Substitution of adenine for guanine in the quadruplex-forming human telomere DNA sequence G(3)(T(2)AG(3))(3)

We have studied the formation and structural properties of quadruplexes of the human telomeric DNA sequence G(3)(T(2)AG(3))(3) and related sequences in which each guanine base was replaced by an adenine base. None of these single base substitutions hindered the formation of antiparallel quadruplexes, as shown by circular dichroism, gel electrophoresis, and UV thermal stability measurements in NaCl solutions. Effect of substitution did differ, however, depending on the position of the substituted base. The A-for-G substitution in the middle quartet of the antiparallel basket scaffold led to the most distorted and least stable structures and these sequences preferred to form bimolecular quadruplexes. Unlike G(3)(T(2)AG(3))(3), no structural transitions were observed for the A-containing analogs of G(3)(T(2)AG(3))(3) when sodium ions were replaced by potassium ions. The basic quadruplex topology remained the same for all sequences studied in both salts. As in vivo misincorporation of A for a G in the telomeric sequence is possible and potassium is a physiological salt, these findings may have biological relevance.     

Stability of human telomere quadruplexes at high DNA concentrations

For mimicking macromolecular crowding of DNA quadruplexes, various crowding agents have been used, typically PEG, with quadruplexes of micromolar strand concentrations. Thermal and thermodynamic stabilities of these quadruplexes increased with the concentration of the agents, the rise depended on the crowder used. A different phenomenon was observed, and is presented in this article, when the crowder was the quadruplex itself. With DNA strand concentrations ranging from 3 µM to 9 mM, the thermostability did not change up to ～2 mM, above which it increased, indicating that the unfolding quadruplex units were not monomolecular above ～2 mM. The results are explained by self‐association of the G‐quadruplexes above this concentration. The ΔG^o₃₇ values, evaluated only below 2 mM, did not become more negative, as with the non‐DNA crowders, instead, slightly increased. Folding topology changed from antiparallel to hybrid above 2 mM, and then to parallel quadruplexes at high, 6–9 mM strand concentrations. In this range, the concentration of the DNA phosphate anions approached the concentration of the K⁺ counterions used. Volume exclusion is assumed to promote the topological changes of quadruplexes toward the parallel, and the decreased screening of anions could affect their stability. © 2013 Wiley Periodicals, Inc. Biopolymers 101: 428–438, 2014.     

A dimeric RNA quadruplex architecture comprised of two G:G(:A):G:G(:A) hexads,G:G:G:G tetrads and UUUU loops

Using CD and NMR, we determined the structure of an RNA oligomer, r(GGAGGUUUUGGAGG) (R14), comprising two GGAGG segments joined by a UUUU segment. A modified quadruplex structure was observed for r(GGAGGUUUUGGAGG) in solution even in the absence of K(+). An unusually stable dimeric RNA quadruplex architecture formed from two strands of r(GGAGGUUUUGGAGG) at low K(+) concentration is reported here. In each strand of r(GGAGGUUUUGGAGG), two sets of successive turns in the GGAGG segments and turns at both ends of the UUUU loops drive four G-G steps to align in a parallel manner, a core with two stacked G-tetrads being formed. Two adenine bases bind to two edges of one G:G:G:G tetrad through the sheared G:A mismatch augmenting the tetrad into a G:G(:A):G:G(:A) hexad. Thus, one molecule of r(GGAGGUUUUGGAGG) folds into a modified quadruplex comprising a G:G:G:G tetrad, a UUUU double-chain reversal loop and a G:G(:A):G:G(:A) hexad. Two such molecules further associate by stacking through the dimeric hexad-hexad interface with a rotational symmetry. The ribose rings of most nucleotides take S (close to C2'-endo) puckering, which is unusual for an RNA. K(+) can increase the stability of this quadruplex structure; the number of bound K(+) was estimated from the results of the titration experiment. Besides G:G and G:A mismatches, a network of hydrogen bonds including O4'-NH(2) and C-H..O hydrogen bonds, and the extensive base stacking contribute to the high thermodynamic stability of R14. Our results could provide the stereochemical and thermodynamic basis for elucidating the biological role of the GGAGG-containing RNA segments abundantly existing in various RNAs. Relevance to quadruplex-mediated mRNA-FMRP binding and HIV-1 genome RNA dimerization is discussed.     

