Commentary

Do ultrastable proteins from hyperthermophiles have high or low conformational rigidity?

Rainer Jaenicke*

Institute of Biophysics and Physical Biochemistry, University of Regensburg, D-93040 Regensburg, Germany

ife on earth has an unbelievable adaptive parisons of their protein inventories with Lcapacity. Except for centers of volcanic those of suitable mesophilic counterparts, a activity, the entire surface of our planet is a wealth of data has been accumulated that biosphere. In this context, the most surpris- indicated that stabilization involves all levels ing discovery in our lifetime was the expan- of the hierarchy of protein structure, i.e., sion from the anthropocentrically defined secondary, supersecondary, tertiary, and ‘‘normal temperature’’ of mesophiles quaternary interactions. The common con- (Ͻ40°C) to the optimum temperature range clusion from model studies was that the of hyperthermophiles around and above the stability of proteins from extremophiles is boiling point of water. That in this class of optimized to maintain corresponding func- microorganisms high temperature is re- tional states under a given set of environ- quired for growth rather than tolerated im- mental conditions. For the standard state at plies that the whole repertoire of their bi- 25°C, enhanced thermal stability of hyper- omolecules must be sufficiently stable to thermophile proteins would then be the allow the cellular microcosm to work. The result of enhanced conformational rigidity Fig. 1. Three-dimensional structure of rubre- strategies nature has used to stabilize the in their folded native state (5). doxin from P. furiosus. Numbered residues mark inventory of the cell, especially proteins, Evidence from recent amide hydrogen the most slowly exchanging hydrogens, close to under extreme conditions are still enig- exchange experiments reported in this is- the two cysteine knuckles (7–9). matic, despite 25 years of active research. sue of PNAS seems to indicate that this What has become clear is that proteins, hypothesis has no general validity. Using independent of their mesophilic or extremo- rubredoxin from the archaeon Pyrococcus Because of the extremophilic character- philic origin, consist exclusively of the ca- furiosus, Herna´ndez et al. (6) report that istics and the presence of a metal-ion clus- nonical 20 natural amino acids; if other conformational opening processes for sol- ter, it is not surprising that there is no way to protein constituents are found, they origi- vent access occurs within milliseconds for obtain quantitative thermodynamic data the nate from covalent chemical modifications all amide positions along the 53-residue easy (calorimetric) way. Therefore, denatur- (1). Thus, enhanced stability can come from polypeptide chain. Considering first the ation kinetics were used to compare the only improved attractive forces: within the object the authors have chosen, Pyrococ- hyperthermophilic protein with its thermo- core, between domains and subunits, or cus rubredoxin seems to be ideally suited philic and mesophilic homologs. Indeed, the from extrinsic protectants such as compat- for the project: the protein contains three unfolding rates of the hyperthermophilic ible solutes, conjugating components, and antiparallel ␤-strands joined by two loops, protein and its mesophilic counterpart differ specific metabolites. The second alternative each containing two cysteine residues li- by two to three orders of magnitude. Re- has been explained in terms of preferential ganded to a metal ion to form C-X-X-C- lated studies with hyperthermophilic dihy- solvation or specific ligand binding (2). For X-X ‘‘knuckles’’ (Fig. 1). In the present drofolate reductase (10) and the 7-kDa the first, a variety of ‘‘rules’’ has been pro- study, for technical reasons, the normally cold-shock protein (Csp) from a mesophile, ϩ posed without providing an unambiguous occurring paramagnetic Fe2 was re- a thermophile, and a hyperthermophile (11) ϩ solution to the problem. The reason for this placed by Zn2 , essentially without any gave similar results. In the latter case, ther- failure is simple: considering the thermody- effect on the native three-dimensional modynamic data confirmed the kinetic anal- namic characteristics of proteins from me- structure. With its conformation unal- ysis; the differences in the equilibrium sta- ⌬ sophiles and extremophiles, the free ener- tered over a wide temperature and pH bilities ( Gstab) were found to increase with gies of stabilization differ only marginally. range, together with an extrapolated de- increasing optimal growth temperature The adaptive changes in terms of free en- naturation temperature close to 200°C (Topt;Fig.2a). The corresponding activation ⌬⌬ ergy differences ( Gstab) correspond to and an estimated global unfolding rate of energies for the unfolding reaction differed Ϫ Ϫ the equivalent of just a few weak intermo- unfolding of 10 6 s 1 at 100°C, the protein by Ϸ7 and 16 kJ⅐molϪ1 (Fig. 2b, open sym- lecular interactions (3). Given the large represents the most thermostable system bols). Interestingly, the refolding kinetics number of small increments from hydrogen presently known (9). Regarding the exper- were shown to differ only slightly, suggesting bonds, as well as hydrophobic, coulombic, imental approach, the protein allows all that the activated states of folding for all and van der Walls interactions in molecules advantages of amide hydrogen exchange three proteins are similarly native-like in with hundreds or even thousands of atoms, measurements to be exploited. Using the their interactions with the solvent (Fig. 2b, there may be an astronomical number of 15N-labeled protein to monitor local con- closed symbols). ⌬⌬ combinations yielding Gstab as a small formational fluctuations at 3–53°C, the difference between large numbers (4). authors succeeded in observing the open– Based on the complete genome sequences closed conformational transitions for all See companion article on page 3166. of hyperthermophiles and systematic com- amide hydrogens of the protein. *To whom reprint requests should be addressed.

