Journal of Molecular Biology
Volume 315, Issue 2, 11 January 2002, Pages 193-211
Journal home page for Journal of Molecular Biology

Regular article
Highly divergent dihydrofolate reductases conserve complex folding mechanisms1

https://doi.org/10.1006/jmbi.2001.5230Get rights and content

Abstract

To test the hypothesis that protein folding mechanisms are better conserved than amino acid sequences, the mechanisms for dihydrofolate reductases (DHFR) from human (hs), Escherichia coli (ec) and Lactobacillus casei (lc) were elucidated and compared using intrinsic Trp fluorescence and fluorescence-detected 8-anilino-1-naphthalenesulfonate (ANS) binding. The development of the native state was monitored using either methotrexate (absorbance at 380 nm) or NADPH (extrinsic fluorescence) binding. All three homologs displayed complex unfolding and refolding kinetic mechanisms that involved partially folded states and multiple energy barriers. Although the pairwise sequence identities are less than 30 %, folding to the native state occurs via parallel folding channels and involves two types of on-pathway kinetic intermediates for all three homologs. The first ensemble of kinetic intermediates, detected within a few milliseconds, has significant secondary structure and exposed hydrophobic cores. The second ensemble is obligatory and has native-like side-chain packing in a hydrophobic core; however, these intermediates are unable to bind active-site ligands. The formation of the ensemble of native states occurs via three channels for hsDHFR, and four channels for lcDHFR and ecDHFR. The binding of active-site ligands (methotrexate and NADPH) accompanies the rate-limiting formation of the native ensemble. The conservation of the fast, intermediate and slow-folding events for this complex α/β motif provides convincing evidence for the hypothesis that evolutionarily related proteins achieve the same fold via similar pathways.

Introduction

The tertiary structures of homologous proteins, i.e. those with similar biological functions that have evolved from a common ancestral gene, are known to be better conserved than their amino acid sequences.1 Reasoning that the conservation of the product of a folding reaction might also require the conservation of key folding intermediates and/or transition states, it has been proposed that folding mechanisms of proteins have been conserved during evolution.2, 3

The validity of this hypothesis has been supported by a number of experimental studies on small (< 100 residues), single-domain proteins that have relatively high levels of sequence similarity (for a review, see Gunasekaran et al.4 and references therein). For the proteins that fold via a simple two-state mechanism, e.g. the family of all-β SH3 domains,5, 6, 7, 8 the properties of the transition state and its position on the reaction coordinate diagram were highly conserved. Even simplifying the amino acid code of the SH3 domains to five letters does not cause a dramatic change in the rate of folding or in the position of the transition state.9 Further, molecular dynamic unfolding and ab initio refolding simulations10 for the SH3 domains were consistent with the experimental findings.

Although a few studies comparing the complex folding mechanisms of larger and/or multi-domain proteins have been published (see Stackhouse et al.11 and Gunasekaran et al.4) it is not yet clear whether rate-limiting steps that occur after the development of chain topology are also conserved. The family of dihydrofolate reductases (DHFRs, EC. 1.5.1.3) is an ideal protein system with which the conservation of folding pathways hypothesis can be tested. DHFRs, which function in vivo by maintaining the pool of dihydrofolate and its derivatives that are essential for the biosynthesis of purines, thymidylates and several amino acids (for reviews, see Blakely12, 13), span the three domains (archae, prokaryote and eukaryote) of life. All representatives whose structures have been solved thus far have an α/β/α topology where the mixed eight-stranded β sheet is flanked on either side by two α helices.14, 15, 16, 17, 18, 19 The complex, doubly wound β-sheet topology is composed of two structurally distinct domains; an adenosine-binding domain (ABD) common to all nucleotide-binding proteins, and a loop subdomain (Figure 1; and see Sawaya et al.17).

The thermodynamic properties and kinetic folding mechanism of Escherichia coli (ec) DHFR have been studied in detail, using a wide variety of biophysical techniques.20, 21, 22, 23, 24, 25, 26, 27, 28 The proposed kinetic folding mechanism for ecDHFR is complex. The first detectable step in folding involves a rapid collapse (<5 ms) of the urea-denatured state to form a burst phase species (IBP) that has significant secondary structure22 and an exposed hydrophobic core.24 The results of pulse-labeling hydrogen exchange experiments in combination with NMR25 have been interpreted to imply that this core involves two clusters of non-polar side-chains that stabilize the native β-sheet topology in 60 % of the population. Mutational analysis28 has demonstrated stereo-specific packing in these non-polar side-chains. The rapid establishment of the β-sheet topology is followed by the formation of a set of hyperfluorescent intermediate states (I1-I4; τ = 200 ms) that convert to the native states (N1-N4) via four parallel folding channels (relaxation times in the range of 1-100 seconds). All of the native states bind the competitive inhibitor methotrexate (MTX), but only one native state (N2) has the ability to bind the cofactor, NADP+. The richness of this kinetic mechanism, including two distinct folding reactions that follow the chain organization step, provides the basis for a stringent test of the conservation of folding mechanisms hypothesis.