Recognition of human telomeric G‐quadruplex DNA by berberine analogs: effect of substitution at the 9 and 13 positions of the isoquinoline moiety

G‐quadruplex forming sequences are widely distributed in human genome and serve as novel targets for regulating gene expression and chromosomal maintenance. They offer unique targets for anticancer drug development. Here, the interaction of berberine (BC) and two of its analogs bearing substitution at 9 and 13‐position with human telomeric G‐quadruplex DNA sequence has been investigated by biophysical techniques. Both the analogs exhibited several‐fold higher binding affinity than berberine. The Scatchard binding isotherms revealed non‐cooperative binding. 9‐ω‐amino hexyl ether analog (BC1) showed highest affinity (1.8 × 10⁶ M^?1) while the affinity of the 13‐phenylpropyl analog (BC2) was 1.09 × 10⁶ M^?1. Comparative fluorescence quenching and polarization anisotropy of the emission spectra gave evidence for a stronger stacking interaction of the analogs compared to berberine. The thiazole orange displacement assay has clearly established that the analogs were more effective in displacing the end stacked dye in comparison to berberine. However, the binding of the analogs did not induce any major structural perturbation in the G‐quadruplex structure, but led to higher thermal stability. Energetics of the binding indicated that the association of the analogs was exothermic and predominantly entropy driven phenomenon. Increasing the temperature resulted in weaker binding; the enthalpic contribution increased and the entropic contribution decreased. A small negative heat capacity change with significant enthalpy–entropy compensation established the involvement of multiple weak noncovalent interactions in the binding process. The 9‐ω‐amino hexyl ether analog stabilized the G‐quadruplex structure better than the 13‐phenyl alkyl analog. Copyright © 2015 John Wiley & Sons, Ltd.     

Formation of G‐wires,bimolecular and tetramolecular quadruplex: Cation‐induced structural polymorphs of G‐rich DNA sequence of human SYTX gene

An exceptional property of auto‐folding into a range of intra‐ as well as intermolecular quadruplexes by guanine‐rich oligomers (GROs) of promoters, telomeres and various other genomic locations is still one of the most attractive areas of research at present times. The main reason for this attention is due to their established in vivo existence and biological relevance. Herein, the structural status of a 20‐nt long G‐rich sequence with two G5 stretches (SG20) is investigated using various biophysical and biochemical techniques. Bioinformatics analysis suggested the presence of a 17‐nt stretch of this SG20 sequence in the intronic region of human SYTX (Synaptotagmin 10) gene. The SYTX gene helps in sensing out the Ca²⁺ ion, causing its intake in the pre‐synaptic neuron. A range of various topologies like bimolecular, tetramolecular and guanine‐wires (nano‐wires) was exhibited by the studied sequence, as a function of cations (Na⁺/K⁺) concentration. UV‐thermal denaturation, gel electrophoresis, and circular dichroism (CD) spectroscopy showed correlations and established a cation‐dependent structural switch. The G‐wire formation, in the presence of K⁺, may further be explored for its possible relevance in nano‐biotechnological applications.     

The inactive X chromosome in the human female is enriched in 5-methylcytosine to an unusual degree and appears to contain more of this modified nucleotide than the remainder of the genome

By employing a procedure that combines ELISA and photoacoustic spectroscopy, we have examined the content of 5-methylcytosine (m⁵C) in DNA of individuals who differed from one another in the number of X chromosomes in their genomes. The results show that the human inactive X chromosome (Xi) contains very high amounts of this modified nucleotide. We estimate that in the 46, XX female there is more m⁵C in Xi (∼ 3.6 x 10⁷) than in all the remaining chromosomes put together (∼2.1 x 10⁷). Our results also suggest that nearly one-fifth of all cytosines in Xi are methylated and that, in addition to CpG methylation, there is extensive non-CpG methylation as well.     

Effects of N2,N2-dimethylguanosine on RNA structure and stability: crystal structure of an RNA duplex with tandem m2 2G:A pairs

Methylation of the exocyclic amino group of guanine is a relatively common modification in rRNA and tRNA. Single methylation (N²-methylguanosine, m²G) is the second most frequently encountered nucleoside analog in Escherichia coli rRNAs. The most prominent case of dual methylation (N²,N²-dimethylguanosine, m²₂G) is found in the majority of eukaryotic tRNAs at base pair m²₂G26:A44. The latter modification eliminates the ability of the N² function to donate in hydrogen bonds and alters its pairing behavior, notably vis-à-vis C. Perhaps a less obvious consequence of the N²,N²-dimethyl modification is its role in controlling the pairing modes between G and A. We have determined the crystal structure of a 13-mer RNA duplex with central tandem m²₂G:A pairs. In the structure both pairs adopt an imino-hydrogen bonded, pseudo-Watson–Crick conformation. Thus, the sheared conformation frequently seen in tandem G:A pairs is avoided due to a potential steric clash between an N²-methyl group and the major groove edge of A. Additionally, for a series of G:A containing self-complementary RNAs we investigated how methylation affects competitive hairpin versus duplex formation based on UV melting profile analysis.     

