate for many years; the question is often posed in the form “Why are enzymes large?” The text-book answer is usually given in terms of bond energies; biological systems operate with two broad classes of chemical bond, firstly “weak” bonds, such as van der Waals Forces (1 kcal/mole), hydrogen bonds (3 - 7 kcal/mole), ionic bonds (5 kcal/mole) and hydrophobic 46 interactions (-1 to -3 kcal/mole) and secondly “strong” covalent bonds (~20 kcal/mole) . The weak bonds are responsible for virtually all the higher orders of biochemical structure, as well as most biochemical interactions that demand a high degree of stereochemical specificity. As these bonds are no more than an order of magnitude greater than the thermal energy of the environment such weak interactions are readily disrupted by the kinetic energy present in living systems. Covalent bonds, on the other hand, are not readily broken at physiological temperatures, enzyme -catalysed reactions are required for such bonds to be rapidly made and ruptured. The rate at which an enzyme catalyses a chemical reaction is probably dependent on the rapidity of a reversible conformational change induced by interaction with a substrate; such conformational changes involve breaking and forming many weak bonds throughout the protein, with the result that a covalent bond is cleaved. Because weak bonds can be broken by the kinetic energy released during the binding of substrates and modulators to enzymes so they can be broken by the addition of kinetic energy in, for example, the increase in environmental temperature. These changes may reduce the catalytic function, or the different tertiary and quaternary structures at the new temperature might have a enhanced functional properties. Thus size, function and thermostability are closely linked in proteins. As the enzyme molecule becomes smaller the number of hydrogen bonds etc. falls, so that less catalytic activity but more stability might be expected. The purpose of this paper is to establish that the size range of naturally occurring proteins with catalytic activity is greater than hitherto thought, secondly that such enzymes exhibit thermal properties that are both consistent with their size and are adapted to the rigorous extracellular environment, and thirdly that their mechanism of action results in substrate and product specificity. There have been a few reports in the literature in which an enzyme activity has been attributed to proteins with a molecular weight of less than 10 kDa (“microenzymes”)(2). A lipase from bovine milk slime with an apparent molecular weight of 7 kDa was reported by Chandan & Shahani (3). The molecular weight was 47 determined by sedimentation velocity and osmotic pressure methods, and the activity exhibited a pH optimum of 9.2 and 370. A rennin from an unidentified thermophilic actinomycete, isolated from soil near Beer Sheva, was found to have a molecular weight of 10.5 kDa, based on analytical ultra-centrifugation and SDS electrophoresis (4). The amino acid composition was determined and the protein was found to have 78 residues, of which 9 were proline, giving a molecular weight of 9.7 kDa. The temperature optimum was 750, and it required calcium ions for activity. Limited proteolytic activity against insulin was noted. Another small proteolytic enzyme was reported by Steele et al. (5) isolated from a novel spiral bacterium, Kurthia spiro/orme. This gram-positive bacterium grew at neutral pH in a thermal spring, but exhibited a wide range of growth temperatures and pH values, ranging from 40 to 470 and a pH range from 7 to 11.5 with an optimum of 300 and pH 10.5. The extracellular protease was alkaline-stable and had an optimum temperature of 600 and optimum pH of 11. An amylase from Bacillus caldolyticus (6) was shown to have active subunits of less than 10 Kda, which associate in the presence of calcium ions to give the classic thermophilic amylase active at 700. The subunits of the thermophilic amylase exist when calcium is absent or low in concentration and are not thermophilic but are still thermostable, showing activity up to 400, but are stable up to 600, as shown by the recovery of activity when the temperature is reduced, or when calcium is added at a higher temperature. We have described two extracellular peptides with esterase activity, one produced by a strain of Candida lipolytica (CMI 92,743) (7) and one from Bacillus stearothermophilus (1). The esterase from the Candida species had a molecular weight of 5 kDa ± 500 as determined by both Sephadex G-100 gel filtration and SDS-polyacrylamide electrophoresis. The enzyme contained 56 amino acid residues, 13 of which were proline, giving a molecular weight of 5717 kDa. The Bacillus esterase was much smaller, with 15 amino acid residues, none of which was proline, giving a molecular weight of 1556 kDa. The composition of the two esterases is shown in Table 1. 