Domains of Pyrococcus furiosus l‑asparaginase fold sequentially and assemble through strong intersubunit associative forces

1 3

ExtremophilesDOI 10.1007/s00792-015-0748-z

ORIGINAL PAPER

Domains of Pyrococcus furiosus l‑asparaginase fold sequentially and assemble through strong intersubunit associative forces

Dushyant K. Garg1 · Rachana Tomar1 · Reema R. Dhoke2 · Ankit Srivastava1 · Bishwajit Kundu1

Received: 27 January 2015 / Accepted: 29 March 2015 © Springer Japan 2015

Keywords Sequential folding · Subunit interaction · Co-operativity · Stability · Domain · Multimeric

AbbreviationsPfA Pyrococcus furiosus l-asparaginaseNPfA N-terminal domain of PfACPfA C-terminal domain of PfASEC Size-exclusion chromatographyCD Circular dichroismWT Wild type

Introduction

Multimeric multidomain proteins fold via different path-ways, often forming complex intermediates (Auton et al. 2007; Powers et al. 2007; Kishore et al. 2012). Understand-ing these pathways is useful in modulating the stability of a protein. Owing to high stability against variety of dena-turing conditions, the thermophilic proteins have interested researchers as models to study protein folding (Luke et al. 2007; Takano et al. 2013). Previously, we had reported a thermostable, pH stable and proteolysis resistant L-aspara-ginase from Pyrococcus furiosus (PfA) (Bansal et al. 2010, 2012). Therefore, for successful clinical and industrial use of this l-asparaginase, understanding its folding and sta-bility is crucial (Kotzia and Labrou 2009). We studied the same using the wild-type (WT) enzyme, its mutants and the isolated domains.

Most proteins undergo cooperative folding in millisec-ond time scales. This complicates the accurate determi-nation of their folding kinetics (Zhang and Chan 2010; Lindorff-Larsen et al. 2011). In these reports, species dis-criminate between unfolded monomer and folded dimer, without the folded monomeric form being measurably

Abstract Here, we report the folding and assembly of a Pyrococcus furiosus-derived protein, l-asparaginase (PfA). PfA functions as a homodimer, with each monomer made of distinct N- and C-terminal domains. The purified individual domains as well as single Trp mutant of each domain were subjected to chemical denaturation/renatura-tion and probed by combination of spectroscopic, chroma-tographic, quenching and scattering techniques. We found that the N-domain acts like a folding scaffold and assists the folding of remaining polypeptide. The domains dis-played sequential folding with the N-domain having higher thermodynamic stability. We report that the extreme ther-mal stability of PfA is due to the presence of high inter-subunit associative forces supported by extensive H-bond-ing and ionic interactions network. Our results proved that folding cooperativity in a thermophilic, multisubunit pro-tein is dictated by concomitant folding and association of constituent domains directly into a native quaternary struc-ture. This report gives an account of the factors responsible for folding and stability of a therapeutically and industri-ally important protein.

Communicated by F. Robb.

Electronic supplementary material The online version of this article (doi:10.1007/s00792-015-0748-z) contains supplementary material, which is available to authorized users.

* Bishwajit Kundu [email protected]

1 Kusuma School of Biological Sciences, Indian Institute of Technology Delhi, Hauz Khas 110016, New Delhi, India

2 CSIR Institute of Microbial Technology Chandigarh, Chandigarh, India

Extremophiles

1 3

populated at equilibrium (Galvagnion et al. 2009). Con-trasting reports of folded monomeric intermediates assem-bling into dimeric molecule are also found (Di Venere et al. 2011; Maity et al. 2005). These reports largely relied on equilibrium denaturation experiments, probed by meas-urement of Trp fluorescence or by far-UV CD. In multid-omain, multisubunit proteins, such simplistic representation is often misleading as in some cases the domains fold first (fast process) followed by the association of subunits (rela-tively slow process) (Cheng et al. 1993). Domain-swapped proteins are exceptions (Rousseau et al. 2001). Unless suf-ficiently long lived, it is almost impossible to detect fold-ing intermediates of these proteins from equilibrium dena-turation experiments, as it gives time-independent averaged values of the detected parameters. In these cases, fast kinetic parameters, determined by complex stop-flow spec-troscopy, are applied to detect the presence of intermediates (Dams and Jaenicke 1999; Zitzewitz et al. 1995), However, by adopting suitable tagging method and clever manipula-tion of intrinsic probes, folding pattern of proteins can be proposed with fairly high degree of confidence by equilib-rium denaturation (Santra et al. 2005; Harris et al. 2014).