The thermodynamic properties and folding kinetics of human (hs) DHFR and Lactobacillus casei (lc) DHFR were elucidated and compared with those for ecDHFR. Although the pairwise sequence identities are less than 30 %, the comparison demonstrated that the complex folding mechanism is conserved throughout the family of DHFRs.

Section snippets

Sequence and structural analysis

The structures of all DHFRs have striking similarity: structural alignments of hsDHFR, ecDHFR and lcDHFR (Figure 2(a)) indicated backbone RMSD values between 1.5 and 2.0 Å. The vertebrate enzymes are approximately 30 residues longer and result in the presence of external loops in the eukaryotic forms.14, 17 Despite the well-conserved backbones pairwise sequence identities are low: hs/ec, 29.2 %; ec/lc, 27.0 %; and lc/hs, 27.8 % (Figure 2(b)). Only ten residues are identical throughout the 78

Similarities and differences in folding mechanisms

The folding mechanisms for ec-, lc- and hsDHFR are fundamentally very similar, despite the low conservation of the amino acid sequence (∼30 % pairwise sequence identity). All three systems fold to the native state via parallel channels with two types of on-pathway kinetic intermediates. The deduced folding mechanisms for all three DHFRs are depicted in Scheme 1.Table 1 highlights the similarity of the unfolding and refolding relaxation times for the ecDHFR and hsDHFR homologs; the refolding

Materials

Ultrapure urea was purchased from ICN Biomedicals Inc. (Aurora, OH) and purified further using a mixed-bead, anion-exchange resin (AG™ 501-X8; Biorad). Folic acid, dihydrofolate, β-NADP+, methotrexate and recombinant human cyclophilin were purchased from Sigma. ANS and β-NADPH were obtained from Molecular Probes (Eugene, OR) and Amersham Life Sciences (Piscataway, NJ), respectively. All other reagents were of analytical grade. The concentrations of ligands were determined spectrophotometrically

Acknowledgements

We gratefully acknowledge the initial research on hsDHFR performed by Dr Jing Zhang. We thank Dr Jill A. Zitzewitz and B. Robert Simler for a critical reading of the manuscript, and Drs Roxana Ionescu and John O’Neill for fruitful discussions. This work was supported by National Science Foundation grants MCB-9604678 and MCB-0081076 to C.R.M. L.A.W. was supported, in part, by an NRF Postdoctoral Fellowship (South Africa).

References (92)

  • F.X. Schmid et al.

    The rate of interconversion between the two unfolded forms of ribonuclease A does not depend on guanidinium chloride concentration

    J. Mol. Biol.

    (1979)
  • O. Kops et al.

    Ssp1, a site-specific parvulin homolog from Neurospora crassa active in protein folding

    J. Biol. Chem.

    (1998)
  • O. von Ahsen et al.

    Cyclophilin-promoted folding of mouse dihydrofolate reductase does not include the slow conversion of the late-folding intermediate to the active enzyme

    J. Mol. Biol.

    (2000)
  • C. Odefey et al.

    Non-prolyl cis-trans peptide bond isomerization as a rate-determining step in protein unfolding and refolding

    J. Mol. Biol.

    (1995)
  • K.F. Fischer et al.

    A rapid test for identification of autonomous folding units in proteins

    J. Mol. Biol.

    (2000)
  • Z. Dosztanyi et al.

    Stabilization centers in proteinsidentification, characterization and predictions

    J. Mol. Biol.

    (1997)
  • L.A. Mirny et al.

    Universally conserved positions in protein foldsreading evolutionary signals about stability, folding kinetics and function

    J. Mol. Biol.

    (1999)
  • H. Roder et al.

    Kinetic role of early intermediates in protein folding

    Curr. Opin. Struct. Biol.

    (1997)
  • A.L. Fink et al.

    Discrete intermediates versus molten globule models for protein foldingcharacterization of partially folded intermediates of apomyoglobin

    Fold. Des.

    (1998)
  • A. Matagne et al.

    The origin of the alpha-domain intermediate in the folding of hen lysozyme

    J. Mol. Biol.

    (1998)
  • P. Wolynes et al.

    Fast-folding experiments and the topography of protein folding energy landscapes

    Chem. Biol.

    (1996)
  • N. Ferguson et al.

    Rapid folding with and without populated intermediates in the homologous four-helix proteins Im7 and Im9

    J. Mol. Biol.

    (1999)
  • J. Andrews et al.

    Nucleotide sequence of the dihydrofolate reductase gene of methotrexate-resistant Lactobacillus casei

    Gene

    (1985)
  • F.C. Bernstein et al.

    The Protein Data Banka computer-based archival file for macromolecular structures

    J. Mol. Biol.

    (1977)
  • B.L. Hillcoat et al.

