NMR study of the human NCK2 SH3 domains structure determination, binding diversity, folding and amyloidogenesis 5

In this study, the malignant hNck2 SH3-1 domain will be expressed in our system to investigate the structure properties and its amyloidogenesis.. And to test whether wild type SH3-1 stil

Chapter 5

Folding and amyloidogenesis study of hNck2 SH3-1 domain

5.1 Introduction

The aggregation and deposition of biomolecules are believed to be underlying causes

of many human diseases, including atherosclerosis, amyloidosis, gallbladder and kidney stones Amyloidosis, characterised by the deposition of abnormal protein fibres in various tissues and organs, is thought to be responsible for critical diseases like Alzheimer’s disease (AD), Parkinson’s disease, Huntington’s disease, progressive muscle atrophy, and prion disease Under appropriate conditions, most proteins

examined seem capable of fibre formation in vitro, leading to the belief that this process is a generic property of polypeptides (Fandrich M, 2001) In vivo, restrained

physiological conditions confine the number of proteins capable of amyloid fibre formation to a few dozen, all of which, unfortunately, lead to some kind of other disease (Tan SY, 1994) Amyloid fibre formation occurs both intra- and extracellular and fibres are often co-aggregated with lipid membranes and calcified (Kidd M, 1963; Kidd M, 1964) The Src-homology region 3(SH3) domains are well characterised, and small protein modules of 60–85 amino acid residues that are found in many proteins involved in intracellular signal transduction The folding and unfolding of this small domain is an apparent two-state reaction under most conditions investigated so far In spite of the apparent two-state character of its folding, NMR-detected hydrogen-deuterium exchange experiments (Sadqi, M., 2002; Sadqi, M., 1999) have indicated that under native conditions, the Spc-SH3 domain undergoes a wide variety of conformational fluctuations, ranging from local motions to extensive structural disruptions affecting its core Oligomerisation of partially unfolded intermediates has been proposed to be playing a crucial role in the initial stages of amyloid formation,

by further evolving into the formation of larger aggregates, protofibrils and finally

amyloid fibrils (Dobson, C M., 2003) In this study, the malignant hNck2 SH3-1 domain will be expressed in our system to investigate the structure properties and its amyloidogenesis

5.2 Materials and methods

5.2.1 Clone and mutagenesis

The V22 insertion was obtained by PCR-based de novo design according to the

AAC04831 NCBI published protein sequence The hNck2 SH3-1 mutation constructs, including V22K, V22D and V22A, were obtained by mutagenesis The primers used

in the mutagenesis are listed below:

V22D_f: 5’-GGAACTGGATATTAAAAAAGATAACGAACGTCTGTGGC-3’ V22D_r: 5’-GCCACAGACGTTCGTTATCTTTTTTAATATCCAGTTCC-3’

V22K_f: 5’-GGAACTGGATATTAAAAAAAAAAACGAACGTCTGTGGC-3’ V22K_r: 5’-GCCACAGACGTTCGTTTTTTTTTTTAATATCCAGTTCC-3’

V22A_f: 5’-GGAACTGGATATTAAAAAAGCGAACGAACGTCTGTGGC-3’ V22A_r: 5’-GCCACAGACGTTCGTTCGCT TTTTTAATATCCAGTTCC-3’

expressed as supernatant components in the BL21 strain and purified under native conditions All proteins were further purified by HPLC

5.2.3 CD and NMR spectra of mutant SH3 domain

For the wild type SH3-1 protein, the CD and NMR samples were prepared by exchange with 5 mM phosphate buffer at pH 6.5 For SH3-1-V22, -A22, -D22 and -K22 mutants, the lyophilised protein powder was dissolved in the deionised water (Millipore, Milli-Q) with the addition of 10% D2O in the NMR samples for spin-lock Since both the high salt concentration buffer and the pH will dramatically affect the solubility of these proteins, most CD and NMR experiments were performed with a 5

buffer-mM salt concentration at pH 4.0 However, the CD and NMR HSQC experiments of the SH3-1-V22 mutant were collected at both pH 4.0 and 6.5 for comparison

