Correlation of hydrogen exchange behaviour and thermal stability of lysozyme

J. Mol. Biol. (1983) I68, 687-692

LETTERS TO THE EDITOR

Correlation of Hydrogen Exchange Behaviour and Thermal Stability of Lysozyme The solvent exchange rates of individual indole NH hydrogens of tryptophan residues of lysozyme have been measured, by using 1H nuclear magnetic resonance spectroscopy, as a function of temperature in the presence of urea and following chemical modification. The results have been interpreted in terms of a low activation energy process which is not dependent on the thermal stability of the protein, and a higher activation energy process that is directly correlated with the thermal stability. The significance of these observations for an understanding of the dynamics of the protein is discussed. The rates at which labile hydrogens in globular proteins exchange with hydrogens of aqueous solvents are often several orders of magnitude slower than for unstructured polypeptides. Nevertheless, even for labile hydrogens which are indicated in crystal structures to be in interior regions of proteins, and apparently completely isolated from solvent, exchange does take place eventually. Fluctuations of the structure must take place to permit exchange of such buried hydrogens to occur, and even for incompletely buried hydrogens fluctuations may be important in determining rates of exchange. Thus, studies of the kinetics and mechanisms of the exchange process give insight into the dynamical properties of proteins (Hvidt & Nielsen, 1966; Englander et al., 1972; Woodward & Hilton, 1979; Barksdale & Rosenberg, 1982). Using IH nuclear magnetic resonance, the rates of exchange of individual hydrogens in a protein can be determined (Glickson et al., 1971; Campbell et al., 1975; Wagner & Wiithrich, 1979). Provided that nuclear magnetic resonances are assigned to specific residues, the exchange process may therefore be studied in detail at different sites in a protein molecule. Recently, we reported individual exchange rates of the tryptophan indole NH hydrogens in lysozyme from hen egg white (Wedin et al., 1982). The results showed that a single process, characterized by a well-defined activation energy, could not explain the temperature dependences of the rates. Instead, two distinct processes, characterized by different activation energies, were found to be involved in the exchange of each hydrogen. The relative importance of each process was found to vary with the hydrogen involved, and to depend on the conditions under which the exchange took place. In conjunction with thermodynamic data, it was proposed that the process of higher activation energy was associated with co-operative thermal unfolding of the protein, and that of lower activation energy with fluctuations of a more local nature (Wedin et al., 1982). Such an interpretation is fully consistent with bulk exchange data for lysozyme (Gregory et al., 1982). Studies of the exchange of individual amide hydrogens of bovine pancreatic trypsin inhibitor and of various derivatives and related proteins have revealed 0022-2836/83/230687-(}6 $03.00/[)

687

9 1983 Academic Pless Inc. (London) Ltd.

688

M. DELEPIERRE E T A L .

that the exchange rates are strongly correlated with the denaturation temperatures of the proteins, the rates decreasing as the thermal stability increases (Wiithrich el al., 1980). A multi-state model involving global fluctuations of hydrophobic clusters provides an explanation for this observation (Wiithrich et al., 1980), but it has also been suggested (Hilton & Woodward, 1979; Woodward & Hilton, 1980) that the correlation arises simply because of contributions to exchange resulting from co-operative unfolding. In this letter, the dependence of exchange rates on thermal stability is further explored by examining its effect on both of the distinct processes involved in the exchange of the tryptophan indole NH hydrogens of lysozyme. Urea, at a concentration of 3 M, was used to decrease the denaturation temperature of lysozyme, and chemical modification was used to produce a derivative of higher thermal stability. This derivative contains an ester cross-linkage between residues Glu35 and Trpl08 and results from oxidation of the protein by 12 (Imoto & Rupley, 1973). The nuclear magnetic resonance spectrum in the presence of 3 M-urea is only slightly different from that in its absence. Small chemical shift changes are consistent with localized perturbations resulting from specific binding of the urea to the protein rather than to an), large-scale conformational changes. This is consistent with the conclusions from X-ray diffraction studies of lysozyme crystals after soaking in urea solutions (Snape et al., 1974: Evans, 1982). Similarly, nuclear magnetic resonance and X-ray diffraction studies of the chemically modified protein indicate that only small conformational differences exist between this and the native protein (Bedell et al., 1975; Dobson, 1977: Cassels et al., 1978). Nuclear magnetic resonance studies of the temperature dependence of the spectra show, however, that the mid-point of denaturation in 2H20 at pH 3"8 is lowered from 73"7_0-3~ to 64-6_2-0~ by the presence of 3 M-urea, but increased to above 85~ by the chemical modification. Similarly, the value of the apparent enthalpy of denaturation derived from the temperature dependence of the reversible denaturation equilibrium (Wedin et al., 1982) was reduced to 76"7 _+5"0 kcal mol- 1 from 119_+5 kcal mol- 1 by the presence of the urea. A value for the modified protein was not determined; the denaturation is not completely revei~ible in this case (Imoto & Rupley, 1973). Exchange rates of the indole NH hydrogens were all measured at pH 3"8 as described (Wedin et al., 1982). Values are reported here for Trpl08, T r p l l l and Trp123 as accurate measurements were possible over a wide temperature range for these residues. Figure 1 shows the results in the presence of 3 M-urea. As for lysozyme in the absence of urea, the exchange kinetics cannot be described by a single activation energy over the whole temperature range (here l0 ~ to 63~ but the experimental data fit well to a model with two processes, each characterized by a specific activation energy (Fig. 1). Comparison between the exchange rates with and without urea (Fig. 2) shows that at temperatures below about 45~ the presence of 3 M-urea has a negligible effect on the exchange behaviour. For Trpl 11 and Trpl23 the exchange rates are marginally faster in the presence of urea. This may be because the K w value is increased in the presence of urea. (Woodward el al., 1975); the exchange rate of tryptophan itself in 3 M-urea was measured at 72~ to be 6"8_+0"5 s -1, just over a factor of two faster than in the absence of urea

