NMR Elucidation of Early Folding Hierarchy in HIV-1 Protease

Abstract

Folding studies on proteases by the conventional hydrogen exchange experiments are severely hampered because of interference from the autolytic reaction in the interpretation of the exchange data. We report here NMR identification of the hierarchy of early conformational transitions (folding propensities) in HIV-1 protease by systematic monitoring of the changes in the state of the protein as it is subjected to different degrees of denaturation by guanidine hydrochloride. Secondary chemical shifts, HN-Hα coupling constants, 1H-15N nuclear Overhauser effects, and 15N transverse relaxation parameters have been used to report on the residual structural propensities, motional restrictions, conformational transitions, etc., and the data suggest that even under the strongest denaturing conditions (6 m guanidine) hydrophobic clusters as well as different native and non-native secondary structural elements are transiently formed. These constitute the folding nuclei, which include residues spanning the active site, the hinge region, and the dimerization domain. Interestingly, the proline residues influence the structural propensities, and the small amino acids, Gly and Ala, enhance the flexibility of the protein. On reducing the denaturing conditions, partially folded forms appear. The residues showing high folding propensities are contiguous along the sequence at many locations or are in close proximity on the native protein structure, suggesting a certain degree of local cooperativity in the conformational transitions. The dimerization domain, the flaps, and their hinges seem to exhibit the highest folding propensities. The data suggest that even the early folding events may involve many states near the surface of the folding funnel. Folding studies on proteases by the conventional hydrogen exchange experiments are severely hampered because of interference from the autolytic reaction in the interpretation of the exchange data. We report here NMR identification of the hierarchy of early conformational transitions (folding propensities) in HIV-1 protease by systematic monitoring of the changes in the state of the protein as it is subjected to different degrees of denaturation by guanidine hydrochloride. Secondary chemical shifts, HN-Hα coupling constants, 1H-15N nuclear Overhauser effects, and 15N transverse relaxation parameters have been used to report on the residual structural propensities, motional restrictions, conformational transitions, etc., and the data suggest that even under the strongest denaturing conditions (6 m guanidine) hydrophobic clusters as well as different native and non-native secondary structural elements are transiently formed. These constitute the folding nuclei, which include residues spanning the active site, the hinge region, and the dimerization domain. Interestingly, the proline residues influence the structural propensities, and the small amino acids, Gly and Ala, enhance the flexibility of the protein. On reducing the denaturing conditions, partially folded forms appear. The residues showing high folding propensities are contiguous along the sequence at many locations or are in close proximity on the native protein structure, suggesting a certain degree of local cooperativity in the conformational transitions. The dimerization domain, the flaps, and their hinges seem to exhibit the highest folding propensities. The data suggest that even the early folding events may involve many states near the surface of the folding funnel. The folding of a protein is conceptually described in terms of a folding funnel (1Dill K.A. Chan H.S. Nat. Struct. Biol. 1997; 4: 10-19Crossref PubMed Scopus (2026) Google Scholar, 2Plotkin S.S. Onuchic J.N. Q. Rev. Biophys. 2002; 35: 111-167Crossref PubMed Scopus (178) Google Scholar, 3Baldwin R.L. Rose G.D. Trends. Biochem. Sci. 1999; 24: 77-83Abstract Full Text Full Text PDF PubMed Scopus (382) Google Scholar, 4Baldwin R.L. Science. 