Non-Native Î±-Helices in the Initial Folding Intermediate Facilitate the

Subscriber access provided by Columbia University Libraries

Article

Non-native #-Helices in the Initial Folding Intermediate Facilitate the Ordered Assembly of the #-Barrel in #-Lactoglobulin Kazumasa Sakurai, Masanori Yagi, Tsuyoshi Konuma, Satoshi Takahashi, Chiaki Nishimura, and Yuji Goto Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00458 • Publication Date (Web): 10 Aug 2017 Downloaded from http://pubs.acs.org on August 13, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Non-native α-Helices in the Initial Folding Intermediate Facilitate the Ordered Assembly of the β-Barrel in β-Lactoglobulin Kazumasa Sakurai*,†,‡, Masanori Yagi§, Tsuyoshi Konuma‡,#, Satoshi Takahashi∥, Chiaki Nishimura⊥, Yuji Goto‡ †

High Pressure Protein Research Center, Institute of Advanced Technology, Kindai University,

930 Nishimitani, Kinokawa, Wakayama 649-6493, Japan ‡

Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan

§

Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-

0871, Japan ∥

Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, 2-1-1

Katahira, Aoba, Sendai, Miyagi 980-8577, Japan ⊥

Faculty of Pharmaceutical Sciences, Teikyo Heisei University, 4-21-2 Nakano, Nakano-ku,

Tokyo 164-8530, Japan *To whom correspondence should be addressed. Phone: +81-736-77-0345, Fax: +81-736-77-4754, E-mail: [email protected].

ACS Paragon Plus Environment

1

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 41

ABSTRACT: The roles of non-native α-helices frequently observed in the initial folding stage of β-sheet proteins have been examined for many years. We herein investigated the residue-level structures of several mutants of bovine β-lactoglobulin (βLG) in quenched-flow pH-pulse labeling experiments. βLG assumes a collapsed intermediate with a non-native α-helical structure (I0) in the early stage of folding, although its native form is predominantly composed of β-structures. The protection profile in I0 of pseudo-wild type (WT*) βLG was found to deviate from the pattern of the “average area buried upon folding” (AABUF). In particular, the protection at the A-strand region, at which non-native α-helices form in the I0 state, was significantly low compared to AABUF. G17E, the mutant with an increased helical propensity, showed a similar protection pattern. In contrast, the protection pattern for I0 of E44L, the mutant with an increased β-sheet propensity, was distinct from that of WT* and resembled the AABUF pattern. Transverse relaxation measurements demonstrated that the positions of the residual structures in the unfolded states of these mutants were consistent with those of the protected residues in the respective I0 states. Based on the slower conversion of I0 to the native state for E44L to that for WT*, non-native α-helices facilitate the ordered assembly of the β-barrel by preventing interactions that trap folding.


2

Page 3 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

The roles of non-native intermediates in protein folding, typically observed as α-helical intermediates mainly for β-sheet proteins, remain controversial.1, 2 Previous studies reported that small proteins exhibit two-state folding involving only the native and unfolded states, whereas larger proteins with more than approximately 100 residues fold via a number of intermediate species.1-5 In a class of large proteins, such as apomyoglobin and cytochrome c, the intermediate states partly share the final native structure, indicating a stepwise folding pathway through obligate folding intermediates. However, another class of large proteins sometimes forms nonnative structures in intermediate states.1,

2, 4, 6

For example, the folding intermediates of β-

lactoglobulin (βLG) involve a significant amount of non-native α-helices. The necessity of the non-native α-helix for the folding of βLG has been occasionally questioned, because βLGs from different species showed different folding behaviors including the positions of the non-native αhelices.7-13 If these non-native interactions are proven to have a certain role for protein folding, the amino acid sequence influences not only the native conformation, but also the order of formation of native interactions. Bovine βLG (hereafter simply βLG unless specifically noted) has been used as a prototypical protein in investigations on non-native intermediates in protein folding. βLG consists of 162 amino acid residues (18 kDa) and contains two disulfide bonds (C66-C160 and C106-C119) and a free thiol (C121). βLG consists of nine β-strands (A-I) and one major α-helix (Figure 1).14-16 The E, F, G, and H strands and C-terminal half of the A strand consist of a stable β-sheet (core sheet), whereas the B, C, and D strands and N-terminal half of the A strand comprise a less stable β-sheet (opposite sheet). These two sheets are stacked with each other, forming an up-anddown β-barrel (Figure 1). Previous studies suggested that βLG folds through the collapse and search mechanism, with the collapse phase containing two steps.7, 8, 10, 17-19 The first step of the


3

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 41

collapse phase occurs within 1 millisecond. This step involves the formation of the folding core of the native structure at the F, G, and H strands and the major helix,10 and a significant amount of non-native α-helices in the region from the A to D strands possessing the local α-helical preference.7,