CD experiments were performed on a Jasco J-810 spectropolarimeter equipped with a thermal controller as described in 2.3 The far-UV CD spectra were collected at a peptide concentration of ~50 μM at 25 ºC The near-UV CD spectra were collected at

a protein concentration of ~200 μM in the absence and in the presence of 8 M urea Five independent scans were collected and the averaged ones were recorded for further analysis

NMR experiments were acquired on an 800 MHz Bruker Avance Spectrometer equipped with pulse field gradient units at 298 K The NMR spectra acquired for both the backbone and side chain assignments included the 15N-edited HSQC-TOCSY, HSQC-NOESY and the 13C-HCCH-TOCSY, as well as triple-resonance experiments (HNCACB, CBCA(CO)NH, HNCO) NOE connectivity was identified from the 15N-

NOESY spectrum

15N T 1 and T 2 relaxation times and {1H}-15N steady-state NOEs experiments were performed on the 800-MHz Spectrometer at 298 K 15N T 1 relaxation experiments were conducted with delays of 10, 100, 200, 300, 400, 500, 600 and 700 msec 15N T2

ones had delays of 10, 60, 100, 130, 160, 200, 230, 260 and 300 msec {1H}–15N steady-state NOEs were obtained by acquiring spectra with and without 1H presaturation of a duration of three seconds and a relaxation delay of five seconds at

800 MHz NMR data were first processed with an NMRPipe and subsequently analysed and fitted by NMRView The solution structure of the Nck2 SH3-1 (2B86) was obtained from PDB and its NMR assignment with an accession code of 6854 was downloaded from BioMagResBank The graphic software MolMol was used for structure display and analysis

5.2.4 Amyloidogenesis condition screening and EM samples preparation

The SH3-1-V22, the 4AlaMut and the wild type SH3-1 at pH3.0 were investigated for their amyloid formations Protein concentration was fixed at 3mg/ml, while a series of sodium chloride concentrations: 10, 20, 50, 100, 200, 300 and 500 mM were chosen for the optimisation of amyloid formation The results were examined by TEM after seven days of incubation EM grids were processed by the Drop-To-Drop method, as follows (1) 5μl of liquid specimen was placed on the parafilm; (2) The carbon-coated

with filter paper, and the grids were placed with specimen-side up Photographs were taken from a JEOL JEM2010F Electron Microscopy equipped with CCD Magnifications were chosen as 8000× and 20000×

5.3 Results

5.3.1 Biological activity of wild type SH3-1

Initially the Nck2 SH3-1 was cloned as malignant form in which an additional Val was inserted at the tip of diverging turn Subsequent protein expressions showed that the SH3-1-V22 went into the inclusion body portion Accordingly, the additional Val residue was removed to testify whether the native tertiary structure can be restored And to test whether wild type SH3-1 still has any biological activity, the CD3ε peptide, which was previously shown to be capable of binding the Nck2 SH3-1 was also expressed as a GST-fusion protein After cleavage, purification and lyophilisation, the HSQC perturbation experiment was performed to verify the binding result Significant HSQC peak shifts were observed for the wild type in the buffer condition (50mM phosphate, pH 6.5), while no chemical shifts were detected for the four insoluble SH3-1 domains in salt-free water, indicating that the extra presence of residue at a diverging turn will abolish the structure’s compactness and biological binding activity