2002; 295: 1657-1658Crossref PubMed Scopus (82) Google Scholar, 5Dobson C.M. Karplus M. Curr. Opin. Struct. Biol. 1999; 9: 92-101Crossref PubMed Scopus (364) Google Scholar, 6Dinner A.R. Sali A. Smith L.J. Dobson C.M. Karplus M. Trends Biochem. Sci. 2000; 25: 331-339Abstract Full Text Full Text PDF PubMed Scopus (422) Google Scholar). The narrow end of the funnel represents the folded native state, and the broad end represents the unfolded state consisting of millions of rapidly inter-converting conformers. As a protein folds from an unfolded state, it goes through one or more partially folded intermediates, which need to be characterized for elucidation of its folding pathways. Experimentally, this is a very challenging task. The most common and direct approach relies on kinetic pulse labeling experiments of amide protons coupled with hydrogen exchange at different time points along the folding reaction of the protein (reviewed in Ref. 7Englander S.W. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 213-238Crossref PubMed Scopus (413) Google Scholar). However, as has been pointed out (8Raschke T.M. Kho J. Marqusee S. Nat. Struct. Biol. 1999; 6: 825-831Crossref PubMed Scopus (106) Google Scholar), this also has limitations. First, it is limited by the necessity of detectable protection against exchange. Second, any lack of protection does not necessarily imply absence of structure. Third, it biases the interpretation of the structure in intermediates toward the native state. Even so, useful information has been obtained in many protein systems (7Englander S.W. Annu. Rev. Biophys. Biomol. Struct. 2000; 29: 213-238Crossref PubMed Scopus (413) Google Scholar). However, this approach is complicated in the case of proteases with autolytic property, because the autolytic reaction interferes with interpretation of the hydrogen exchange data. Therefore, in these systems it becomes necessary to look for alternative experimental avenues. Several equilibrium, real-time NMR, and kinetic pulse labeling studies (8Raschke T.M. Kho J. Marqusee S. Nat. Struct. Biol. 1999; 6: 825-831Crossref PubMed Scopus (106) Google Scholar, 9Dyson H.J. Wright P.E. Annu. Rev. Phys. Chem. 1996; 47: 369-395Crossref PubMed Scopus (138) Google Scholar, 10Killick T.R. Freund S.M.V. Fersht A.R. Protein Sci. 1999; 8: 1286-1291Crossref PubMed Scopus (24) Google Scholar, 11Balbach J. Steegborn C. Schindler T. Schmid F.X. J. Mol. Biol. 1999; 285: 829-842Crossref PubMed Scopus (44) Google Scholar, 12Dobson C.M. Hore P.J. Nat. Struct. Biol. 1998; 5: 504-507Crossref PubMed Scopus (81) Google Scholar, 13Barbar E. Biopolymers. 1999; 51: 191-207Crossref PubMed Scopus (38) Google Scholar, 14Shortle D. Ackerman M.S. Science. 2001; 293: 487-489Crossref PubMed Scopus (546) Google Scholar, 15Hughson F.M. Wright P.E. Baldwin R.L. Science. 1990; 249: 1544-1548Crossref PubMed Scopus (718) Google Scholar, 16Jennings P.A. Wright P.E. Science. 1993; 262: 892-896Crossref PubMed Scopus (776) Google Scholar, 17Raschke T.M. Marqusse S. Nat. Struct. Biol. 1997; 4: 298-304Crossref PubMed Scopus (236) Google Scholar, 18Dabora J.M. Pelton J.G. Marqusee S. Biochemistry. 1996; 35: 11951-11958Crossref PubMed Scopus (63) Google Scholar, 19Arai M. Kuwajima K. Adv. Protein Chem. 2000; 53: 209-282Crossref PubMed Scopus (430) Google Scholar, 20Chamberlain A.K. Marqusee S. Adv. Protein Chem. 2000; 53: 283-328Crossref PubMed Scopus (74) Google Scholar, 21Chyan C.L. Wormald C. Dobson C.M. Evans P.A. Baum J. Biochemistry. 1993; 32: 5681-5691Crossref PubMed Scopus (184) Google Scholar, 22Schulman B.A. Redfield C. Peng Z.Y. Dobson C.M. Kim P.S. J. Mol. Biol. 1995; 253: 651-657Crossref PubMed Scopus (188) Google Scholar) have indicated that in many proteins the local structural features of the kinetic intermediates have many things in common with the partially unfolded states created by the use of chemical denaturants such as urea, guanidine, or extreme pH conditions (15Hughson F.M. Wright P.E. Baldwin R.L. Science. 1990; 249: 1544-1548Crossref PubMed Scopus (718) Google Scholar, 16Jennings P.A. Wright P.E. Science. 1993; 262: 892-896Crossref PubMed Scopus (776) Google Scholar, 17Raschke T.M. Marqusse S. Nat. Struct. Biol. 1997; 4: 298-304Crossref PubMed Scopus (236) Google Scholar, 18Dabora J.M. Pelton J.G. Marqusee S. Biochemistry. 1996; 35: 11951-11958Crossref PubMed Scopus (63) Google Scholar, 19Arai M. Kuwajima K. Adv. Protein Chem. 2000; 53: 209-282Crossref PubMed Scopus (430) Google Scholar, 20Chamberlain A.K. Marqusee S. Adv. Protein Chem. 2000; 53: 283-328Crossref PubMed Scopus (74) Google Scholar, 21Chyan C.L. Wormald C. Dobson C.M. Evans P.A. Baum J. Biochemistry. 