Protein hydrophobicity and stability support the thermodynamic theory of protein degradation

Btochlmlca et Btophyslca A~ta, 788 (1984) 17-22

17

Elsevter BBA31915

PROTEIN HYDROPHOBICITY AND STABILITY SUPPORT THE THERMODYNAMIC THEORY OF PROTEIN DEGRADATION DAVID F MANN, KAREN SHAH, DAVID STEIN and GARY A SNEAD

Department of Biochemistry, Chtcago College of Osteopathic Medicine, 1122 East 53rd Street, Chicago, IL 60615 (U S A ) (Received September 13th, 1983) (Revised manuscript recewed March 13th, 1984)

Key words Hydrophobtctty, Protem degradatton, Thermodynamic theory

Rat liver enzymes were used to study the relationship between their in vivo half-lives and their apparent hydrophobicity or their resistance to inactivation by mechanical shaking. The apparent hydrophobicity of these enzymes, measured as the percent of the protein recovered from an octyl-Sepharose column, is correlated with their known half-lives (r = 0.75, P < 0.01). The presence of specific ligands which are known to increase compactness by impeding unfolding of proteins decreased the apparent hydrophobicity of fructose-l,6-bisphosphatase, glucose-6-phosphate dehydrogenase, glyceraldehyde-3-phosphate dehydrogenase, and pyruvate kinase. Resistance of enzymes to inactivation by mechanical shaking correlated well with their in vivo half-lives (r --- 0.90, P < 0.01). When the shaking experiments were done in the presence of substrates, fructose-l,6-bisphosphatase, glucose-6-phosphate dehydrogenase, glyceraldehyde-3-phosphate dehydrogenase and lactate dehydrogenase were protected from inactivation.

Introduction The half-lives of water-soluble mammalian, bacterial and plant proteins appear to be dependent on their conformatlonal and physical properties [1-3]. Proteins with shorter half-hves are more likely to have the following characteristics: larger subunit molecular weight [4,5], lower isoelectnc point [6], affinity for hydrophobic surfaces [7-11], a greater susceptibility to protemases [12] a n d / o r a greater degree of instability as measured by thermal or acid treatment [13,14]. Proteins are known to exist in several conformational states which are in equilibrium (e.g., folded to unfolded or partially unfolded conformations) [15]. McLendon and Radany [15] have proposed that the rate of protein degradation is directly proportional to the amount of protein in the unfolded state. The properties of molecular weight, susceptibility to 0167-4838/84/$03 00 © 1984 Elsevier Science Pubhshers B V

proteinases and general stability are in fact associated with the rate of unfolding and therefore consistent with the thermodynamic theory of protein degradation [15]. Hydrophobicity has been shown to be associated with the half-life of cytoplasmic proteins by several groups of investigators. Bohley et al. [7] first reported that short-lived proteins were selectively accumulated in the apolar phase of phasepartition experiments, and these results were later corroborated by others using hydrophobic affimty chromatography [8-10]. Mann and Moreno [16] have recently shown that hydrophob~city, as measured by association with octyl-Sepharose, is related to the ease with which a protein unfolds. This paper shows that hydrophobicity and sensitwlty to protein denaturation via mechanical shaking of several proteins is correlated with the halflife of a protein as predicted by the thermody-

18 namlc theory of protein degradation. A preliminary report of some of these results has appeared previously [10]. Methods