2962–2964 ͉ PNAS ͉ March 28, 2000 ͉ vol. 97 ͉ no. 7 Downloaded by guest on September 29, 2021 Fig. 2. Unfolding/folding mechanism of the Csp from Bacillus subtilis, Bacillus caldolyticus, and Ther- motoga maritima with Toptϭ 52, 72, and 90°C, respectively. (a) Guanidinium chloride (GdmCl)-induced equilibrium unfolding transitions at 25°C, monitored by intrinsic fluorescence. Least-squares fit according to two-state model U 7 N yields ⌬Gstab ϭ 11.3, 20.1, and 26.2 kJ/mol for B. subtilis Csp, B. caldolyticus Csp, and T. maritima Csp, respectively. (b) Unfolding kinetics (open symbols) and refolding kinetics (closed symbols) of B. subtilis Csp (‚, Œ), B. caldolyticus Csp (ᮀ, ■), and T. maritima Csp (E, F), respectively. The apparent rate constant ␭ is plotted against the denaturant concentration. Fits according to the two-state model (5, 11). [Reproduced with permission from ref. 11 (Copyright 1998, Nat. Struct. Biol.).] Fig. 3. Hydrogen–deuterium exchange of glyc- eraldehyde phosphate dehydrogenase from T. ma- ritima (open symbols) and rabbit (closed symbols), The most remarkable findings for Pyro- chain. For a mumber of ultrastable proteins, measured at pH 6.0 (E, F) and pH 7.0 (‚, Œ), plotted coccus rubredoxin are the following. (i) The X-ray analysis and classical hydrogen– as relaxation spectra. (a) Measurements at 25°C; native hydrogen bonding of the protein deuterium exchange kinetics clearly indi- increased X values reflect increased rigidity. (b) close to its temperature of maximal thermo- cated anomalous conformational rigidity Measurements for the rabbit enzyme at 25°C and for the Thermotoga enzyme at 68°C. Coincidence dynamic stability is disrupted for all amide paralleled by decreased biological activity. of the curves indicated similar flexibility (5). [Re- hydrogens in less than a second. (ii) At 28°C, As an example, the upper profile in Fig. 3a produced with permission from ref. 5 (Copyright conformational fluctuations for solvent ac- illustrates the slow hydrogen–deuterium ex- 1998, Curr. Opin. Struct. Biol.).] cess occur in the millisecond time range or change of an ultrastable protein from Ther- faster through the entire protein. (iii)At motoga maritima compared with the lower alkaline pH, the activation energy analysis curve for its mesophilic counterpart from terations of its dynamics because of local of hydrogen bonded amides yields maxi- rabbit, both measured at constant (low) structural changes (18). mum enthalpy contributions of less than 5% temperature; coming closer to physiological On the other hand, also the ‘‘rigidity

of the activation enthalpies commonly ob- conditions for both proteins (Fig. 3b), ex- hypothesis’’ must not be taken dogmati- COMMENTARY served for protein unfolding (12). (iv) The change rates become superimposable in cally. For instance, in the case of phage T4 corresponding distribution of amide protec- agreement with Somero’s “corresponding lysozyme, a topological separation of tion factors is indistinguishable from data states” concept (13). Regarding the mech- functional and stabilizing regions was ob- reported for typical mesophilic homologs. anisms of orthologous enzymes with differ- served—functional amino acids being op- (v) The slowest local unfolding process in- ent temperature optima, it is interesting to timized for flexibility rather than stability; volves a tertiary hydrogen bonding interac- note that their active site residues are always however, mutating residues involved in tion, in accordance with similar configura- conserved such that differences in the cat- catalysis or substrate binding may enhance tions in other thermostable proteins. alytic properties must be caused by substi- stability at the cost of reduced activity, For those who have been working in the tutions elsewhere in the molecules (5, 14, supporting a relationship between stabil- field for some time, these findings are un- 15). Analyzing the stabilities and activities of ity and function (19). In other cases, part expected, because in short, they seem to large numbers of random mutants, it turns of the polypeptide backbone (predomi- show that moderate-scale conformational out that the two properties are not neces- nantly secondary structure) was found to form a stable scaffold while more flexible dynamics in the millisecond time range are sarily inversely correlated. However, muta- regions are involved in catalysis (15). A ubiquitous throughout the polypeptide tions that increase thermostability while drastic exception has been reported for a chain of the most thermostable globular maintaining low-temperature activity are number of thermophile enzymes with un- protein presently known. The critical com- extremely rare (16, 17). stable metabolic intermediates. Here, na- ment of the authors questioning ‘‘the com- What can be done to resolve the different ture suspended the corresponding-states mon hypothesis that enhanced conforma- views? Experienced protein chemists know concept for good reasons, because only tional rigidity in the folded native state that proteins show individual features, like extremely high catalytic activity (even at underlies the increased thermal stability of different species; thus, more examples need low temperature) can salvage short-lived hyperthermophile proteins’’ might be to be studied to draw general conclusions. substrates or intermediates (20). viewed by some as going a bit too far, having Apart from this one swallow does not a Thus, as usual in nature, we are left with in mind the catalog of stabilizing increments summer make argument, the chemist won- a whole spectrum of different solutions to that has been worked out during the past ders whether even rubredoxin under the our problem. In connection with the decade (5). Based on a vast amount of extreme experimental conditions might be present rubredoxin study, it would be most structural and thermodynamic data, the damaged, e.g., by oxidation of cysteine res- desirable to see more such data under view is now widely accepted that the anom- idues or perturbation of the metal-ion clus- conditions closer to physiological ones to alous stability of hyperthermophile proteins ter; similarly, the physicochemist might ar- come closer at last to the solution of one correlates with strong local interactions gue that, close to pH 12, the protein has 14 of the most challenging problems in phys- and/or improved packing of the polypeptide negative charges that may cause drastic al- ical biochemistry today.