L E T T E R S TO T H E E D I T O R

01

--d

689

--

--I

--5--

--6

--7

--

--

2-8

I

2"9

I

5"0

I

5"1

I

5"2

I

5'3

I

5"4

I

5,5

5"6

io'/t Ft(:. 1. Temperature dependence of rate constants ( s - ' ) for exchange of the indole NH hydrogens of Trpl08 (+), Trpl 11 (O) and Trpl23 (O) of lysozyme in 2H20 at pH 3"8 in the presence of 3 M-urea. Experimental points have an uncertainty of about _ ()'10 log units. The least-squares fits of the data to a two-process model (see text) are indicated by continuous lines. The activation energies for the high activation energy process obtained h'om the fits were 77 to 86 kcal mol- t in the presence of 3 M-urea compared with 92 to 97 kcal mol-1 in the absence of urea (Wedin el al., 198.2).

where the rate is 2"8• s-1 (Wedin et al., 1982). For Trpl08, the exchange rates are very slightly slower in the presence of urea. This is attributed to the binding of urea in the vicinity of this residue; in crystalline lysozyme urea molecules have been found to bind to residues 107 and 109 (Evans, 1982). Comparison between the exchange rates in the chemically modified lysozyme with those in the unmodified protein is also possible from Figure 2. The exchange from residue 108 is considerably slower after modification, but for this residue the chemical structure as well as the environment of the indole ring has been altered. The exchange rates for Trpl 11 and Trp123, however, are at temperatures below about 55~ essentially the same in the two cases. At these lower temperatures, the exchange is dominated by the low activation energy process. The data therefore show clearly that there is no correlation between thermal stability and exchange of hydrogens by this process. Any fluctuations involved in the exchange are not involved in the co-operative unfolding process. As it has recently been found that individual exchange rates at these lower temperatures are similar in lysozyme crystals to those measured in solution (G. A. Bentley, M. Delepierre, C. M. Dobson, S. A. Mason, F. M. Poulsen & 1%. E. Wedin, unpublished results) these fluctuations are likely to be highly localized and of limited amplitude. Examination of the data given in Figure 2 reveals that at higher temperatures the exchange rates of each hydrogen are markedly increased by the presence of 3 M-urea. Under these conditions the high activation energy process is dominant and for each residue a given exchange rate is observed to occur at a temperature 9 to 12~ lower in the presence of 3 M-urea. As noted above, the presence of the urea causes the denaturation temperature to be decreased by about 9~ For the

M. D E L E P I E R R E

690

ETAL.

(o)

o~

-4-

+

.J

+

-5-

4-

+

-6I

I

I

I

I

I

I

+k~0

(b)

-2

~

~

-5 -4 .J

-5 -6 -7 -I

I

I

I

I

I

I

I

(cl -2 -5 -4 ~ o

-5 -6 -7 2-8

I

2.9

I

3-0

I

3.1

I

3-2 103/t

I

5"3

I

3.4

I

3-5

5-6

Fit:. 2. ('oml)arison of the data given ill Fig. I (O) tbr (a) Trpl08, (b) T r p l l l and (c) Trp123 with similar data obtained with the modified lysozyme (+). The continuous lines here represent the best fit of the data for the unmodified lysozyme described previously (Wedin et al., 1982).

modified protein, the displacement of the high activation energy process is in the opposite direction; this can be seen for T r p l l l in Figure 2. The correlation of the high activation energy process with thermal stability is in accord with a model which interprets this process as the total co-operative unfolding of the protein (Woodward et al., 1975). If it is assumed that exchange from the unfolded form takes place with a rate identical to that of model compounds (e.g. tryptophan), the observed exchange rate at high temperature under the conditions described here would correspond simply to the rate of unfolding (Wedin el al., 1982). On this model, the correlation with thermal stability would indicate that the rate of unfolding increases as the thermal stability decreases. There is, however, evidence that this picture is not complete.