8, 17, 18

The B- to H-strand regions have been suggested to assume a compact

domain, whereas the A-strand region may fluctuate with the non-native α-helical conformation.10, 19 The second collapse occurring in ~30 ms is the attachment of the fluctuating A-helix to the preformed compact domain. The initially and secondarily collapsed states have been denoted as I0 and I1, respectively.19 After the collapse stage, the search stage occurs in ~100 ms, in which non-native α-helices are substituted with the native β-structure stabilized by nonlocal interactions (the α-to-β transition). Despite extensive experimental investigations, direct evidence for the formation mechanism and roles of non-native intermediates in the folding of βLG is limited.

Figure 1. Ribbon diagram of the βLG structure in the native state (PDB code: 2AKQ). Residues in the α-helix, core β-sheet, and opposite β-sheet are indicated by red, deep blue, and light blue,


4

Page 5 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

respectively. Labels α and A–I denote the positions of the major α-helix and each strand, respectively. The figure was prepared using MolFeat (FiatLux, Tokyo, Japan).

We recently proposed the significance of the non-native intermediate of βLG based on the findings of folding experiments on two mutants with different secondary structure propensities; one is G17E, the α-helical propensity of which at the A strand is raised, and the other is E44L, the β-sheet propensity of which at the B strand is raised.20 The folding of E44L took longer than the pseudo-wild type (WT*), even though the folding intermediate of E44L contains a similar amount of β-structures in the native state. These findings suggest that the intermediate of E44L involves interactions that trap folding, and α-helices detected for WT* may suppresses the formation of these interactions. The suggestion further implies that the non-native intermediate of βLG has a significant role for the formation of the β-barrel in βLG. A competent method to characterize the interactions involved in initial folding intermediates at the amino acid residue level is the quenched-flow pH-pulse labeling experiment. This method can show clearly the positions of locally structured regions in the intermediate state through a protection pattern against the hydrogen/deuterium (H/D) exchange reaction. Extensive investigations on apomyoglobin and its variants indicated that the formation of local structures is mainly driven by hydrophobic interactions, and the location of these clusters may be predicted from the “average area buried upon folding” (AABUF) scores of the amino acid sequence.6, 21-24 In contrast, the relationship between AABUF and the structures of non-native intermediates has not yet been elucidated. If the protection pattern for the non-native intermediate of βLG coincides with AABUF, the formation of the non-native intermediate may be triggered by a hydrophobic interaction. However, if the protection pattern deviates from AABUF, driving


5

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 41

forces distinct from the hydrophobic interaction may also contribute to the formation of nonnative structures. In an attempt to characterize the non-native intermediates of βLG mutants at the amino acid residue level, we herein performed quenched-flow pH-pulse labeling experiments on WT* and mutant βLGs and compared protection profiles to their AABUF patterns. We also measured transverse relaxation rates in the unfolded states of these mutants because residual interactions, even those observed in the unfolded states, may contribute to the structure of the initial collapse state.25-27 The obtained data were also compared with other sequence-based scores for the propensities of secondary structures of respective residues. The results obtained revealed that the distributions of local structures in the intermediate states of WT* and G17E were not consistent with the AABUF patterns, particularly at the A-strand region, whereas that of E44L was. This discrepancy may contribute to the smooth folding of βLG by controlling the order of formation of native interactions. Our results will provide an insight into the general molecular strategy used in the folding of relatively large proteins.


6

Page 7 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

MATERIALS AND METHODS Expression and purification of WT* βLG and its mutants. In the present study, C121A βLG was treated as the pseudo-wild type (denoted as WT*),28 and the double mutants C121A/G17E and C121A/E44L were denoted as G17E and E44L, respectively. The procedure for the expression and purification of βLG mutants was reported previously.20 One change from the previous procedure is the usage of M9 synthetic medium containing [15NH4]2SO4 as the sole nitrogen source for 15N-labeled protein samples. Briefly, we used the pAED4 vector.29 The mutants were expressed as inclusion bodies in E. coli BL21(DE3) (Novagen, Inc., Madison, WI, USA) in M9 media. After the solubilization of the inclusion body by 8 M urea solution, protein solutions were diluted by 30 mM Tris-HCl (pH 8.0) containing 2 mM glutathione (oxidized form), 2 mM glutathione (reduced form), and 1 mM EDTA and incubated at room temperature for up to 1 week for refolding. The proteins were then purified with CM Sepharose Fast Flow (GE Healthcare). The eluted protein solutions were dialyzed against 4 mM HCl and lyophilized. The signal assignment of WT* in the native conditions was performed by Yagi et al.30 That in the 8 M urea conditions was conducted in this study by measuring HNCACB, CBCACONH, HNCO, and HNCACO with a Bruker AVANCE III 950 MHz instrument. The data analysis was performed with Sparky (Goddard, T.D., and Kneller, D.G., SPARKY 3, University of California, San Francisco, CA, USA). The 1H–15N HSQC signals of E44L and G17E were assigned by comparing the signals with those of WT*. The chemical shifts in the native and urea-unfolded conditions are listed in Table S1 and S2, respectively.