This report provides insights on the folding mechanism of PfA by adopting such methods. PfA is a homodimeric protein in which each monomer comprises of a relatively large N-terminal domain (residues 1–182), and a C-termi-nal domain (residues 202–326). Interestingly, each domain contains a single Trp residue, giving us the necessary spectroscopic probe to monitor the folding of individual domain. We report that association between the monomers of PfA begins after the N-terminal domain of each subunit attains partially folded structures. Complete folding of the N-domain proceeds in a cooperative manner concomitant with the folding of the entire molecule. The C-terminal domain, however, shows poor co-operativity by itself and relies on the assistance provided by the N-domain. Thus, we showed that non-covalent associative forces between mono-mers can be strong enough to dictate folding co-operativity. Moreover, such forces may supersede the forces contribut-ing to stability of individual domains within each monomer.

Materials and methods

Cloning, expression and purification of proteins

WT PfA clone earlier developed in our lab was used as a template for developing the domain-specific Trp mutants (Bansal et al. 2010). PCR-based site-directed mutagenesis was carried out using primer 5′CTTATCCAGCCAGAAG ATTTTGTAGATCTTGCTGAAAC3′ and its complement, and primer 5′ACAGTAACAAAGCTCATGTTTATTCTAG GCCACACAA3′ and its complementary pair for obtaining

W60F (N-terminal Trp mutant) and W301F (C-terminal Trp mutant), respectively. For cloning individual domains, the primers, restriction sites and methodology used are pre-sented elsewhere (Tomar et al. 2013, 2014). Transformed E. coli cells (Rosetta DE3), grown by shaking at 37 °C in LB medium (HiMedia) containing 50 µg ml−1 kanamycin and 17 µg ml−1 chloramphenicol (Sigma), were induced with 1 mM IPTG (Sigma) (at A600 of 0.6) and harvested 14 h post-induction by centrifugation, followed by sonica-tion. Expression was checked on 12 % SDS-PAGE. All the proteins were purified by standard Ni–NTA-affinity proce-dure explained elsewhere (Bansal et al. 2012; Tomar et al. 2013, 2014). The molecular weights of purified proteins were verified by MALDI TOF mass spectrometry (Bruker). Protein concentrations were determined by UV absorb-ance at 280 nm using molar extinction coefficients (ε280)—27515 and 21890 M−1 cm−1for the WT and Trp mutants, respectively, whereas for NPfA and CPfA, ε280 used were 14,440 and 12,950 M−1 cm−1, respectively.

Determination of secondary structure

For determining the secondary structures, each pro-tein sample (0.2 mg ml−1) was taken in a quartz cuvette of 1 mm path length and their far-UV CD spectrum was recorded from 250 to 200 nm in a spectropolarimeter (Jasco J-815). A spectral band width of 5 nm was used. The WT protein and Trp mutants were prepared in 25 mM sodium phosphate (pH 7.4) containing 20 mM NaCl, while domains were prepared in 25 mM Tris buffer (pH 9.0) con-taining 50 mM NaCl. An average of three scans was plotted against wavelength.

Activity assay

Activity of the proteins was measured by standard Nessleri-zation method as detailed elsewhere (Wriston 1985). Enzy-matic activity was obtained as a measure of absorbance at 480 nm. One international unit (IU) of l-asparaginase activity was defined as the amount of enzyme liberating 1 μmol NH3 in 1 min. Specific activity of l-asparaginase was defined as units per milligram protein.