1993; 32: 5681-5691Crossref PubMed Scopus (184) Google Scholar, 22Schulman B.A. Redfield C. Peng Z.Y. Dobson C.M. Kim P.S. J. Mol. Biol. 1995; 253: 651-657Crossref PubMed Scopus (188) Google Scholar). That means characterization of the partially unfolded states created by denaturants can provide useful insights into the structural features of the kinetic intermediates. Further, the progressive folding of a protein will be associated with significant changes in its internal dynamics at all time scales (pico- to milleseconds). The fully unfolded state is highly dynamic with motions occurring mostly on picosecond time scales. Any restriction in the motions implies transient ordering of the polypeptide chain. As the protein starts to fold, more and more structure-forming breaking events (milli- to microsecond time scale) occur, and these lead to an increase in the slow motions. Thus a systematic monitoring of these graded changes in the motional characteristics as well as in the residual structures along the polypeptide chain under different conditions of denaturation provides very valuable information on the hierarchy of folding propensities in a protein. Following these ideas, we describe below NMR identification of the hierarchy of folding propensities in HIV-1 1The abbreviations used are: HIV, human immunodeficiency virus; HSQC, heteronuclear single quantum coherence; NOE, nuclear Overhauser enhancement. protease, using a tethered dimer construct of the protein in which the two monomers are joined head-to-tail by a flexible linker GGSSG; this protein, hereafter referred to as HIVTD, folds similarly to the native homodimer in vitro (23Pillai B. Kannan K.K. Hosur M.V. Proteins. 2001; 43: 57-64Crossref PubMed Scopus (60) Google Scholar, 24Kumar M. Kannan K.K. Hosur M.V. Bhavesh N.S. Chatterjee A. Mittal R. Hosur R.V. Biochem. Biophys. Res. Commun. 2002; 294: 395-401Crossref PubMed Scopus (33) Google Scholar) and also has similar activity toward the substrates. 2V. Prashar, M. Kumar, B. Pillai, A. Chatterjee, N. S. Bhavesh, R. Mittal, R. V. Hosur, and M. V. Hosur, unpublished results. NMR Experiments—Isotopically 15N-labeled HIV-1 protease-tethered dimer for the NMR experiments was prepared as described earlier (25Bhavesh N.S. Panchal S.C. Mittal R. Hosur R.V. FEBS Lett. 2001; 509: 218-224Crossref PubMed Scopus (25) Google Scholar). 1H-15N HSQC spectra were recorded with 1.0–1.2 mm protein samples in NMR buffer (50 mm acetate, 150 mm dithiothreitol, 5 mm EDTA, pH 5.2) containing different concentrations of guanidine hydrochloride at 32 °C. For the high resolution data required for coupling constant measurements, and points were along and For the relaxation and points were along the two For measurements, the were and and spectra at and were by the to the 1H-15N heteronuclear data were with a relaxation and of the The was with a relaxation The were as is the in the with and is the in the The experiments were out using the pulse described by R. C.M. T. Biochemistry. PubMed Scopus Google Scholar). the experiments were on a and the data were using on a of the structure of to hydrophobic clusters was using the C. V. T. FEBS Lett. PubMed Scopus Google Scholar). NMR of the of the changes in the state of the protease guanidine was from to m the 1H-15N HSQC was of an unfolded state. As the guanidine was to 5 the m were The protein m guanidine, a of folded partially and unfolded was as the was the protein showing protease chain of were in the HSQC were by The autolytic activity as the guanidine was Therefore, to into the early folding events in the protein we the structural and dynamic characteristics of the protein at and 5 m guanidine under which conditions the protein of a of the 1H-15N HSQC of the protein at and 5 m guanidine The of in the m is as be from the of residues in the two of the tethered the in the m are also in the 5 m at a which small an of (6 m have been (25Bhavesh N.S. Panchal S.C. Mittal R. Hosur R.V. FEBS Lett. 2001; 509: 218-224Crossref PubMed Scopus (25) Google Scholar). are which suggest the of partially folded in that the and partially folded states are highly dynamic and one can that the partially folded states be in slow exchange with the m one may these to their be more one partially folded