5.3.2 Mutagenesis study of SH3-1-V22

To clarify whether different types of residue at the same position of Val have the same deteriorating effect, three other residues with distinguished side-chain properties were introduced by mutagenesis; namely, SH3-1-K22, -A22, and -D22 The 15N

labelled samples were prepared as described before HSQC spectra of these mutants which were dissolved in pure water were acquired as shown in Figure 5.1 A-D The SH3-1-V22 HSQC spectrum at pH2.0 and pH 4.0 are superimposed in Figure 5.1 F For comparison, the wild type SH3-1, which was dissolved in the phosphate buffer at

pH 6.5, is also illustrated here A narrow dispersion ~1 ppm at 1H dimension and 18 ppm at 15N dimension indicate that SH3-1-V22, -A22, -K22 and -D22 are in an unstructured state Meanwhile, these mutants and the wild type SH3-1 were also studied by CD, as shown in Figure 5.2 The mutants were all dissolved in pure water and far-UV spectra were acquired from 260nm to 190nm The wild type SH3-1 dissolved in the buffer condition was also acquired All mutants show unstructured states according to the CD profiles, while the wild type SH3-1 shows a β sheet dominant structure From both CD and NMR preliminary studies, it showed that regardless of the residue types at the tip of a diverging turn for all mutants, it would prevent the SH3-1 domain from achieving a well folded native structure

Figure 5.1 HSQC of 4 mutants and refolded WT SH31 The SH31A22, D22, K22 and

-V22 that are dissolved in pure water HSQC spectra are shown in Figures A-D -V22del (buffer, pH6.5) HSQC is illustrated in Figure E SH3-1-V22 HSQC spectra at pH2.0 and pH 4.0 are compared in Figure F

Figure 5.2 CD Comparison of mutants and WT SH3-1-A22, -D22, -K22, -V22 and wild

type CD spectra are compared with each other Different colours are designated to different mutants and WT, as is shown in the right-hand corner

5.3.3 Assignment, secondary CS analysis and dynamics study of SH3-1-V22

As mentioned in the previous section, the insertion at the tip of the diverging turn of the SH3-1 significantly changes the structural properties and in addition, one residue

at this critical point retains the protein in a denatured state, according to CD and NMR studies It is fundamentally interesting to investigate the structural and dynamics properties of such an insoluble protein The double-labeled SH3-1-V22 sample was first prepared to conduct a sequential assignment The assigned HSQC spectrum of the SH3-1-V22 and the V22△ at 8M urea are illustrated in Figure 5.3 To reflect the real denatured situation, the V22△ was dissolved in 8M urea and the assignment was done to provide a reference for SH3-1-V22 chemical shift The secondary chemical shifts of the SH3-1-V22 for 1Hα, 13Cα, 13Cβ and 13CO were successfully obtained

To analyse the secondary chemical shifts of SH3-1-V22 and the tendency to form a secondary structure element, as well as the denatured status after the introduction of additional residue at the tip of a diverging turn, the averaged helix/strand chemical shifts both in buried and exposed states that are obtained from Franc Avbelj (Franc

Avbelj, 2004) are used for comparison As shown in Figure 5.4, the proton and carbon

secondary chemical shifts of the SH3-1-V22 at pH 4.0 and WT SH3-1 at pH 6.5 from the Wagner Group were compared (Figure 5.4 A-B) The proton and carbon secondary chemical shifts of SH3-1-V22 are also compared with those of WT SH3-1

at pH 2.0 (Figure 5.4 C-D) The alpha proton secondary chemical shifts of SH3-1-V22 are compared with those of helix at both buried and exposed states (Figure 5.4 E-F)

In the cases of the buried and exposed helix, high similarities are observed for some residues with an exposed secondary chemical shifts that are closer to the secondary chemical shift of the SH3-1-V22 From both the proton and carbon secondary

chemical shift analysis, it can be seen that the SH3-1-V22 has a high tendency to form

a helix conformation, but a reduced degree compared with WT SH3-1 at pH 2.0 Meanwhile, due to the small deviation of chemical shift in contrast with the one of

WT SH3-1, it could be concluded that the tertiary contacts are disrupted because of the introduced additional Valine residue

In order to obtain the residue specific dynamic properties of the SH3-1-V22, the 15N NMR backbone relaxation data of the SH3-1-V22 with a protein concentration of