Equilibrium unfolding/refolding studies

For all unfolding and refolding studies, an 8 M stock of ultrapure guanidine hydrochloride (Sigma-Aldrich St Louis, MO, USA) was used. For unfolding, protein samples were prepared in respective buffers (25 mM Tris, 50 mM NaCl pH 8.0; for CPfA, pH 9.0), containing varying concentra-tions of GdnCl, followed by 3 days incubation at 60 °C. For refolding, protein samples unfolded in 6-M GdnCl for 24 h were stepwise diluted by adding to corresponding refolding

Extremophiles

1 3

buffer, to reach a final protein concentration of 0.2 mg ml−1. Fluorescence scans of each sample taken in a 1-cm cuvette were recorded between 310 and 400 nm, after excitation at 295 nm, in a LS 55 spectrofluorimeter (Perkin Elmer, USA). The excitation and emission slits were fixed at 5 nm each. The ratio of fluorescence emission at 353 and 333 nm was calculated and plotted against increasing concentration of GdnCl. For the WT protein, the spectral center of mass from the fluorescence scans as well as the change in ellipticity at 222 nm was plotted against increasing GdnCl. The standard free energy change of unfolding (G0

D) was obtained by linear extrapolation of the plot of free energy change and denatur-ant concentration to 0 M.

Data analysis

The folding pattern of dimer to monomer unfolding can be approximated by the following equation:

where N2 and U are the molar concentration of dimers and monomers, respectively, and Ku is the unfolding constant.

Using a nonlinear least-squares method, the entire data of each denaturant-induced transition curve were analyzed for G0

D, mg and Cm using the relation (Santoro and Bolen 1988):

where yN and yD are optical properties of the native and denatured protein molecules, G0

D is the value of free energy change in the absence of denaturant, m is the slope, R is the universal gas constant and T is the temperature in Kelvin. The percentage fraction of unfolded protein (Fu %) was calculated using the equation:

Fluorescence spectral center of mass (SCM) was calcu-lated from intensity data measured over the entire spectrum.

where λ is the wavelength and Iλ is fluorescence intensity at λ.

Acrylamide quenching experiments

The accessibility of tryptophan in full-length proteins was determined by acrylamide quenching studies. For this study, we used protein samples that were denatured with

N2Ku↔ 2U,

(1)y = yN + yDx Exp [−(G0

D − m[GdnCl])/

RT]

/(1 + Exp[

− (G0D − m[GdnCl])/RT])

(2)GD = (G0D) + m[GdnCl],

(3)Fu% = [(y − yN )/yD − yN )] × 100.

(4)SCM =∑

�.I�/I�,

various concentrations of GdnCl (as mentioned above). Acrylamide from a stock solution of 8 M was added to these samples to obtain final concentrations varying from 100 to 500 mM. Samples were incubated for 5 min before taking fluorescence scans from 310 to 400 nm with exci-tation at 295 nm. A ratio of fluorescence intensities in the absence (F0) and presence of quencher (F) was plotted against increasing quencher concentrations. The dynamic quenching constants were computed from the Stern–Vol-mer equation (Eftink and Ghiron 1976):

where [Q] is the quencher concentration and KSV is the dynamic quenching constant.

Size‑exclusion chromatography

To determine the oligomeric nature and subunit associa-tions, the WT protein was incubated as earlier and ana-lyzed by SEC. Samples at 0.2 mg ml−1 were passed through a pre-equilibrated Superdex 200 column fitted to an Akta FPLC system (GE healthcare) at a flow rate of 0.5 ml min−1. The molecular weight of species correspond-ing to each elution peak was compared with standard curve. In addition, to distinguish between monomeric and dimeric PfA, SDS-denatured sample was also run through the same column pre-equilibrated with SDS-containing buffer.

SEC‑MALS analysis

To confirm the molecular weight of the species corre-sponding to the 5-M GdnCl-denatured protein, 100 µl of 1 mg ml−1 of the sample was loaded on a superdex 200, 10/300 SEC column operated at 25 °C and was coupled to a DAWN HELIOS II™ MALS instrument (Wyatt Technol-ogy). The MALS was equipped with an 18-angle light scat-tering detector and operated at 60 °C using the inbuilt peltier controller system. The flow rate of elution was 0.5 ml min−1. BSA denatured with 5 M GdnCl and non-denatured pro-teins were run as controls. The dn/dc value for protein in the native buffer was taken as 0.185 ml g−1. For the sample denatured with 5-M GdnCl, the corresponding dn/dc value was determined using inbuilt differential refractometer.