in and the to different The of these that the state by the in the two spectra have in the dynamic characteristics under and 5 m guanidine is also from the that the resolution in the 5 m is which be to coupling or because of conformational exchange or We have these using a of NMR which provide valuable into the folding conformational transitions in the protein. in m Folding of in the that the state is not a residual structure or at local structural in an dynamic F.M. J. Adv. Proteins. Scholar). of may exhibit in the V. Adv. Protein Chem. PubMed Scopus Google Scholar). These are to be the folding or the the folding events in the polypeptide chain. characterization of the residual structures in the states of proteins is obtained from chemical from HN-Hα coupling constants, amide and in the amide against exchange H.J. Wright P.E. 2001; PubMed Scopus Google Scholar). the secondary chemical and HN-Hα coupling the secondary structural propensities, the are more and even very small in the On the the amide hydrogen and report on the of structures H.J. Wright P.E. 2001; PubMed Scopus Google Scholar). the case we the secondary chemical shifts, the HN-Hα coupling constants, and the amide in HIVTD, and a of their has been described earlier (25Bhavesh N.S. Panchal S.C. Mittal R. Hosur R.V. FEBS Lett. 2001; 509: 218-224Crossref PubMed Scopus (25) Google Scholar). As in that the S. T.R. J. Wright P.E. H.J. J. Chem. 2001; PubMed Scopus Google Scholar) chemical were the most of secondary structures because of their to the of the of for the chemical These from Ref. N.S. Panchal S.C. Mittal R. Hosur R.V. FEBS Lett. 2001; 509: 218-224Crossref PubMed Scopus (25) Google Scholar) along with the for coupling and amide are in the secondary are to be for many residues and sequence in to the the for the to the we propensities for the and and propensities for the and The secondary structure elements in the native protein are on the of the for all the propensities and the in the are and the which are are this we may of the two monomers by the flexible linker in has any influence on the structural propensities. may to these at the of the and the of the the NMR spectra we for the residues and from the two the secondary for these residues are also as indicated by in However, the for the residues are very small for the the for the residues and are for that the propensities are by the sequence and the has a small at However, as we the may have influence on the motional because of local of the residues the The HN-Hα coupling in in m guanidine from the structures of the in the HSQC spectra are in The from to with an close to is of and the small are the from sequence of these coupling Redfield C. J. Smith Dobson C.M. Smith L.J. J. Mol. Biol. 1997; PubMed Scopus Google Scholar). Thus the coupling here not any secondary structural which is in to the from secondary chemical However, such have been in many to the that the coupling are to small in the and are H.J. Wright P.E. 2001; PubMed Scopus Google Scholar). the amide in the protein in m For most residues the are in the with an at is of amide protons to the and not in any of Thus it that are structures in the protein in m of the local structural from the secondary chemical in the protein in the of its structure an many of the are to be associated with the are at and and seem to in the On the all the are associated with propensities. it that in HIVTD, the may a in the folding events the folding starts from an unfolded state. The different native and non-native folding described along with the locations of the are on the native protein structure in Interestingly, these the residues at the dimer and in the hinge of the native protein structure. that folding of active residues and active may be an early in the folding of in m heteronuclear and 15N transverse relaxation provide most valuable information on the motions of the at different time scales. the heteronuclear are most to motions on the to time the transverse relaxation are also to slow motions and conformational transitions occurring on time many these have valuable insights into motional and in which in provide to the folding Freund S.M.V. Fersht A.R. V. Sci. S. A. 2001; PubMed Scopus Google Scholar, J. M. J. E. T. T. Smith L.J. Dobson C.M. Science. 