Exploring the binding dynamics of etoricoxib with human hemoglobin: A spectroscopic,calorimetric, and molecular modeling approach

Etoricoxib, widely used for the treatment of osteoarthritis, rheumatoid arthritis, ankylosing spondylitis, and related conditions has ample affinity to bind with globular proteins. Here, the molecular interaction between purified human hemoglobin (HHb), a major heme protein and etoricoxib, a cyclooxygenase-2 inhibitor was studied by various spectroscopic, calorimetric, and molecular modeling techniques. The binding affected hypochromic changes in the Soret band of hemoglobin (Hb) and induced remarkable quenching of the intrinsic fluorescence property of protein molecules. Synchronous fluorescence studies revealed alterations in tryptophan (Trp) and tyrosine (Tyr) microenvironments of HHb molecule in presence of etoricoxib. Flouremetric and isothermal titration calorimetric studies suggested two binding sites in HHb for etoricoxib at three different temperatures (298.15, 303.15, and 310.15 K). Negative values of Gibbs energy change (ΔG⁰) and enthalpy change (ΔH⁰) strongly suggest that it is spontaneous and exothermic reaction, mainly stabilized by hydrogen bonding as evidenced by sucrose binding assay. These findings support our in silico molecular docking study, which specified the binding site and the non-covalent interactions involved in the association. Moreover, the interaction impacts on structural integrity and functional aspects of HHb as confirmed by decreased α helicity, increased free iron release, increased rate of co-oxidation, and decreased rate of esterase activity. Overall, these studies conclude that etoricoxib leads to a remarkable alteration in the conformational aspects of binding to HHb.

Communicated by Ramaswamy H. Sarma.     

A preliminary analysis of the DNA and diet of the extinct Beothuk: a systematic approach to ancient human DNA

We have used a systematic protocol for extracting, quantitating, sexing and validating ancient human mitochondrial and nuclear DNA of one male and one female Beothuk, a Native American population from Newfoundland, which became extinct approximately 180 years ago. They carried mtDNA haplotypes, which fall within haplogroups X and C, consistent with Northeastern Native populations today. In addition we have sexed the male using a novel-sexing assay and confirmed the authenticity of his Y chromosome with the presence of the Native American specific Y-QM3 single nucleotide polymorphism (SNP). This is the first ancient nuclear SNP typed from a Native population in the Americas. In addition, using the same teeth we conducted a stable isotopes analysis of collagen and dentine to show that both individuals relied on marine sources (fresh and salt water fish, seals) with no hierarchy seen between them, and that their water sources were pooled or stored water. Both mtDNA sequence data and Y SNP data hint at possible gene flow or a common ancestral population for both the Beothuk and the current day Mikmaq, but more importantly the data do not lend credence to the proposed idea that the Beothuk (specifically, Nonosabasut) were of admixed (European-Native American) descent. We also analyzed patterns of DNA damage in the clones of authentic mtDNA sequences; there is no tendency for DNA damage to occur preferentially at previously defined mutational hotspots, suggesting that such mutational hotspots are not hypervariable because they are more prone to damage.     

Binding of ibuprofen to human hemoglobin: elucidation of their molecular recognition by spectroscopy,calorimetry, and molecular modeling techniques