Analysis of PfA dimeric interface

The structural coordinates of WT PfA was analyzed using the recently reported crystal structure (Pdb: 4q0 m). The 2D-GraLab and PDB-PISA programs were utilized to obtain information on the non-covalent interactions pre-sent at the dimeric and monomeric interface of the protein (Zhou et al. 2009; Krissinel and Henrick 2007). Further analysis and representations of the structures were done

(5)Fo/F = 1 + Ksv[Q],

Extremophiles

1 3

using the Discovery studio and VMD software packages (Humphrey et al. 1996).

Results

Protein size, secondary structure and activity

The molecular weights of all the purified proteins were ana-lyzed on SDS-PAGE and confirmed by MALDI mass spec-trometry. Both the WT and mutant (W301F and W60F) proteins showed peak at ~36.5 kDa, corresponding to mono-mer. Additional low intensity peak corresponding to dimeric full-length protein of molecular weight ~74 kDa was also observed. Refolded NPfA and CPfA domains displayed peaks of 22.1 and 16.1 kDa, respectively (Fig. 1a). Far-UV CD spectra of the WT and the single Trp mutants overlapped, whereas those of isolated NPfA and CPfA domains were dif-ferent and predominated by β-sheet character (Fig. 1b). No activity was detected for the individual domains whereas the specific activities of the single Trp mutants remained same as the WT. This indicates that the mutations have no effect on the secondary and tertiary structure of proteins.

Folding mechanism of PfA

Equilibrium unfolding experiments

To understand the mechanism of folding of PfA, equilib-rium unfolding/refolding of proteins was monitored as

plots of Trp fluorescence against increasing denaturant concentrations. To nullify the concentration dependency of denaturation midpoint (Cm), WT PfA was monitored at three different protein concentrations, from which an inter-mediate working concentration of 0.2 mg ml−1 was chosen for all the proteins (data not shown). Unlike the W60F and the isolated C-domain, the overlayed unfolding and refold-ing curves of other proteins displayed hysteresis when ana-lyzed after overnight incubation at 25 °C (Supp. Fig. 2a, b). However, upon further incubation at 60 °C for 3 days, the transition phase of the unfolding curves moved toward lower GdnCl concentrations and began to overlap with the refolding curves, indicating attainment of equilibrium. The refolding curves remained unchanged regardless of the period of incubation (Supp. Fig. 2c). In the non-denatured condition, the emission λmax for WT, W301F and W60F was 333, 332 and 338 nm, respectively, with relatively lower fluorescence yield of the W60F (Fig. 2a–c). The red-shifted λmax and lower fluorescence intensity of W60F indi-cated that the Trp of the C-terminal is not only more solvent exposed but natively quenched by the surrounding residues. Each protein displayed a threshold of GdnCl concentration, beyond which their fluorescene λmax exhibited a distinct red shift. The shift was at 3.5 M for the W301F mutant but was at 3 M for the W60F mutant (Fig. 2a, b). However, for the WT, this shift was found at 3.25 M, indicating an averaged λmax representing Trps of both the domains (Fig. 2c). This 0.25 M stepwise shift in λmax indicates the relative stabili-ties of each domain in the context of the full-length protein. The N-terminal domain (W301F), being the most stable,

Fig. 1 Molecular mass, second-ary structure and activity of pro-teins. a MALDI mass spectra showing peaks corresponding to molecular weights in Dalton. b CD spectra showing overlap-ping curves of WT, W301F and W60F, indicating structural similarity and non-overlapping curves of CPfA and NPfA. c Activity of WT, W301F and W60F showing similar profiles (~26 U mg−1). Isolated NPfA and CPfA domains did not show activity