7

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 41

H/D exchange experiments. H/D exchange experiments were initiated by dissolving lyophilized protein into 500 µl of 20 mM Glycine-HCl (pH reading of 2.0) in D2O. The reaction was monitored by recording a series of 1H–15N HSQC spectra at 37°C for at least one month. In each spectrum, peak intensities were calculated using Sparky and analyzed assuming single exponential decay with time by Igor Pro (WaveMetrics, Lake Oswego, OR, USA). Protection factors (PFs) were calculated as a ratio between the intrinsic exchange rate constant predicted from the amino acid sequence and the observed rate constant.31, 32

D/H exchange pulse labeling. Quenched-flow pH-pulse labeling experiments were performed at 4°C with an SFM-400/QS instrument (Bio-Logic, Claix, France). Samples containing ~1.5 mg/mL protein in the presence of 3.84 M GdnHCl and 50 mM glycine-HCl (pH reading of 2.0) in D2O were incubated overnight in advance of labeling to substitute all amide protons with deuteron. Refolding was initiated by an 8.5-fold dilution into refolding buffer in H2O at pH 3.0, where the exchange of amide groups is too slow to occur. The solution was kept for 9.8 ms, which is the minimal duration at this point for the apparatus used, for the refolding of proteins. After the refolding period, the pH of the solution was increased to 9.7 by mixing with 0.60 M NaOH to accelerate D/H exchange reactions. The durations of the labeling pulses were 6.2, 8.8, 17.7, and 25.9 ms. At this pH, the exchange of the amide group is very fast and all of the deuterons in the amide groups of unstructured and exposed residues will be rapidly exchanged with protons during this period. The pH of the solution was then reduced to 3.2 by mixing with 0.44 M HCl to quench the D/H exchange reaction. Under these conditions, the refolding reaction was completed. GdnHCl concentrations were kept at 0.45 M throughout the refolding reaction.


8

Page 9 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Samples were then concentrated by ultrafiltration and buffer-exchanged using PD-10 desalting columns (GE Healthcare Life Sciences) equilibrated with diluted HCl at pH 2.5 at 4°C. The solution was frozen in liquid nitrogen and stored at -80°C until NMR measurements. Prior to NMR measurements, frozen solutions were melted, and 50 µL of 4 mM DCl in D2O was added. The volume of the solution was adjusted to 500 µL by adding 4 mM HCl. The HSQC spectra of these samples were measured with a Bruker AVANCE III HD 800 MHz instrument at 30°C. The proton occupancies for each residue were calculated by normalizing the signal intensities obtained with those from the initial and final reference samples. The proton occupancies were dependent on the durations of the labeling pulses due to a slight D/H exchange under the EX1 regime in the structured species. They were extrapolated to zero duration to obtain A0 values, free from the effects of the D/H exchange during the pulse labeling (see Figure S1).6, 21, 23 The sample not treated in D2O was prepared as the initial reference state (A0 = 0). A sample pulselabeled 2 h after the initiation of refolding was also prepared as the final reference (A0 = 1). The proton occupancies obtained in the present procedure will directly reflect the populations of the structured amide groups of respective residues after the 10 ms folding time. These data analyses were also performed with Sparky and Igor Pro.

Measurements of R2 in the denatured state. Measurements for the rate constants of transverse relaxation (R2) of the 15N nucleus of backbone amides were performed with the pulse sequence described by Farrow et al.

33

using a Bruker DRX500 instrument. This experiment

included a series of 10 experiments with transverse decay times ranging between 7.2 ms and 244.8 ms.


9

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 41

Previous studies reported that the R2 rates of the ith residue of fully unfolded polypeptides containing disulfide bonds (R2,unfolded) are empirically expressed by the following equation34, 35:

, = , ∑

| |

+ ,

()* |$%&' | !" ∑+ −

(1)

where R2,int is a measure of the intrinsic relaxation rate, λ is a measure of the persistence length of the polypeptide chain in residues, N is the total chain length of the polypeptide in residues, R2,exch is the amplitude of the exchange contribution from the disulfides, Cysk is the kth cysteine residue, and NCys is the total number of residues that are linked by disulfide bonds. In the present study, λ, R2,int, and R2,exch were assumed to be 6, 0.3 s-1, and 4.2 s-1, respectively.