B~ochemzcals. The Sepharose CL-4B and the octyl-subsututed Sepharose CL-4B were purchased from Pharmacla Free Chemicals. The reagents for the urea determination were purchased from American Hospital Supply Corporation. Unless specified, all other biochemicals were obtained from Sigma Chemical Company. Homogenate preparatton. Male Sprague-Dawley rats (150-300 g) were fasted for 16 h before being killed by decapitation. The livers were removed and homogenized m 5 vol. ice-cold 0.05 M TrisHC1 buffer, or 0.05 M potassium phosphate (pH 7.5), using a Potter-Elvehjem homogemzer with a Teflon pestle. The homogenate was centrifuged at 100000 × g for 10 min and the hplds which were on the surface were carefully removed using a Pasteur pipet. The supernatant was then decanted and centrifuged a second time at 100000 × g for 60 nun to remove the hydrophobic mlcrosomes. Chromatography The unsubst~tuted Sepharose or the octyl-Sepharose (non-charged) was packed in columns (5.0 × 1.1 cm) equihbrated with the homogenization buffer at room temperature. The 100000 × g rat liver supernatant (0.5 ml containing approx. 6.0 mg protein) was placed on the column and eluted with the homogenization buffer. The three tubes with the highest protein concentration as determined by the absorbance at 280 nm or using the Bio-Rad Protein Assay ~ procedure were pooled. Enzyme assays were done on both the pooled fractions and on the original 100000 × g supernatant. An amount of gel was used which absorbed 80% of the protein. Shaking expertments The 100000 × g rat liver supernatants were desalted by passage through a Sephadex G-25 column. The protein concentration was determined by the bluret procedure [17] and the column eluates were diluted so that the final protein concentration was always 2.0 m g / m l , since it was found that the rate of inactivation decreased as the protein concentration increased. The shaking experiments were done in a cold room (2°C) using a TCS shaker (Technical Consulting Services,

Southampton, PA) according to the modified method of Asakura et al. [18]. 2 ml of the freshly prepared 100000 × g supernatant were placed m vials (10 × 50 mm) and shaken for various time periods at a frequency of 30 Hz During the first 5 mIn of shaking, the temperature of the vial contents increased to approx. 14°C and remained there for the duration of the expenment. The enzymes were assayed after shaking. Enzyme assays The following enzymes were assayed as described m the indicated references. tyrosme aminotransferase (EC 2.6.1.5) [19], serlne dehydratase (EC 4.2.1.13) [20]: glucose-6-phosphate dehydrogenase (EC 1 1.1 49) [21]: pyruvate klnase (EC 27.1.40) [22]; fructose-1,-6-blsphosphatase (EC 3.1.3.11) [23]: aspartate aminotransferase (EC 2.6.1.1) [24]: alanme aminotransferase (EC 2.6.1.2) [25]; lactate dehydrogenase (EC 1.1.1.27) [26], malate dehydrogenase (EC 1.1.1.37) [27]: arginase (EC 3.5.3 1) [28]; catalase (EC 1.11 1.6) [29], aldolase (EC 4.1.2.13) [30], and glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1 12) [31] Results The apparent hydrophoblcIty of the enzymes studied, as measured by their recovery from octylSepharose columns was Inversely related to their half-life (r = 0.75, P < 0.01) (Table I). The hydrophoblclty of glyceraldehyde-3-phosphate dehydrogenase, glucose-6-phosphate dehydrggenase, fructose-l,6-bisphosphatase and pyruvate klnase was significantly decreased in the presence of specific ligands of these enzymes (Table II) Since NAD but not nicotinic acid adenine dinucleotide, a NAD analogue, altered the hydrophobiclty of glyceraldehyde-3-phosphate dehydrogenase, NAD probably has a specific effect on the apparent hydrophoblcity of t~s enzyme. The significance of the protective effects of the NAD analogue on pyruvate kinase are unknown (Table II) NADP significantly decreased the hydrophobiclty of both glucose-6-phosphate dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase whereas the hypoxanthme analogue of NADP had no effect on glyceraldehyde-3-phosphate dehydrogenase. The hypoxanthme analogue of N A D P also decreased appreciably the hydrophoblclty of both

19 TABLE I T H E HALF-LIVES A N D P E R C E N T RECOVERY F R O M O C T Y L - S E P H A R O S E C O L U M N S F O R T W E L V E R A T LIVER ENZYMES The values for the percent recovery from the octyl-Sepharose columns were normahzed by considering the recovery from the unsubst,tuted Sepharose column as 100% and adjusting the recovery from the octyl-Sepharose accordingly Each value is the average of four to twenty determinations The numbers m the parentheses md~cate the appropriate references Enzyme