LETTERS TO THE EDITOR

691

For example, the exchange rates and activation energies should be identical for different residues. Whilst this is a reasonable approximation at the highest temperatures (see, for example, Fig. 1) values do differ slightly for different residues (Wedin el al., 1982; Delepierre, 1983). Further, total unfolding of the modified protein would result in irreversible breakage of the ester linkage between Glu35 and Trpl08. For residue 108 in the modified protein, a high activation energy process contributes to exchange at the highest temperatures at which exchange was measured (Fig. 2). The nuclear magnetic resonance spectrum, however, shows t h a t no such breakage occurs under these conditions. This suggests t h a t exchange can take place from states t h a t are not completely unfolded, y e t which are related to the process of co-operative unfolding. This conclusion is also consistent with measurements of the t e m p e r a t u r e and pH dependence of the amide hydrogens of lysozyme (Delepierre, 1983). The experiments described here provide i m p o r t a n t confirmation of the significance of unfolding processes in hydrogen exchange, and suggest a means of testing whether or not these processes are i m p o r t a n t under specified conditions. I t is hoped t h a t extension of such studies will provide information about the mechanisms of protein folding and unfolding. This work is supported by the Science and Engineering Research Council. Fellowships are acknowledged from the Royal Society and C.N.R.S. (to M.D.), from the National Science Foundation and the Danforth Foundation (to R.E.W.), and h'om the Carlsberg Foundation (to F.M.P.). One of us (('.M.D.) is a member of the Oxibrd Enzyme Group. Inorganic (!hemistkw Laboratory South Parks Road Oxford OX1 3QR England Department of ('hemistry (!arlsberg Laboratory Gamle Carlsbergvej ll) DK-2500 Copenhagen Valby Denmark

M. DELEP! Eft.RE

C. M. DOBSO.X" ~. SELVARAJAH R. E. WED[N

F. M. POt'I,SEN

Received 24 March 1983 REFERENCES Barksdale. A. D. & Rosenberg, A. (1982). Melh. Biochem. Anal. 28, l-i 13. Beddel], C. R., Blake, C. ('. F. & Oatley, S. J. (1975). J. Mol. Biol. 97,643-654. Campbell. I. D., Dobson, ('. M. & Williams, R. J. P. (1975). Proc. Roy. Soc. ser. B. 189, 4855O2. Cassels, R., Dobson, C. M.. Poulsen, F. M. & Williams, R. J. P. (1978). Eur. J. Biochem. 92, 81-97. Delepierre, M. (1983). Th~se d'Etat. Universitd de Paris-Sud. Dobson, C. M. (1977). In N M R in Biology (Dwek, R. A., Campbell, I. D., Richards, R. E. & Williams. R. J. P., eds). pp. 63-94, Academic Press, London. Englander, S. W., Downer, N. W. & Teitelbaum, H. (1972). Annu. Rev. Biochem. 41,903924. Evans, S. T. (1982). Part II thesis. Oxford University.

692

M. D E L E P I E R R E ETAL.

Glickson, J. D., Phillips, W. D. & Rupley, J. A. (1971). J. Amer. Chem. Soc. 93, 4031-4038. Grego,-y, R. B., Knox, D. G., Percy, A. J. & Rosenberg, A. (1982). Biochemist~T], 21, 65236530. Hilton, B. D. & Woodward, C. K. (1979). Biochemistry, 18, 5834-5841. Hvidt, A. & Nielsen, S. O. (1966). Advan. Protein Chem. 21, 287-386. Imoto, T. & Rupley, J. A. (1973). J. Mol. Biol. 80, 657-667. Snape, K. W., Tjian, R., Blake, C. C. F. & Koshland, D. E. (1974). Nature (London,), 250, 295-298. Wagner, G. & Wiithrich, K. (1979). J. Mol. Biol. 134, 75-94. Wedin, R. E., Delepierre, M., Dobson, C. M. & Poulsen, F. M. (1982). Biochemistry, 21, 1098-1103. Woodward, C. K. & Hilton, B. D. (1979). Annu. Rev. Biophys. Bioeng. 8, 99-127. Woodward, C. K. & Hilton, B. D. (1980). Biophys. J. 32, 561-572. Woodward, C. K., Ellis, L. M. & Rosenberg, A. (1975). J. Biol. Chem. 250, 432-439. Wiithrich, K., Wagner, G., Richarz, R. & Braun, W. (1980). Biophys. J. 32, 549-558. Edited by G. A. Gilbert