Ibuprofen, used for the treatment of acute and chronic pain, osteoarthritis, rheumatoid arthritis, and related conditions has ample affinity to globular proteins. Here we have explored this fundamental study pertaining to the interaction of ibuprofen with human hemoglobin (HHb), using multispectroscopic, calorimetric, and molecular modeling techniques to gain insights into molecular aspects of binding mechanism. Ibuprofen-induced graded decrease in absorption spectra indicates protein disruption along with sedimentation of HHb particle. Red shifting of absorption peak at 195 nm indicates alteration in the secondary structure of HHb upon interaction with ibuprofen. Flouremetric and isothermal titration calorimetric (ITC) studies suggested one binding site in HHb for ibuprofen at 298.15 K. However, with increase in temperature, ITC revealed increasing number of binding sites. The negative values of Gibbs energy change (ΔG⁰) and enthalpy change (ΔH⁰) along with positive value of entropy change (ΔS⁰) strongly suggest that it is entropy-driven spontaneous exothermic reaction. Moreover, hydrophobic interaction, hydrogen bonding, and π–π interaction play major role in this binding process as evidenced from ANS (8-anilino-1-napthalenesulphonic acid), sucrose binding, and molecular modeling studies. The interaction impacts on structural integrity and functional aspects of HHb as confirmed by CD spectroscopy, increased free iron release, increased rate of co-oxidation and decreased rate of esterase activity. These findings suggest us to conclude that ibuprofen upon interaction perturbs both structural and functional aspects of HHb.     

A guide to the effects of a large portion of the residues of triosephosphate isomerase on catalysis,stability, druggability,and human disease

Triosephosphate isomerase (TIM) is a ubiquitous enzyme, which appeared early in evolution. TIM is responsible for obtaining net ATP from glycolysis and producing an extra pyruvate molecule for each glucose molecule, under aerobic and anaerobic conditions. It is placed in a metabolic crossroad that allows a quick balance of the triose phosphate aldolase produced by glycolysis, and is also linked to lipid metabolism through the alternation of glycerol‐3‐phosphate and the pentose cycle. TIM is one of the most studied enzymes with more than 199 structures deposited in the PDB. The interest for this enzyme stems from the fact that it is involved in glycolysis, but also in aging, human diseases and metabolism. TIM has been a target in the search for chemical compounds against infectious diseases and is a model to study catalytic features. Until February 2017, 62% of all residues of the protein have been studied by mutagenesis and/or using other approaches. Here, we present a detailed and comprehensive recompilation of the reported effects on TIM catalysis, stability, druggability and human disease produced by each of the amino acids studied, contributing to a better understanding of the properties of this fundamental protein. The information reviewed here shows that the role of the noncatalytic residues depend on their molecular context, the delicate balance between the short and long‐range interactions in concerted action determining the properties of the protein. Each protein should be regarded as a unique entity that has evolved to be functional in the organism to which it belongs. Proteins 2017; 85:1190–1211. © 2017 Wiley Periodicals, Inc.     

Determining the binding site and binding affinity of estradiol to human serum albumin and holo-transferrin: fluorescence spectroscopic,isothermal titration calorimetry and molecular modeling approaches

The interactions between estradiol and two carrier proteins, i.e. human serum albumin (HSA) and holo-transferrin (HTF) in aqueous solution at pH = 7.4 were studied by three-dimensional fluorescence emission spectroscopy, isothermal titration calorimetry (ITC), zeta-potential, resonance light-scattering and molecular modeling. Extensive fluorescence quenching was observed throughout the interaction between the drug and both proteins. Moreover, conformational changes were determined by observing the rearrangement of Trp residues during binding of estradiol with HSA and HTF at different concentrations. ITC experiments revealed that, in the presence of estradiol, both van der Waals forces and hydrogen bonding became predominant. In addition, other binding parameters such as enthalpy and entropy changes were determined by the zeta potential method. Molecular modeling suggested that estradiol was situated within sub-domain IB sited in the hydrophobic cluster in Site I, whereas the drug was located in the N-terminal of HTF where it was hydrogen bonded with Ala 670.     

Molecular binding of toxic phenothiazinium derivatives,azures to bovine serum albumin: A comparative spectroscopic,calorimetric, and in silico study

In this paper, the comparative binding behavior of antimalarial drug azure A, azure B and azure C with bovine serum albumin (BSA) has been studied. The interaction has been confirmed by multispectroscopic (UV, fluorescence, Fourier transform infrared (FT‐IR), and circular dichroism) and molecular docking techniques. The experimental results show that azure B has the highest BSA binding affinity followed by azure A and azure C. The experimental evidence of binding showed a static quenching mechanism in the interaction azures with BSA. The isothermal titration calorimetry result reveals that the binding was exothermic with positive entropy contribution in each case. The thermodynamic parameters ΔH, ΔG, and ΔS at 25°C were calculated, which indicates that the weak van der Waals forces and hydrogen bonding rather than the hydrophobic effect played an important role in the interaction. According to the theory of Förster nonradiative energy transfer, the distance (r) between the donor (BSA) and acceptor azures found to be <7 nm in all the case. The circular dichroism and FT‐IR studies show that the content of α‐helix structure has increased for the azures‐BSA system. Overall, experimental studies characterize the interaction dynamics and energetics of the binding of three toxic analogs towards the physiologically relevant serum albumins. We hope, the outcome of this work will be most helpful for synthesizing a new type of phenothiazinium derivatives of the better therapeutic application.     