Half-hfe (h)

Percent recovery from octyl-Sepharose +-SE

Tyrosme ammotransferase (2 6 1 5) Serme dehydratase (4 2 1.13) Glucose-6-phosphate dehydrogenase (1 1 1 49) Catalase (1 11 1 6) Pyruvate kmase (2 7 1 40) Fructose-l,6-blsphosphatase (3 1 3 11) Aldolase (4 1 2 13) Aspartate ammotransferase (2 6 1 1) Glyceraldehyde-3-phosphate dehydrogenase (1 2 1 12) Alanme ammotransferase (2 6 1 2) Lactate dehydrogenase (1 1.1 27) Malate dehydrogenase (1 1 1 40)

3 (36) 4 (37) 15 (37) 25 (38) 30 (37) 44 (39) 68 (40) 72 (37) 75 (40) 82 (41) 84 (40) 96 (37)

58 +- 2.2 12 + 1 7 58-+23 76+-8 6 45 +- 2 6 53-+3 2 71 +- 3 7 102+-4 2 83 -+ 3 2 89+-6 1 87 -+ 2 2 78 +_5.7

~ ~ ~ ~ ~ ~ "

These half-hves were determined under steady state condltmns

T A B L E II T H E E F F E C T OF SELECTED SUBSTRATES A N D A N A L O G U E S ON E N Z Y M E RECOVERY D U R I N G H Y D R O P H O B I C AFFINITY CHROMATOGRAPHY One octyl-Sepharose cQlumn was equdlbrated with potassium phosphate buffer, whale the second column was equlhbrated wxth the same buffer containing one of the substrates or analogues The enzyme preparation used on the second column also contained the substrate or analogue at the same concentration The chromatography with and without N A D was done at 37°C All other experiments were done at room temperature (21°C). Results are means + S.E for three or four experiments Abbrewatlons. F16DP, fructose-l,6-blsphosphatase; G6PD, glucose-6-phosphate dehydrogenase; G3PD, glyceraldehyde-3-phosphate dehydrogenase, PK, pyruvate l~nase, LDH, lactate dehydrogenase, and MD, malate dehydrogenase Column buffer

Percent recovery F16DP

G6PD

G3PD

PK

LDH

MD

Protein

Noaddxtlon 03mMNAD

165:56 28± 87

11+_51 11+29

425:34 72+ 98 a

15_+ 0.6 21_+ 35

73+ 76 86+_ 4.9

65+_ 5 6 6 8 + 2.9

14+0.7 155:09

Noaddltlon 0 3 m M nicotinic acid adenmedmucleotlde

5 2 + 85

32_+29

47+__144

2 9 ± 1.7

85+ 58

72+11.5

185:1.5

53_+ 1.1

35+142

49+ 52

55+ 92 a

75+ 12

84-+ 2 0

175:1.0

Noaddltlon 12mMNADP

45+_ 7 6 44+_ 4 1

34-+54 54+3.7 a

59+_ 9 9 76+_ 3 0 a

22-+ 3 9 28+- 4.2

92+_ 7.6 85+_ 8 9

-

17±0.9 18±19

Noadditlon 1 2 m M mcotmarmde hypoxanthme dmucleotlde phosphate

68+- 6 4

44+-38

49+-104

35+- 38

78_+ 1.2

71_+ 9 2

19_+18

87+_11.0

63+-64

37+-115

48+-133

72+-101

79+-17.3

20-+05

Noaddltlon 1 m M fructose-l,6-blsphosphate

58+- 5 7 78+- 3 2 a

47+_62 45+-29

64+_ 7 7 81_+ 3 8

18+- 1 0 28+- 2 8 a

86_+ 3.5 96+- 3 8

-

20+-03 23+_03

a These values represent slgmficantly less hydrophoblc~ty than the control ( P < 0 05) using the Kruskal-Walhs H test, a non-parametric statistic.