Extremophiles

1 3

the C-terminal domain (W60F) was found as the least sta-ble domain, while the WT showed intermediate stability. In the case of isolated domains, emission λmax of NPfA and CPfA was 338 and 342 nm, respectively, indicating that in these cases, the Trps are relatively more exposed com-pared to their respective full-length counterparts (Fig. 2d, e). Fluorescence scans of NPfA displayed slight transition at 2-M GdnCl above which there was a gradual decrease in fluorescence intensity along with red shift (Fig. 2d). In the case of CPfA, no such abrupt transition was observed, rather a monotonous decrease and red shift were visible (Fig. 2e). These observations point toward the occurrence of low extent of cooperative unfolding in the case of NPfA, while in the case of CPfA co-operativity was even lesser.

The optical signals normalized to percentage unfolded fractions were plotted as a function of denaturant concen-trations and fitted using the non linear Eq. (1) (Fig. 3). As per the fit, the WT, W301F and the W60F mutants showed Cm of 3.04, 3.18 and 2.8 GdnCl, respectively (Table 1). To achieve unfolding, in contrast to W60F, the W301F required higher denaturant concentration (0.2 M) when compared with that of the WT protein. This clearly indi-cated that the N-domain was more stable. We reached the same inference when unfolding of individual domains was monitored (Fig. 3b). Both the N and CPfA started to

show unfolding transition from the beginning and reached an apparent plateau by 5-M GdnCl. The curves of the WT as well as the Trp mutants were consistent with a two-state folding behavior. This was further confirmed by monitoring the unfolding using triple spectral probes where far-UV CD spectra, 353/333 fluorescence intensity ratio and fluores-cence spectral center of mass were found overlapping for the WT protein (Fig. 3c). The poorly resolved pre-transition baselines of the isolated domains indicated low cooperativ-ity of folding within the domains. They were, however, fit-ted using a dynamic fit wizard of Sigma plot (Fig. 3b).

The various thermodynamic parameters calculated using Eq. (1) are given in Table 1. The higher m and G0

D values of the W301F mutant and the isolated NPfA indicated higher co-operativity and higher stability, respectively, of the N-domain as compared to the C-domain (Table 1). Moreo-ver, in isolation, each domain displayed poor co-operativity and exhibited lower m values as compared to the WT and the mutant proteins.

Fluorescence quenching experiments

Information obtained from equilibrium denaturation experiments were confirmed by Trp fluorescence quench-ing experiments. In the case of the W301F mutant, the

Fig. 2 Tryptophan fluorescence of PfA and its variants. Trp emission scans of each protein (0.2 mg ml−1) incubated with varying GdnCl concentration (indicated by colored symbols) monitored after excita-

tion at 295 nm. The λmax of the native and fully unfolded samples is shown in each curve

Extremophiles

1 3

quenching pattern displayed Stern–Volmer plots with two distinct clusters of slopes; one from 0 to 3.0 M, and the other above 3.5-M GdnCl with an intermittent slope at 3.25 M (Fig. 4a). Similarly, for the W60F mutant, an inter-mittent Ksv value at 2.75-M GdnCl distributed between two clusters of 2.5 and below and 3 M and above was noted (Fig. 4b). However, the WT PfA displayed two distinct clusters; one between 0 and 3 M and the other between 3.25 and 5 M (Fig. 4a). The two distinct clusters indicate differ-ential access of the quencher to the Trps, with no access up to 3.0 M, and full access beyond 3.25 M. The cluster-ing also represents folded and unfolded states, suggesting a two-state unfolding pattern for WT PfA. These data were in full agreement with the data obtained from equilibrium denaturation experiments. Transition points in equilib-rium denaturation curves obtained from fluorescence data (Figs. 2a–c, 3a) matched with the transition in Ksv values obtained from quenching plots (Tables 1, 2).