2002; 295: PubMed Scopus Google Scholar, D. J. H.J. Wright P.E. Biochemistry. 2000; PubMed Scopus Google Scholar, J. J. D. Wright P.E. H.J. Biochemistry. 2001; PubMed Scopus Google Scholar, D. J. H.J. Wright P.E. Nat. Struct. Biol. 1998; 5: PubMed Scopus Google Scholar, S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar). close was the of motions and the locations of hydrophobic clusters D. J. H.J. Wright P.E. Biochemistry. 2000; PubMed Scopus Google Scholar, J. J. D. Wright P.E. H.J. Biochemistry. 2001; PubMed Scopus Google Scholar, D. J. H.J. Wright P.E. Nat. Struct. Biol. 1998; 5: PubMed Scopus Google Scholar, S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar), which to the that transient hydrophobic by local chain may the early folding a was also the of and locations of the small amino residues and which to the that these residues as hinges in the folding of the protein S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar). The 1H-15N heteronuclear for the different residues in in m guanidine are in As against a of fully we a of and which to suggest the of as well as highly flexible along the polypeptide chain. that a of at residues of similar motional in close may be to a local domain, we of motions and of and that has two this has been because of the of the These of motions are of chain which local transient ordering and in the folding of the protein. Interestingly, the and residues exhibit motional may that because these of residues are in a in the dimerization of the native protein structure may be a for dimer even in m it is that because of the of the two monomers in HIVTD, which the of the close to the of the may be local in restriction of motions. may be the the because the polypeptide chain has an to a The 15N transverse relaxation for the different residues in in m guanidine are in we in motional The from to with an high motions on time scales and also from conformational transitions on to microsecond time As in the case of 1H-15N we of on the of the and with and from NOE, and with of high flexibility from that conformational transitions occurring on the to microsecond time which are not by the 1H-15N significant to the The in all the is the that all of have from slow motions. in m urea, pH was for to microsecond time the of from 1H-15N and 15N were S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar). the of earlier D. J. H.J. Wright P.E. Biochemistry. 2000; PubMed Scopus Google Scholar, J. J. D. Wright P.E. H.J. Biochemistry. 2001; PubMed Scopus Google Scholar, D. J. H.J. Wright P.E. Nat. Struct. Biol. 1998; 5: PubMed Scopus Google Scholar, S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar) that hydrophobic the we the hydrophobic clusters along the sequence of the protein using the C. V. T. FEBS Lett. PubMed Scopus Google Scholar). These are by at the in The and in and in with many of the hydrophobic partially in a and hydrophobic is in with earlier on proteins and that chain a in local transient ordering of the protein chain. the in and in not with any These seem to to the of protein folding in which an hydrophobic of in the case and of along the polypeptide chain are The motional restriction in and in to through of local transient structure residues in exhibit These are of very high this with the of and by and along the chain. to the by S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar) that Gly and residues as hinges in the folding of a protein. along the Folding the denaturing conditions are the protein along the folding funnel. As at 5 m guanidine, the m in the HSQC are and to partially folded appear. that the secondary structural propensities of the and 5 m guanidine, be However, be dynamics which on the conformational transitions on the folding funnel. We have these from the changes in the HN-Hα coupling and 15N transverse relaxation as described The in these parameters are also in the changes in the coupling along the sequence of the protein as we from to 5 m The changes are not and to be a in the coupling constant for all the residues a a increase in the for most residues in 5 m with we that the which has very structural in m guanidine, a residues near the active and a near the hinge in 5 m is a on the hierarchy of the folding The 15N transverse relaxation a increase in 5 m guanidine conditions for of the and the was by of this