20 TABLE Ill HALF-LIVES A N D THE TIME R E Q U I R E D TO ACHIEVE 50% INACTIVATION OF EIGHT RAT LIVER ENZYMES AFTER MECHANICAL S H A K I N G AT 30 Hz The time required for 50% reactivation was determined by plotting the log of the percent inactivation with respect to the time each sample was shaken 3-5 t~me periods were used for each determination and each of the values reported here are the average of 6-20 such determinations The numbers m the parentheses indicate the approprmte references Enzyme

Half-hfe (h)

Tyrosme almnotransferase Serme dehydratase Glucose-6-phosphatedehydrogenase

3 (36) 4 (37) 15 (37)

Pyruvate klnase Fructose-l,6-bisphosphatase Glyceraldehyde-3-phosphate dehydrogenase

30 (37) 44 (39) 75 (40)

20.9 _+ 4 6 104 9 + 8 3 115 8 +_ 12 5

Lactate dehydrogenase Argmase

84 (40) 96 (42)

256 6 _+ 17 3 402 0 + 110 1

glucose-6-phosphate dehydrogenase and fructose1,6-bisphosphatase, affecting the former in nearly the same manner as had the N A D P . These results however, were not significant. Fructose 1,6-bisphosphate also decreased the apparent hydrophobicity of both fructose-l,6-blsphosphatase and pyruv~ite klnase (P < 0,05). Fructose 1,6-bisphosphate is a positive allosteric modifier of pyruvate kinase [32]. Similar hydrophobic affinity chromatography experiments using 2 mM pyridoxal phosphate in the column buffer did not

Time required for 50% macttvatlon (mm)_+S E 20 8 _+ 0 24+ 92 +

25 0 05 12

result m any significant differences m the apparent hydrophobiclty of aspartate ammotransferase, tyrosine ammotransferase, alanine aminotransferase or serine dehydratase (data not shown). Protein stability can be measured by following the rate of enzyme inacnvat~on by mechanical shaking. The time required for 50% mactivation by this method is correlated with the in vivo half-hves of the enzymes used (r = +0.90, P < 0.01) (Table III). Ad&tional shaking experiments showed that glucose-6-phosphate dehydrogenase, fructose-l,6-

TABLE IV THE EFFECT OF SUBSTRATES A N D COFACTORS ON THE INACTIVATION OF CERTAIN RAT LIVER ENZYMES Enzyme

Substrate or Cofactor

Shaking time for 50% reactivation (mm)___S E W~thout

Tyroslne armnotransferase Senne dehydratase Glucose-6-phosphate dehydrogenase Pyruvate kmase Fructose 1,6-blsphosphate Glyceraldehyde-3-phosphate dehydrogenase Lactate dehydrogenase

0 12 mM pyndoxal phosphate 0 6 mM pyndoxal phosphate 03mMNADP 1 mM fructose 1,6-blsphosphate 0 75 mM ADP 0 5 mM fructose 1,6-b~sphosphate 0.3mMNAD 03 mM N A D 0 05 M pyruvlc acid

22 5 + 0 22 + 87 + 14 7 -+

With substrate or cofactor 28~ 0 04 15 29

103 6 _+10 1310 _+18 281 _+26

a These values are calculated as stated in Table 1II and are the mean_+ S E of three separate expenments b Using the Student t-test this value is slgmficant at P < 0.05 No loss m activity after shaking for 180 nun

25 5 + 10 5 0 32 + 05 610 + 19 }' 6 5 -+ 13 75 -+ 32 500 _+ 80 5 h - ~ 7182 _+6288 b 41 + 59

21 bisphosphatase and glyceraldehyde-3-phosphate dehydrogenase or lactate dehydrogenase were slgmficantly stabilized by NADP, fructose 1,6-blsphosphate and NAD respectively (Table IV). These results are consistent with those obtained with hydrophobic affinity chromatography in which the apparent hydrophoblclty of all studied enzymes except lactate dehydrogenase was decreased in the presence of these substrates (Table II).