48 One feature of the Candida esterase is the high percentage of proline, a feature also noted with the rennin from a thermophilic actinomycete, although at 23.7 mole% Amino acid Candida Bacillus Ala 5 1 Asp 3 3 Arg 1 1 Cys 1 - Glu 8 1 Gly 8 2 His 3 2 Ileu 1 - Leu 3 - Lys 1 - Pro 13 - Phe 1 - Ser 1 4 Thr 2 1 Tyr 1 - Val 4 - Total 56 15 Table 1 Amino acid composition of the extracellular esterases from Candida lipolytica and Bacillus stearothermophilus. Amino acid analysis was carried out on an Applied Biosystems 420H Amino acid Analyser using phenylthiocarbamyl derivitisation. Calibration used an internal standard of norleucine. the esterase has twice the proline content. Proline has been associated with increased protein thermostability (8, 9) on the basis of the different thermostability of five Bacillus oligo-1,6- glucosidases. Comparisons were made of aminoacid composition and structural parameters and from the analysis, in conjunction with the strong site specificity of proline residues for b-turns (10, 11, 12) it was proposed that enhanced stability could be gained by increasing the frequency of proline occurrence at b-turns and the total number of hydrophobic residues (8) This appeared to be given support by Matthews et al (13) where the thermostability of bacteriophage lysozyme was increased by replacing alanine with 49 proline at position 82 of a b-turn so as to decrease the backbone entropy of unfolding. Other results have correlated increased proline with increase thermostability with different enzymes (14, 15, 16) The correlation with both pullulanases and oligo-1.6-glucosidases between proline percentage and Tm (the temperature at which the enzyme was half inactivated in 10 minutes at pH6.8) was linear between 3% and 9% proline and 450 to 980. The Candida esterase does not fall on that line, although its Tm, at 700 is higher than expected for a mesophilic organism. The rennin (3) similarly does not coincide with the Bacillus data, but its stability (Tm 75 0) is greater than the optimum growth temperature (52 0). Mode of Action: The two esterases show different behaviour towards a range of inhibitors (Table 2) Addition Candida lipolytica Bacillus stearothermophilus None 100 100 NaN3 35 90 KCN 29 85 p-chloro-mercuribenzoate 29 79 CuSO4 100 0 FeSO4 100 80 1,10 phenanthroline 20 113 EDTA 30 89 Table 2. The effect of inhibitors on the activity of the esterases from Candida lipolytica and Bacillus stearothermophilus.The additions were made at 5mM. The Candida esterase shows a pattern of inhibition consistent with iron chelation, with the cytochrome oxidase inhibitor azide, p-chloro-mercuribenzoate and 1,10 phenanthroline being potent inhibitors, and general metal chelators such as EDTA 50 also effective. The Bacillus esterase shows slight inhibition with some of these inhibitors, but with 1,10 phenanthroline shows slight activation. The most striking inhibitor of the Bacillus enzyme was copper sulphate which had an effect considerably greater than other inorganic salts, for example ferrous sulphate. Inorganic salts are known to weaken electrostatic forces within the protein molecule (17), but the complete inhibition could not be attributed to this. Cu 2+ is known to form complexes with amino and amide groups, as in the classic Biuret test for proteins. If the enzymes are dialysed against EDTA then a loss of activity is observed in both cases (Fig 1). The rate of activity loss is similar in both cases although the temperature at which the Candida enzyme (280) and the Bacillus enzyme (700) were treated was different. 100 50 0 - 123 Ti m e o f di al ysi s (ag ai nst ED TA) Figure 1. The effect of dialysis on esterase activity. The enzyme preparations were dialysed against 1mM EDTA at 70 0 in the case of Bacillus and 28 0 in the case of Candida for four hours. The results are expressed as a percentage of the starting 0 4 51 activity as determined by fluorometric assay using fluorescein dibutyrate at 10-7M as substrate. Attempts to restore the activity were made with a number of cations, added at 5 mM to the dialysed solution. For the esterase from Candida lipolytica 90% of the original activity was restored by ferric ions, other ions such as zinc, ferrous ions, nickel and copper restored about 20%, while monovalent ions, calcium and magnesium either did not have any effect or further reduced the residual