Jaenicke PNAS ͉ March 28, 2000 ͉ vol. 97 ͉ no. 7 ͉ 2963 Downloaded by guest on September 29, 2021 1. Jaenicke, R. (1998) Biochemistry (Moscow) 63, 9. Hiller, R., Zhou, Z. H., Adams, M. W. W. & Structure (London) 6, 769–781. 312–321. Englander, S. W. (1997) Proc. Natl. Acad. Sci. USA 16. van den Burg, B., Vriend, G., Veltman, O. R., 2. Carpenter, J. F., Clegg, J. S., Crowe, J. H. & 94, 11329–11332. Venema, G. & Eijsink, V. G. H. (1998) Proc. Natl. Somero, G. N. (1993) Cryobiology 30, 201–241. 10. Dams, T. & Jaenicke, R. (1999) Biochemistry 38, Acad. Sci. USA 95, 2056–2060. 3. Dill, K. A. (1990) Biochemistry 29, 7133–7155. 9169–9178. 17. Giver, L., Gershenson, A., Freskgard, P.-O. & 4. Jaenicke, R. (1991) Eur. J. Biochem. 202, 715–728. 11. Perl, D., Welker, C., Schindler, T., Schro¨der, K., Arnold, F. H. (1998) Proc. Natl. Acad. Sci. USA 95, 5. Jaenicke, R. & Bo¨hm, G. (1998) Curr. Opin. Struct. Marahiel, M. A., Jaenicke, R. & Schmid, F. X. 12809–12813. Biol. 8, 738–748. ) Nat. Struct. Biol. 5 (1998 , 229–235. 18. Cavagnero, S., Debe, D. A., Zhou, Z. H., Adams, 6. Herna´ndez, G., Jenney, F. E., Jr., Adams, 12. Pfeil, W. (1998) Protein Stability and Folding: A M. W. W. & Chan, S. I. (1998) Biochemistry 37, M. W. W. & LeMaster, D. M. (2000) Proc. Natl. Collection of Thermodynamic Data (Springer, Ber- 3369–3376. Acad. Sci. USA 97, 3166–3170. lin). 19. Shoichet, B. K., Baase, W. A., Kuroki, R. & 7. Day, M. W., Hsu, B. T., Joshua-Tor, L., Park, J. B., 13. Somero, G. N. (1978) Annu. Rev. Ecol. Syst. 9, Zhou, Z. H., Adams, M. W. W. & Rees, D. C. 1–29. Matthews, B. W. (1995) Proc. Natl. Acad. Sci. USA (1992) Protein Sci. 1, 1494–1507. 14. Field, P. A. & Somero, G. N. (1998) Proc. Natl. 92, 452–456. 8. Blake, P. R., Park, J. B., Zhou, Z. H., Hare, D. R., Acad. Sci. USA 95, 11476–11481. 20. Sterner, R., Kleemann, G. R., Szadkowski, H., Adams, M. W. W. & Summers, M. F. (1992) 15. Auerbach, G., Ostendorp, R., Prade, L., Korndo¨r- Lustig, L., Hennig, M. & Kirschner, K. (1996) Protein Sci. 1, 1508–1521. fer, I., Dams, T., Huber, R. & Jaenicke, R. (1998) Protein Sci. 5, 2000–2008.

2964 ͉ www.pnas.org Jaenicke Downloaded by guest on September 29, 2021