    Dihydrofolate reductase of Streptococcus faecalis. I. Purification and some properties of reductase from the wild strain and from strain A

    J. Biol. Chem.

    (1966)
  • B. Birdsall et al.

    Correction for light absorption in fluorescence studies of protein-ligand interactions

    Anal. Biochem.

    (1983)
  • M.G. Rossmann et al.

    Protein folding

    Annu. Rev. Biochem.

    (1981)
  • K.W. Plaxco et al.

    Topology, stability, sequence, and lengthdefining the determinants of two-state protein folding kinetics

    Biochemistry

    (2000)
  • A.R. Viguera et al.

    Thermodynamic and kinetic analysis of the SH3 domain of spectrin shows a two-state folding transition

    Biochemistry

    (1994)
  • V.P. Grantcharova et al.

    Folding dynamics of the src SH3 domain

    Biochemistry

    (1997)
  • K.W. Plaxco et al.

    The folding kinetics and thermodynamics of the Fyn-SH3 domain

    Biochemistry

    (1998)
  • J.C. Martinez et al.

    The folding transition state between SH3 domains is conformationally restricted and evolutionarily conserved

    Nature Struct. Biol.

    (1999)
  • D.S. Riddle et al.

    Functional rapidly folding proteins from simplified amino acid sequences

    Nature Struct. Biol.

    (1997)
  • D.S. Riddle et al.

    Experiment and theory highlight role of native state topology in SH3 folding

    Nature Struct. Biol.

    (1999)
  • T.M. Stackhouse et al.

    Folding of homologous proteinsconservation of the folding mechanism of the alpha subunit of tryptophan synthase from Escherichia coli, Salmonella typhimurium, and five interspecies hybrids

    Biochemistry

    (1988)
  • R.L. Blakely

    Dihydrofolate reductases

  • R.L. Blakely
  • J.F. Davies et al.

    Crystal structures of recombinant human dihydrofolate reductase complexed with folate and 5-deazafolate

    Biochemistry

    (1990)
  • M.R. Sawaya et al.

    Loop and subdomain movements in the mechanism of Escherichia coli dihydrofolate reductasecrystallographic evidence

    Biochemistry

    (1997)
  • N.A. Touchette et al.

    Folding of dihydrofolate reductase from Escherichia coli

    Biochemistry

    (1986)
  • C. Frieden

    Refolding of Escherichia coli dihydrofolate reductasesequential formation of substrate binding sites

    Proc. Natl Acad. Sci. USA

    (1990)
  • K. Kuwajima et al.

    Transient intermediates in the folding of dihydrofolate reductase as detected by far-ultraviolet circular dichroism spectroscopy

    Biochemistry

    (1991)
  • P.A. Jennings et al.

    A reexamination of the folding mechanism of dihydrofolate reductase from Escherichia coliverification and refinement of a four-channel model

    Biochemistry

    (1993)
  • B.E. Jones et al.

    Development of nonpolar surfaces in the folding of Escherichia coli dihydrofolate reductase detected by 1-anilinonaphthalene-8-sulfonate binding

    Biochemistry

    (1994)
  • B.E. Jones et al.

    Early intermediate in the folding of dihydrofolate reductase from Escherichia coli detected by hydrogen exchange and NMR

    Protein Sci.

    (1995)
  • C. Clarke et al.

    Native Escherichia coli and murine dihydrofolate reductases contain late-folding non-native structures

    J. Mol. Biol.

    (1999)
  • Cited by (40)

    • Investigation of folding unfolding process of a new variant of dihydrofolate reductase protein from Zebrafish

      2016, International Journal of Biological Macromolecules
      Citation Excerpt :

      In the current study an attempt has been made to facilitate an in depth knowledge into the structure-function relationship of zDHFR through investigation of its GdnHCl mediated unfolding process and observing the process of refolding. The mechanisms for dihydrofolate reductases (DHFR) from human (hs), Escherichia coli (ec) and Lactobacillus casei (lc) were elucidated and compared using intrinsic Trp Fluorescence and Fluorescence-detected 8-anilino-1-naphthalenesulfonate (ANS) binding [33]. All three homologs displayed complex unfolding and refolding kinetic mechanisms that involved partially folded states and multiple energy barriers.

    • Intermediates in protein folding

      2012, Comprehensive Biophysics
    • Structural-based analysis of dihydrofolate reductase evolution

      2011, Molecular Phylogenetics and Evolution
    • NMR analysis of a kinetically trapped intermediate of a disulfide-deficient mutant of the starch-binding domain of glucoamylase

      2011, Journal of Molecular Biology
      Citation Excerpt :

      The average value of the half-life is 11 h. The intermediate was found to be unexpectedly stable when compared with typical folding intermediates with much shorter half-lives on the microsecond-to-millisecond timescale.3,5–13 These results indicate that the transition from the intermediate to the native state is highly cooperative and requires gross structural rearrangement.

    View all citing articles on Scopus
    1

    Edited by P. E. Wright

    View full text