activity. For the Bacillus esterase the only ion found with any significant effect was Fe 3+, but only 40% activity was restored. When dialysis was carried out against water instead of EDTA the activity was lost more slowly, but 80% of the activity could be restored by mixing the dialysed enzyme with the dialysis solution, unlike ferric ions where again 40% activity was restored. Effect of pH. The two esterases showed quite distinct pH profiles, the Candida esterase showed a flat profile until pH values below 2, while the Bacillus enzyme showed an optimum at pH 9.0 which is above the growth optimum for the organism (pH 6 to 8 ). E ffe c t o f p H o n e n zym e a c tiv ity 10 0 90 80 70 60 50 40 30 20 10 0 *-- C a n d i d a • K u r t h i a pH 52 Figure 2. The effect of pH on esterase activity. In the range 2 to 4 acetate buffer was used, phosphate buffer from 5 to 7.5 and tris-HCl from 7.5 to 11. All buffer concentrations were 0.1M; activity was determined using flurorescein dibutyrate with a blank control at the same pH and buffer condition. The enzymes were stable to a range of pH values from 5 to 10 for several hours. Effect of Temperature. The effect of temperature on esterase activity is shown in Fig. 3. The two esterases show very different temperature optima as would be expected from a mesophilic and a thermophilic organism. In the case of the mesophilic enzyme the growth temperature and the enzyme optimum are the same at 28 0, but in the case of B. stearothermophilus the growth optimum is 55 0 , but the temperature optimum is 1200. 1 0.0 2 .5 0 .0 75 Tem per ature Figure 3. The effect of temperature on the activity of the esterases from C. lipolytica and B. stearothermophilus. For the Candida esterase the optimum temperature was determined at pH 6.0 using a fluorimetric assay, the sample was equilibrated for one minute before activity 0 25 50 10 0 12 5 15 0 53 measurements were made. The optimum for the Bacillus enzyme was determined by hydrolysis of tributyrin in a sealed tube for two hours, followed by titration of the acid liberated, using a control without enzyme to correct for thermal hydrolysis at elevated temperatures. The optimum temperature of 1200 is exceptionally high,and compares with that of the extreme thermophile Sulpholobus solfataricus (18) where 5’methylthio-adenosine-phosphorylase had an optimum of 120 0 and Pyrococcus furiosus protease at 115 0. The Arrhenius plot (Fig 4.) was discontinuous and concave upwards, a feature unusual in enzyme-catalysed reactions, although at temperatures above 120 0 rapid inactivation occurs . 1.30 1.15 1.10 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 1/T x 10 3 ( 0K-1) Figure 4. Arrhenius plot for B. stearothermophilus esterase. Such plots are observed when a reaction system consists of two parallel reactions with different efficiencies, where the activation energies are such that one dominates at a higher temperature and the other at a lower temperature (19). 54 The energy of activation (Ea) from 300 to 700 was 355 cal mole-1 and from 800 to 1200 was 948 cal mole -1. The thermostability of these small esterases is high, although the thermophilic enzyme is more stable, as might be expected Figure 5 shows the stability over a 100 hour period, with the Candida esterase having a half life of 2 hours at 500, very stable for a mesophilic organism, and the Bacillus esterase retained 95% of its activity at 700 at 100 hours. Even at 900 the enzyme had a half life of 12 hours. Increased thermostability is the sort of property that might be useful in the extracellular environment. Robust enzymes which persist will maximise the substrate returns on the investment in extracellular proteins. From the economic viewpoint a small thermostable extracellular enzyme is efficient providing its activity is similar to that of large enzymes. 10 0 90 80 70 60 50 40 30 20 10 25 50 Tim e (h) C andi da 5 0 B acillus 7 0 B acillus 8 0 B acillus 9 0 75 10 0 Figure 5 Thermostability of the esterases. Specificity. The two esterases differ both in their specificity towards triglycerides and towards the chain length of the fatty acid chains. 