10

Page 11 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

RESULTS H/D exchange on native states. In order to obtain structural information on the folded states of WT* βLG and the two mutants, we performed conventional H/D exchange experiments. WT* βLG showed the same pattern of protection factors (PF) as that previously reported30; the residues that composed the secondary structures showed stronger protection (Figure 2a). The residues with the highest PF values gathered on the C, F, G, and H strands (Figure 2a, indicated by *), confirming that the folding core regions are kept rigid in the native structure. While the PF pattern of βLG was found to be largely similar to its AABUF pattern, PF values at the AB-loop (residues 35-45) and I strand (residues 145-155) were lower than those expected from the AABUF profiles. These regions are involved in the dimer interface of the βLG homodimer. Under the present experimental conditions involving an acidic pH, βLG assumes a monomer form, and these regions are exposed to the solvent.36, 37. Accordingly, the PF pattern of WT* βLG in the native state is consistent with its folded structure. The two mutants demonstrated the contrasting effects of the mutations on their native conformations. G17E showed a significantly lower protection pattern than WT* (Figure 2b, d). This result was consistent with our previous findings showing that the native state of G17E is significantly destabilized and is in equilibrium with a small amount of the equilibrium intermediate state even in the absence of a denaturant.20 However, G17E also showed significantly stronger protection at the C, G, and H strands than at the other regions (Figure 2b, indicated by *), indicating the formation of core regions similar to those of WT*. The signals for most residues on the A strand (K14-A23) were not observed, even in the first HSQC measurement, due to their rapid exchange, indicating the complete loss of protection at these residues (Figure 2b). This may have occurred because of local structural perturbations and/or the


11

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 41

overstabilization of the A helix caused by the introduction of glutamate residue at this position. In contrast, although slight decreases in PF were observed on the A and B strands, the PF pattern and the distribution of highly protected residues of E44L was very similar to that of WT* (Figure 2c, d), indicating that the E44L mutation does not affect the structure or stability of the native state. The result is consistent with previous findings showing that the native state of E44L is more stable than WT*.20

Figure 2. (a-c) PF patterns for WT* (a), G17E (b), and E44L (c) in the native state at pH* 2.0 obtained from H/D exchange experiments. The lines indicate the AABUF scores calculated from the amino acid sequences. The blue bars, red bar, and circles on the upper side of each panel


12

Page 13 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

indicate the positions of the β-strands, α-helix, and residues involved in the dimer interface, respectively. (d) Ratios of PFs for E44L (blue) and G17E (red) with respect to PFs for WT*. In panels (a)-(c), we calculated the mean and the standard deviation of log(PF), and σ, respectively, for all the measured residues of each mutant. The residues having PFs of more than 10+1.1σ were labeled by * to identify significantly protected residues.

pH-pulse D/H exchange on initial intermediate states. In order to obtain structural information on the folding intermediate states of WT* and the two mutants, we performed quenched-flow pH-pulse D/H exchange experiments at a refolding time of 9.8 ms, in which βLG is in the initially collapsed state (I0). The procedures used to calculate A0 from the data obtained are shown in Figure S1. Figure 3 shows the A0 values obtained for each residue. The normalized A0 value reflects the fraction of structured conformation at respective residue positions. In the present procedure used for pH-pulse D/H exchange experiments, an A0 of 0 means that the amide group is completely protected from exchange, while an A0 of 1 means no protection. In order to compare A0 values with the AABUF pattern, we plotted A0 values on inversed vertical axes running from 1 to 0, indicating that the higher and lower positions of the plots mean more and less fraction of protected conformations in the I0 state, respectively. It is noted that the AABUF patterns of E44L and G17E are quite similar to that of WT* (Figure S2).


13

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 41

Figure 3. Results of quenched-flow pH-pulse D/H exchange experiments. (a-c) Markers are normalized values of zero-extrapolated intensities (A0) for each residue with pulse durations from 6.2 to 25.9 ms, which indicate the degree of structural formation in the folding intermediate of WT* (a), G17E (b), and E44L (c) in the I0 state. Solid lines indicate AABUF calculated from the βLG sequence. (d) Differences in the A0 of E44L (blue) and G17E (red) with respect to that of WT*.

The intermediate states of WT* showed relatively higher populations of protected conformation at the C, G, and H strands, some of which are in the folding core regions, and lower populations of protected conformation at the A, B, and I strands. This result is consistent


14

Page 15 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

with previous findings reported by Kuwata et al.,10 showing that measurable protection was not observed at a refolding time of 10 ms on the A-strand residues. It is noted that the burst phase intermediate formed at ~140 µs was suggested to possess the non-native α-helix on the A strand; however, the PF values of the helix were marginal (

Non-Native Î±-Helices in the Initial Folding Intermediate Facilitate the

Recommend Documents