Quantitative analysis of mRNA expression levels and DNA methylation profiles of three neighboring genes: FUS1, NPRL2/G21 and RASSF1A in non-small cell lung cancer patients

Background

Tumor suppressor gene (TSG) inactivation plays a crucial role in carcinogenesis. FUS1, NPRL2/G21 and RASSF1A are TSGs from LUCA region at 3p21.3, a critical chromosomal region in lung cancer development. The aim of the study was to analyze and compare the expression levels of these 3 TSGs in NSCLC, as well as in macroscopically unchanged lung tissue surrounding the primary lesion, and to look for the possible epigenetic mechanism of TSG inactivation via gene promoter methylation.

Methods

Expression levels of 3 TSGs and 2 DNA methyltransferases, DNMT1 and DNMT3B, were assessed using real-time PCR method (qPCR) in 59 primary non-small cell lung tumors and the matched macroscopically unchanged lung tissue samples. Promoter methylation status of TSGs was analyzed using methylation-specific PCRs (MSP method) and Methylation Index (MI) value was calculated for each gene.

Results

The expression of all three TSGs were significantly different between NSCLC subtypes: RASSF1A and FUS1 expression levels were significantly lower in squamous cell carcinoma (SCC), and NPRL2/G21 in adenocarcinoma (AC). RASSF1A showed significantly lower expression in tumors vs macroscopically unchanged lung tissues. Methylation frequency was 38–76 %, depending on the gene. The highest MI value was found for RASSF1A (52 %) and the lowest for NPRL2/G21 (5 %). The simultaneous decreased expression and methylation of at least one RASSF1A allele was observed in 71 % tumor samples. Inverse correlation between gene expression and promoter methylation was found for FUS1 (rs = −0.41) in SCC subtype. Expression levels of DNMTs were significantly increased in 75–92 % NSCLCs and were significantly higher in tumors than in normal lung tissue. However, no correlation between mRNA expression levels of DNMTs and DNA methylation status of the studied TSGs was found.

Conclusions

The results indicate the potential role of the studied TSGs in the differentiation of NSCLC histopathological subtypes. The significant differences in RASSF1A expression levels between NSCLC and macroscopically unchanged lung tissue highlight its possible diagnostic role in lung cancer in situ recognition. High percentage of lung tumor samples with simultaneous RASSF1A decreased expression and gene promoter methylation indicates its epigenetic silencing. However, DNMT overexpression doesn’t seem to be a critical determinate of its promoter hypermethylation.     

Effects of phospholipase A(2) and its products on structural stability of human LDL: relevance to formation of LDL-derived lipid droplets

Hydrolysis and oxidation of LDL stimulate LDL entrapment in the arterial wall and promote inflammation and atherosclerosis via various mechanisms including lipoprotein fusion and lipid droplet formation. To determine the effects of FFA on these transitions, we hydrolyzed LDL by phospholipase A(2) (PLA(2)), removed FFA by albumin, and analyzed structural stability of the modified lipoproteins. Earlier, we showed that heating induces LDL remodeling, rupture, and coalescence into lipid droplets resembling those found in atherosclerotic lesions. Here, we report how FFA affect these transitions. Circular dichroism showed that mild LDL lipolysis induces partial β-sheet unfolding in apolipoprotein B. Electron microscopy, turbidity, and differential scanning calorimetry showed that mild lipolysis promotes LDL coalescence into lipid droplets. FFA removal by albumin restores LDL stability but not the protein conformation. Consequently, FFA enhance LDL coalescence into lipid droplets. Similar effects of FFA were observed in minimally oxidized LDL, in LDL enriched with exogenous FFA, and in HDL and VLDL. Our results imply that FFA promote lipoprotein coalescence into lipid droplets and explain why LDL oxidation enhances such coalescence in vivo but hampers it in vitro. Such lipid droplet formation potentially contributes to the pro-atherogenic effects of FFA.     