~700 μM were collected in salt-free water (pH 4.0) at 25℃ on an 800 MHz Bruker

Avance NMR Spectrometer As shown in Figure 5.5, the HetNOE, T 1 and T 2

parameters are presented in figures A, B and C sequentially The relaxation parameters of the wild type SH3-1 are also presented and compared with those of the SH3-1-V22, as shown in Figure 5.5

A reduced spectral density mapping was calculated at 0, 1H and 15N frequencies Big differences can be observed between the SH3-1-V22 and WT at 0 and 1H frequencies, especially at the N- and C-terminals of SH3-1

Figure 5.3 Assignments of SH3-1-V22 at pH 4.0 and V22△ in 8M urea The upper panel is

the SH3-1-V22 assignment and the lower panel is the V22△ assignment The stars stand for the his-tag or unknown residues The residue-plus-bar stands for the residues in the his-tag

Figure 5.4 Secondary chemical shift comparison A) Comparison of alpha proton secondary

CS between SH3-1-V22 and the one from the Wagner Group B) Comparison of SCS_Cβ between SH3-1-V22 and the one from Wagner Group C) Comparison of alpha proton secondary CS between SH3-1-V22 and WT@ph2.0 D) Comparison of SCS_Ca- SCS_Cβ between SH3-1-V22 and WT@ph2.0 E) Comparison of alpha proton SCS between SH3-1-V22 and averaged SCS of buried helices F) Comparison of alpha proton SCS between SH3-1-V22 and averaged SCS of exposed helices (SCS = secondary chemical shift)

SCS_Ca-Figure 5.5 Dynamics parameter comparison of SH3-1-V22 with wild type SH3-1 at pH 2.0 The 15 N NMR backbone relaxation data of SH3-1-V22 with a protein concentration of

~700 μM were collected in salt-free water (pH 4.0) at 25 ºC on an 800 MHz Bruker Avance NMR spectrometer TOP: { 1 H}- 15 N steady-state NOE intensities MIDDLE: 15N T 1

(longitudinal) relaxation times BOTTOM: 15N T 2 (transverse) relaxation times SH3-1-V22 data are shown in grey Wild type SH3-1 relaxation parameters are also presented here for

5.3.4 Identification of non-native medium-range NOEs and native-like range NOEs

long-To gain an insight into structural and packing properties, we have acquired both the

15N- and 13C-edited NOESY spectra of the denatured SH3-1-V22 An NOE analysis shows to some extent the persistence of the tertiary contacts Unfortunately, due to a severe signal overlap arising from resonance degeneration, it was impossible to assign the 13C-NOESY spectrum Nevertheless, the NOEs from the 15N-NOESY spectrum can still be assigned due to the good dispersion in the 1H and 15N dimensions of the spectra Interestingly, in addition to intra-residual ones, sequential and medium-range NOEs can be observed for almost all residues In particular, 23 residues (~40% of the molecule) still owned long-range NOEs A detailed comparison

of the persistent NOEs in the denatured SH3-1-V22 with those of the published native Nck2 SH3-1 (Park S, 2006) led to a classification of the remaining NOEs into non-native and native-like categories (Tables 5.1 and 5.2) A close examination of the non-native NOEs revealed that they were all sequential and medium-range NOEs located almost all over the molecule As shown in Table 5.1, many sequential NH-NH and α/βH(i)-NH(i+2) NOEs that are characteristic of the loop/turn/nascent-helix conformations suddenly show up on five β-strands These NOEs were totally inconsistent with well-formed and rigid β-strands in the published SH3-1 structure (Park S, 2006), thus clearly indicating that in the denatured SH3-1-V22, the five β-stands became distorted, or/and lost its rigidly extended backbone conformation Similarly, non-native NOE patterns were also observed in the two RT-loop strands

On the other hand, many native-like medium-range NOEs over the β-turn/loop regions could be identified Specifically, except for the absence of long-range NOEs