Discussion Our data support the tenets of the thermodynamic theory of protein degradation as proposed by McLendon and Radany [15] and others who have suggested spontaneous denaturation (unfolding) a n d / o r subumt dissociation as the rate-limitmg step in lntracellular protein degradation [1,2,33]. First, the half-lives of the twelve proteins studied are mversely correlated with apparent hydrophoblcity (Table I). Moreover, several of the proteins examined experience a decrease in their apparent hydrophobicity when chromatographed in the presence of their substrates (Table II). Begley et al. [43] have reported a decrease m apparent hydrophobicity with cobalamin binding to either intrinsic factor or transcobalamin II. However, Mann and Moreno [16] have m~rpreted these changes in apparent hydrophoblclty to be the result of a tighter protein structure, which may explain why proteins in association with their ligands are more stable and less susceptible to proteolys~s. Another measure of protein unfolding is the inactivation of enzymes by mechanical shaking [18]. The time required for inactivation (Table III) is correlated with the in vivo half-hfe of several proteins. A notable exception is tyrosine atmnotransferase winch for both octyl-Sepharose chromatography and the mechanical shaking exhibit a more stable nature than would be predicted from its In vivo half-life. The reason for this is unknown. In addition, the respective substrates of glucose-6-phosphate dehydrogenase, fructose-l,6bisphosphatase, glyceraldehyde-3-phosphate dehydrogenase and lactate dehydrogenase protected these enzymes from denaturation during shaking (Table IV). Pyridoxal phosphate did not protect either tyrosine aminotransferase or serine dehy-

drase from inactivation by mechanical shaking, nor did it alter the apparent hydrophoblcity of these enzymes as measured with octyl-Sepharose chromatography (data not shown). The tyrosine aminotransferase data are consistent with the resuits of Lee et al. [42], who found that the apoenzyme was inacnvated at the same rate m vivo as the holoenzyme form, suggesting that pyndoxal phosphate binding did not serve to stabilize the enzyme conformation. The majority of the serine dehydrase is present as the holoenzyme initially and it would seem unlikely to be affected by the presence or absence of additional coenzyme [45]. It is not clear why fructose 1,6-bisphosphate sigmficantly affected pyruvate kmase retennon on octyl-Sepharose but not the rate of enzyme denaturation via shaking. Retention of proteins on hydrophobic affinity resins was previously attributed to the presence of hydrophobic amino-acid residues on the surface of the protein [35,36]. Studies with cross-hnked proteins, hgands and sucrose strongly suggest that hydrophoblcity as measured by retention on octyl-Sepharose is actually due to protein unfoldmg [16]. Therefore the correlation between half-hfe and hydrophobicity suggests that hydrophoblcity is in the same class as three other characteristics singled out by McLendon and Radany [15] as being associated with unfolding and half-life, which include spontaneous denaturation, susceptibility to proteolync degradation and molecular weight. However, for the proteins studied here, there was no correlation found between the subunit molecular weight and the half-life or hydrophoblcity (data not shown). In conclusion, the data presented m this paper suggest that the ease with which a protein unfolds as measured by hydrophobicity and denaturation by shaking is a determining factor in protein turnover as predicted by the thermodynamic theory of protein degradation [15].

Acknowledgements Tins work was supported by the Chicago College of Osteopathic Medtcine. We wish to thank Drs. W. Farnsworth, D Norwell and L. Van Winkle for reading the manuscript and Mrs. Yvette E. Baker for her typing.