increase may from of the two into the which may be because of the head-to-tail through the linker in transient in a is on the that as the guanidine is the protein starts autolytic activity that a dimer is even in the of certain concentrations of the However, as can be in the changes in the in from to 5 m guanidine are and this on the hierarchy of conformational transitions or folding propensities in the protein. the unfolded state, which represents an of rapidly inter-converting are very exchange to the transverse relaxation As the protein starts folding for the conformational transitions and to the in the can in a As more and more the of slow to increase and along the under conditions which are near the unfolded state the protein is dynamic and the fully folded state also has conformational exchange to the the of the changes in in from to 5 m guanidine, along the sequence of the polypeptide on the to local transient or to folding conformational transitions. will structures are and be exchange to we can the residues on the of increase in to and can be to the hierarchy of folding propensities. However, the is similar to the in and may not be it that residues have the highest propensities for conformational transitions, by residues at the of these residues with the hydrophobic clusters in that conformational transitions may be by changes in the chain hydrophobic residues the concentrations are these folding in a on the structure of the native protein. is that the residues in these conformational transitions are contiguous or are in close proximity on the structure, suggesting a of cooperativity in the folding These conformational transitions involve of native as well as of the non-native Interestingly, the dimerization domain, the flaps, and their hinges the highest to conformational transitions in the protein. of that the for the dimerization may be a of the of the two which the of the near the of the to in the of a transient as in the native protein structure. the conformational transitions may lead to the of partially folded as in the data any on the folding of the protease is a or a this is of for the of folding of the protein. this is by monitoring the folding using a of such as the that it is a the folding from all the be A.R. and in Protein to and Scholar). the of a folding by any of the conventional in the kinetic or in the is severely hampered because of the of the protein. we that in 5 m guanidine, the HSQC spectra many with the m and most of these not to the native state of the protein. means that is at one non-native state that with the m state. is more that are non-native states that are different from the m state, and these from different protein at the surface of the folding different along the folding funnel. is with the different of folding propensities along the sequence of the polypeptide chain. Thus we that the protein has a folding and even the early events involve many states near the surface of the folding funnel. may different and may be local cooperativity of folding transitions along We have here to useful insights into the early events in the folding of HIV-1 protease, using a tethered dimer construct as its a of NMR parameters under conditions of guanidine the early folding have been These include native as well as non-native secondary and the native structures the residues at the active site, the and the dimerization domain. Interestingly, proline residues seem to have significant influence on the local structural in the state of the protein. the seem to on the motional along the and these may be in the folding is also that and conformational and may as hinges in the folding as earlier S. Wright P.E. H.J. Biochemistry. 2002; PubMed Scopus Google Scholar). As the is partially folded and the protein to autolytic The at and 5 m guanidine denaturing conditions that the dimerization domain, the flaps, and the hinges or the of the protein have the highest to conformational transitions. The graded propensities also suggest a certain degree of local cooperativity in these conformational transitions. from the of more one non-native even in 5 m guanidine and the hierarchy of the conformational transitions, it that the protein folds a with different different pathways. this is the experimental of the early events in the folding of HIV-1