Oligomeric nature and association properties

WT PfA when passed through size-exclusion column eluted at 14.1 ml, corresponding to a dimer. In the presence of increasing GdnCl up to 3.0 M, the protein eluted as sin-gle peak (peak I) with a gradual shift toward lower elution

volumes (Fig. 5a). This trend represents a homogeneous population having progressively increasing Stoke’s radius. At 3.25-M GdnCl, a second, higher molecular weight peak appeared (Fig. 5a, Peak II). With further increase in GdnCl, the intensity of peak II increased with a concomitant decrease in the intensity of the peak I. Interestingly, we did not observe any peak at elution volumes higher than 14.1 ml. The peak II therefore, may represent either a swollen dimer that never dissociated into monomers, or a monomer denatured to an extent that its Stoke’s radius is higher than the native PfA. To determine which of these possibilities exist, PfA was incu-bated overnight with 1 or 2 % SDS-containing buffer and monitored for its migration pattern on both SDS-PAGE as well as on SEC. When electrophoresed on SDS-PAGE, the WT PfA migrated as monomer of ~37 kDa (Fig. 5b). The position of the migrated samples was the same with or with-out boiling. When the same sample was passed through the SEC column pre-equilibrated with 1 or 2 % SDS-containing buffer, it eluted as a single peak at 12.9 ml, which was equiva-lent to the elution volume of a 5-M GdnCl-incubated sample (Fig. 5c). Together, these experiments indicated that the 5-M GdnCl-incubated sample is actually a population of unfolded monomer and not an un-dissociated dimer.

Since the SEC could not be performed at 60 °C, the size predictions were verified by re-analyzing the samples

Fig. 3 Equilibrium unfolding profile of proteins. Fractional unfold-ing curves of a WT, W301F and W60F proteins and b of individual domains (indicated by symbols as given in respective panels) as a function of denaturant concentration. The curves were fitted into a

two-state model using Eq. (2) (solid lines). c Percentage normalized equilibrium denaturation curves of WT PfA deduced by CD, SCM and 353/333 ratio (indicated by respective symbols) showing com-plete overlap, indicating two-state unfolding behavior

Table 1 Thermodynamic parameters calculated from the two-state fits

The raw optical signals were fitted using non linear Eq. (1) and ΔG, Cm and m values were derived using Eq. (3), as described in materials and methods section. In all cases, the protein concentration used was 0.2 mg ml−1

Parameters WT W301F W60F NPfA CPfA

ΔGH2O (Kcal mol−1) 15.74 ± 0.62 19.72 ± 0.44 11.93 ± 0.81 5.7 ± 0.44 1.6 ± 0.35

D1/2 (M GdnCl) 3.04 ± 0.01 3.18 ± 0.03 2.82 ± 0.1 2.01 ± 0.13 1.99 ± 0.08

m (Kcal mol−1 M−1 GdnCl) 5.12 ± 0.62 6.21 ± 0.62 4.23 ± 0.25 2.9 ± 0.21 0.9 ± 0.05

Extremophiles

1 3

through MALS, operated at elevated temperature. Iden-tical to the SEC data, we observed a single peak of 76.3 ± 2.2 kDa, matching with the molecular weight of a dimeric native protein (Fig. 5d). For the 5-M GdnCl-incu-bated sample, a single peak (peak II) corresponding to the molecular weight of a monomeric species was observed. From these experiments, we concluded that during GdnCl-induced unfolding the dimeric PfA dissociates directly into an unfolded monomer.

Nature of associative forces contributing to PfA stability

The protein resisted thermal denaturation and retained activity at high temperature (Fig. 1c). This indicated the

presence of high intersubunit associative forces between the constituent domains. Analysis of the PfA structure showed that the monomers aligned in a head-to-tail fash-ion with the N-domain of one monomer NPfAI in close proximity to the C-domain of another monomer CPfAII (Suppl Fig. 1a). The two C-domains, CPfAI and CPfAII were found to align centrally Analysis revealed that for individual domains, the intra-monomeric H-bonding and hydrophobic interactions are more as compared to inter-monomeric domain-wise interactions (Table 3). However, when all the inter-monomeric interactions are combined (NPfAI-CPfAII, CPfAI-CPFAII and NPfAII-CPfAI), they far exceeded the intra-monomeric interactions. The num-bers of salt bridges between the monomers were also found to be more than those within each monomer (NPfA-CPfA). Together with hydrophobic interactions, these possibly contributed to the observed thermal stability and associa-tion-induced folding of N-terminal domain.