0 0 55 Figure 5 shows that the Candida esterase shows a preference for shorter chain lengths, with a peak at butyric although it is still effective against palmitic acid side chains with about 20% of the relative activity shown towards butyrate esters. This preference is seen both with fluorescein diesters and glycerol esters. Products of the hydrolysis of triglycerides were separated by HPLC on a Lichrosorb Si60 column with a mass detector (Sedex 55) using a mobile phase of toluene-hexane (1:1) (solvent A) and toluene-ethyl acetate (3:1) plus 1.2% formic acid (solvent B). A gradient of 1 - 50% B over 10 minutes and 50 - 100% B from 10 to 15 minutes, with 100% B from 15 to 40 minutes was used. The Candida esterase showed a peak of 2,3 diglyceride in the early stages of hydrolysis , with 1,3 diglyceride in small quantities appearing after all triglyceride was hydrolysed. The Bacillus esterase on the other hand showed 1, 3 diglyceride only. In both cases the diglyceride was slowly hydrolysed further to monoglyceride and butryric acid . 50 40 U 30 20 10 3 Candida Bacillus 6 8 10 12 14 16 Carbon chain length of triglycerides 18 0 2 4 56 Figure 6 Esterase specificity Conclusions. It is apparent that the lower limit on the size of naturally occurring enzymes is well below 10 kDa, the Bacillus esterase described here is only 1.57 kDa and is little more than a peptide. The range of activities so far seen is limited to simple hydrolysis, which is not unexpected given the small size; the scope for the binding of substrates with recognition between similar groups must be limited. Esterases, lipases and proteases have been described, but so far not carbohydrate or nucleic acid degrading activities. Despite the limited substrate binding possibilities both esterases show specificity towards triglycerides, and these enzymes, with their considerable thermostability and reasonable activity, are of interest in industrial processes. Stability can be further enhanced by immobilisation which retains activity. It is not known how widespread such enzymes are but preliminary screens show many thermophiles have enzyme activities in the micorenzyme size range. It is noteworthy that the two esterases described appear to have different mechanisms of activity. The larger esterase from C.lipolytica has an ferric iron active site, where the metal catalyses an acid hydrolysis of the substrate. This is demonstrated by the pH dependence and the effect of dialysis. The Bacillus esterase appears to use the metal ion to maintain the conformation of the peptide, but the pH optimum is not consistent with metal catalysis, nor is the effect of dialysis consistent with an active site metal ion, although the effect of chelators such as EDTA is consistent with the involvement of metal ions in an important functional role. The active site sequences for a number of esterases show great similarity with the sequence Gly-Glu-Ser*-Ala-Gly being conserved and Glu/Asp -Ser being very common in the next two positions. The amino acid composition of the microenzyme from B stearothermophilus contains the seven amino acids found in the active site 57 sequence although the sequence is not known. The peptide is N terminal blocked, possibly by a formyl group. If this hypothesis is substantiated the mechanism whereby a confomrational change in such a small protein could cleave the ester bond will be of interest. The possible number of hydrogen bonds etc. is small but the role of metal ions in stabilisation may be significant. References [I] D de C. M. Simoes, D. McNeill, B. Kristiansen, M. Mattey, Biotech. Lett. 1995, 17, 9, 953 - 958. [2] R. U. Schenk, J. Bjorksten, Finska Kemists. Medd. 1973, 82, 26 - 46. [3] R. C. Chandan, K. M. Shahani, J. Dairy Sci. 1963, 46, 275 - 283. [4] S. H. Laxer, A. Pinsky, B. Bartoov, Biotech. Bioeng. 1981, 23, 2483 - 2492. [5] D. B. Steele, M. J. Fiske, B. P. Steele, V. C. Kelley, Enz. Microb. Tech. 1992 , 14, 5, 358 - 360 [6] W. Heinen, A. M. Lauwere, Proc. Int. Symp. Enzymes and Proteins from Thermophilic Microorganisms Zurich 1975 78 - 89 (Birkauser Verlag, Basel und Stuttgart) [7] G. Adoga, M. Mattey, FEMS Microbiol. Lett. 1979, 6, 61 - 63. [8] Y. Susuki, K. Oishi, H. Nakano, T. Nagayama, Appl. Microbiol. Biotechnol., 1987, 26, 546 - 551. [9] Y. Susuki, Proc. Jpn. Acad. Ser. B. 1977, 65, 146 - 148. [10] P. Y. Chou, G. D. Fasman, J. Mol.Biol. 1977, 115, 135 - 175. [II] P.N. Lewis, F.A. Momany, H.A. Scheraga, Biochem. Biophys. Acta 1973 , 303, 211 -229. [12] M. Levitt, Biochemistry 1978 , 17, 4277 - 4285. [13] B. W. Mathews, H. Nicholson, W. J. Becktel, Proc. Natl. Acad. Sci USA 1987, 84, 6663 - 6667. [14] H. S. Tesfay, R. E. Amelunxen, I. D. Goldberg, Gene 1989, 82 , 237 - 248. [15] U. Karst, H. Schutte, H. Baydoun, H. Tsai, J.Gen Microbiol. 1989, 135, 1305 - 1313. [16] Y. Suzuki, K. Hatagaki, H. Oda, Appl. Microbiol. Biotechno.l 1991, 34, 707 - 714. [17] M. P. Brosnan, C. T. Kelly, M. W. Forgaty, Eur. J. Biochem. 1992, 203, 225 - 231. [18] G. Cacciapuoti, M. Porcelli, C. Bertoldo, M. Rosa, V. Zappia, J. Biol. Chem. 1994 , 269, 24762 - 24769. [19] M. H. Han, J. Theor. Biol. 1972, 35 , 543 - 568.