22

References 1 Goldberg, A L and Dice, J F (1974) Annu Rev Blochem 43, 835-869 2 Goldberg, A L and St John. A C (1976) Annu Rev Blochem 45. 747-803 3 Cooke, R J and Davies, D D (1980) Blochem J 192 499-506 4 Dehhnger, P J and Schlmke, R T. (1971) Btochem Biophys Res Commun 40, 1473-1480 5 Dice, J F , Dehhnger, P J and Schlmke, R T (1973) J B~ol Chem 248, 4220-4228 6 Dice, J F and Goldberg, A,L (1975)Proc Natl Acad Scl U.S A 72, 3893-3897 7 Bohley, P, IOrschke, H, Langner, J , Raemann, S, Wlederanders, B., Ansorge. S and Hanson, H (1976) m Intracellular Protein Catabohsm (Turk, V and Marks, N , eds.), pp 108-110, Plenum Press, New York 8 Bohley, P and Rlemann. S (1977) Acta Blol Med Germ 36, 1823-1827 9 Segal, H L, Rothstem, D M and Wmkler, J R (1976) Blochem Blophys Res Commun 73, 79-84 10 Mann, D F and Shah. K (1979) Fed Proc 38, 414 11 Dean, R T (1975)Blochem Btophys Res Commun 67, 604-609 12 Bond, J S (1971)Biochem Blophys Res Commun 43, 333-339 13 Bond, J S (1976) Blochlm Blophys Acta 451,238-249 14 Bond, J S (1975) m Intracellular Protein Turnover (Schlmke, R T and Katunuma, N,, eds), pp 281-294, Academic Press, New York 15 McLendon, G and Radany, E (1978)J Biol Chem 253, 6335-6337 16 Mann, D,F and Moreno, R,O (1984) Prep Blochem 14, m the press 17 Layne, E. (1957) Methods m Enzymology (Colowlck, S P and Kaplan, N.O, eds), Vol II1, pp 447-454, Academic Press, New York 18 Asakura, T, Hemdge, P L., Ghory, P K and Adachl, K (1977) J Blol Chem 252, 1829-1830

19 20 21 22 23 24 25 26 27 28 29

Dlamondstone, T I (1966)Anal B~ochem 16, 395 401 Peramo, C (1967)J Biol Chem 242, 3860-3867 Langdon, R G (1966) Methods Enzymol 9. 126 131 Pon, N G and Bondar, R J L (1967)Anal Blochem 19, 272-279 Kolb, H J (1975) Arch Biochem Blophys 170, 710-715 Amador, E and Wacker, W E C (1962) Chn Chem 8, 343 350 Swlck, R W, Barnstem, P L and Stange, J L (1965) J Blol Chem 240, 3334 3340 Kornberg, A (1955) Methods Enzymol 1,441-443 Ochoa, S (1955) Methods Enzymol 1,735-739 SchLmke, R T, (1962) J Blol Chem 237, 459-468 Beers, R F and Slzer, 1 W (1952) J Blol Chem 195,

133-140 30 Blostem, R and Rutter, W J (1963~ J B~ol Chem 238, 3280-3285 31 Byers, L D , She, H S and Alayoff, A (1979) Biochemistry 18, 2471-2480 32 Freeland, R A (1968) Life ScJ 7, Part II, 499-503 33 Dean, R T (1980)Fed Pro~, 30, 15-19 34 Shaltlel, S (1974) Methods Enzymol 34, 126 140 35 Morns, C J O R (1976) Trends Blochem Scl 1, N207-N208 36 Kenne~, F.T (1962)J Blol Chem 237, 3405-3498 37 Dice, J F and Goldberg, A L (1975) Arch Blochem B~ophy', 170, 213-219 38 Price. V,E, Sterhng, W R, Tarantola, V A, Hartley, R W, Jr. and Rechclgl, M, Jr (1962) J B l o l Chem 237, 3468-3475 39 Zaht~s, J G and Pltot. H C (1979) Arch B~ochem Blophys 194, 620 631 40 Kuehl, L and Sumslon, E N (1980) J Biol Chem 245, 6616 6623 41 Lee, K, Darke, P L and Kenney, F T (1977) J BJol Chem 252, 4958-4961 42 Schlmke, R T (1964)J Blol Chem 239, 3808-3817 43 Begley, J A , Heckman, SM and Hall, C A (1981) Blochem Blophys Res Commun 103 434-441 44 Hunter, J E and Harper, A E (1976) J Nutr 106, 653-664