Discussion

The folding mechanism of multimeric, multidomain pro-tein is complex. The complexities arise from the relative stabilities of constituent domains, domain–domain interac-tion and the presence/absence of linker. In this study, we selected PfA, a thermophilic, homodimeric, two-domain protein and elucidated its mechanism of folding. In our studies, we used isolated domains as well as the single Trp mutant as structural reporter of each domain. In multi-domain proteins, one of the domains has been shown to assist the folding of the other domain. This could either be due to the stabilization effect of one domain over the other or, to internal chaperoning effect (Mills et al. 2007;

Fig. 4 Stern–Volmer plots of acrylamide quenching experiments. Each linear plot represents fluorescence quenching data with increas-ing acrylamide concentration. Before adding the quencher the protein sample W301F (a), W60F (b) and WT (c) were incubated with differ-ent concentrations of GdnCl as indicated against each curve. Quench-ing plots of all the 3 proteins were distributed in 2 clusters with one

intermittent plot corresponding to 3.25 M in W301F and 2.75 M in W60F. Ksv values were derived from the slopes of each curve. Maxi-mal exposure of Trp was achieved at 3.25-M GdnCl in case of WT, 3.5 M in case of W301F and 3 M in case of W60F. Further quantita-tive information of the quenching curves is given in Table 2

Table 2 Ksv values of WT and Trp mutants as a function of GdnCl concentration, as calculated using Eq. (5) (“Materials and methods” section)

WT column 1 Column 2 W301F W60F W60F column3

GdnCl (M)

Ksv (M−1)

GdnCl (M)

Ksv (M−1)

GdnCl (M)

Ksv (M−1)

0 3 0 0.34 0 3.1

1 3.26 1 0.51 2 4

2 5.17 3 0.621 2.25 3.95

2.5 4.95 3.25 5.5 2.5 5.05

3 6.68 3.5 12.1 2.75 5.27

3.25 10.68 3.75 11 3 10.53

3.5 10.4 4 12 3.5 11.81

3.75 11 5 12.8 4 10.76

4 10.89 5 11.17

5 11.1

Extremophiles

1 3

Spitzfaden et al. 1997; Oliver et al. 2003; Suhanovsky and Teschke 2013). In both the cases, one of the domains must begin to fold, before influencing the other. In our case, comparison of Ksv values at 0-M GdnCl between the W60F and W301F mutant indicated that in the native PfA, the Trp of the C-terminal domain is relatively more exposed (Fig. 4). This was confirmed by the structural data (Fig. 5; Table 2). From unfolding curves, it was found that initiation of unfolding transition of the W301F protein hap-pened at a higher GdnCl concentration compared to W60F, clearly indicated that the N-domain is more stable com-pared to the C-domain (Fig. 3). In co-translational folding, it is assumed that a part of polypeptide forming an N-ter-minal domain would fold spontaneously, extending a stable framework for the rest of the chain to fold. In a way, the N-domain serves as an internal guiding factor for the rest of the polypeptide to fold (Alexandrov 1993). However, contrasting reports where C-terminal domain folds first are also available (He et al. 2005).

In our case, multiple data support that the N-domain folds and forms a stable core. The relatively high Ksv value at 3-M GdnCl for the W60F compared to the W301F as well as by unfolding/refolding experiments performed on isolated domains pointed toward this result (Fig. 3a). In addition, the thermodynamic parameters clearly placed NPfA as the more stable partner (Table 1). The unfolding pattern of W301F was more in sync with WT than W60F, suggesting that the overall folding of PfA is primarily governed by the N-terminal domain while the C-terminal domain plays negligible role (Fig. 3). Consequently, during folding, the N-domain is expected to form the native-like contacts easily. This was indeed reflected in the refolding curves of the domains where a completely refolded protein was formed without the requirement of exten-sive incubation (Supp. Fig. 2). This information along with our previous finding that the N-domain acts as a folding assistance to the C-terminal domain made us to conclude that in the case of PfA the domains fold sequentially; N-domain folds first fol-lowed by the C-domain (Tomar et al. 2013).