A new, model-free calculation method to determine the coordination modes and distribution of copper(II) among the metal binding sites of multihistidine peptides using circular dichroism spectroscopy

A new calculation method to determine microscopic protonation processes from CD spectra measured at different pH and Cu(II):ligand ratios was developed and used to give the relative binding strengths for the three histidines of hsPrP(84-114), a 31-mer polypeptide modeling the N-terminal copper(II) binding region of human (homo sapiens) prion protein. Mutants of hsPrP(84-114) with two or one histidyl residues have also been synthesized and their copper(II) complexes studied by CD spectroscopy. The 1-His models were analyzed first, and the molar CD spectra for the different coordination modes on the different histidines were calculated using the general computational program PSEQUAD. These spectra were deconvoluted into the sum of Gaussian curves and used as a first parameter set to calculate the molar spectra for the different coordination modes (3N and 4N coordination) and coordination positions (His85, His96 and His111) of the 2-His peptides. The calculation method therefore does not require the direct use of CD spectra measured in the smaller peptide models. This is a significant improvement over earlier calculation methods. In the same runs, the stepwise deprotonation pK(mic) values were refined and the pH-dependent distribution of copper(II) between the two histidines was determined. The results revealed the high, but different copper(II) binding affinities of the three separate histidines in the following order: His85 < His96His111. The calculation also showed that molar CD spectra which belong to the same coordination mode and coordination position in different ligands have very similar transition energies but different intensities. For this reason, direct transfer of molar CD spectra between different ligands may be a source of error, but the pK(mic) values and the copper(II) binding preferences are transferable from the 2-His peptides to the 3-His hsPrP(84-114).     

Analysis of the human replication protein A:Rad52 complex: evidence for crosstalk between RPA32, RPA70, Rad52 and DNA

The eukaryotic single-stranded DNA-binding protein, replication protein A (RPA), is essential for DNA replication, and plays important roles in DNA repair and DNA recombination. Rad52 and RPA, along with other members of the Rad52 epistasis group of genes, repair double-stranded DNA breaks (DSBs). Two repair pathways involve RPA and Rad52, homologous recombination and single-strand annealing. Two binding sites for Rad52 have been identified on RPA. They include the previously identified C-terminal domain (CTD) of RPA32 (residues 224-271) and the newly identified domain containing residues 169-326 of RPA70. A region on Rad52, which includes residues 218-303, binds RPA70 as well as RPA32. The N-terminal region of RPA32 does not appear to play a role in the formation of the RPA:Rad52 complex. It appears that the RPA32CTD can substitute for RPA70 in binding Rad52. Sequence homology between RPA32 and RPA70 was used to identify a putative Rad52-binding site on RPA70 that is located near DNA-binding domains A and B. Rad52 binding to RPA increases ssDNA affinity significantly. Mutations in DBD-D on RPA32 show that this domain is primarily responsible for the ssDNA binding enhancement. RPA binding to Rad52 inhibits the higher-order self-association of Rad52 rings. Implications for these results for the "hand-off" mechanism between protein-protein partners, including Rad51, in homologous recombination and single-strand annealing are discussed.     

On the mechanism of human stefin B folding: I. Comparison to homologous stefin A. Influence of pH and trifluoroethanol on the fast and slow folding phases

The folding of human stefin B has been studied by several spectroscopic probes. Stopped-flow traces obtained by circular dichroism in the near and far UV, by tyrosine fluorescence, and by extrinsic probe ANS fluorescence are compared. Most (60 ± 5%) of the native signal in the far UV circular dichroism (CD) appeared within 10 ms in an unresolved “burst” phase, which was followed by a fast phase (t = 83 ms) and a slow phase (t = 25 s) with amplitudes of 30% and 10%, respectively. Similar fast and slow phases were also evident in the near UV CD, ANS fluorescence, and tyrosine fluorescence. By contrast, human stefin A, which has a very similar structure, exhibited only one kinetic phase of folding (t = 6 s) detected by all the spectroscopic probes, which occurred subsequent to an initial “burst” phase observed by far UV CD. It is interesting that despite close structural similarity of both homologues they fold differently, and that the less stable human stefin B folds faster by an order of magnitude (comparing the non-proline limited phase). To gain more information on the stefin B folding mechanism, effects of pH and trifluoroethanol (TFE) on the fast and slow phases were investigated by several spectroscopic probes. If folding was performed in the presence of 7% of TFE, rate acceleration and difference in the mechanism were observed. Protein 32:296–303, 1998. © 1998 Wiley-Liss, Inc.