Fig. 5 Size-exclusion chromatography and multi-angle light scat-tering profiles. a Chromatograms for the WT PfA during unfolding at different GdnCl concentrations (indicated against the chromato-grams). Samples were incubated at 60 °C for 3 days to attain equi-librium. b SDS-PAGE of the WT PfA boiled in 2 % SDS-contain-ing Laemmli sample buffer (lane 2), incubated at room temperature in Laemmli buffer containing 1 % SDS (lane 3) and 2 % SDS (lane

4). c Comparison of SEC profile of native PfA with PfA incubated in 1 % SDS, 2 % SDS and 5-M GdnCl. Buffer used was 25 mM sodium phosphate, pH 7.4, containing 20 mM NaCl. d SEC-MALS analysis of native PfA (solid line), PfA unfolded at 5 M (dashed dotted line) GdnCl concentration. The calculated molecular weight has been shown for each peak (solid dots). Data suggest that the peak II cor-responds to monomeric species

Extremophiles

1 3

Multiple reports suggest that multimeric proteins of meso-philic origin when subjected to denaturation release mono-mers before showing any secondary structural changes (Zhu et al. 2003; Wójciak et al. 2003). This is thought to occur primarily because the non-covalent interactions within each monomer outnumber those present between the monomers (Zhanhua et al. 2005). In the case of PfA, we found a grad-ual shift in peak I in SEC toward population of larger stokes radius up to 3-M GdnCl. Together, with the equilibrium unfolding data we reasoned that this shift is due to gradual unfolding of the C-domain, without affecting the N-domain (Fig. 5a). Beyond 3 M, however, the unfolding transition of

W301F (Fig. 3a) correlated with the emergence of the second peak in SEC (Fig. 5a), suggesting global unfolding involv-ing both the domains. The match in the elution profile of the 5-M GdnCl-incubated sample with that of 1 % SDS-incu-bated sample confirmed that this globally unfolded species is a monomer. The differential unfolding of domains possi-bly results from a higher number of cooperative non-covalent interactions in the N-domain as compared to the C-domain. Thus, the peaks I and II were clearly of partially denatured dimer and fully denatured monomer, respectively. Structural analysis suggested that the network of interactions between monomers superseded those within monomers (Fig. 5).

Table 3 Interaction between amino acids involved in the interface of the two domains of WT PfA within the monomers (NPfAI and CPfAI) and between the monomers (CPfAI-CPfAII, CPfAI-NPfAII)

Bonding Monomer interface Dimer interface

NPfA I-CPfA I CPfA I-CPfA II NPfA I-CPfA II CPfA I- NPfA II

H-bonding MET 96-ASP 195 VAL 204-LYS 206′ GLU 18-ARG 277′ MET 264-VAL 155′ARG 98-THR 193 LYS 206-VAL 204′ TYR 21-ARG 272′ TYR 265- VAL 155′ARG151-ASP 289 ARG151-GLY 322 TYR 21-LYS 274′ TYR 265- THR 157′SER 153-GLU 293 VAL 155-MET 264′ ARG 272- TYR 21′GLU 164-TYR 265 ASN167-ARG 325 VAL 155-TYR 265′ LYS 274- TYR 21′ILE 171-PHE 183 THR 157-TYR 265′ ARG 277- GLU 18′

Salt bridges ARG151-ASP 289 LYS 206-GLU 230′ ASP 59-ARG 240′ ARG 240- ASP 59′

LYS 201-ASP 214′

ASP 214-LYS 201′

GLU 230-LYS 206′

Hydrophobic interactions PRO 169-PHE 183 ILE 205-LEU 203′ TYR 88-TYR 232′ TYR 239-LEU 54′VAL 181-ILE 184 ILE 208-VAL 204′ TYR 21-TYR 273′ ILE 208-TYR 88′LEU 148-PRO 185 ILE 205-ILE 205′ TYR 88-ILE 208′ TYR 232-TYR 88′TYR 68-VAL 192 VAL 204-ILE 208′ TYR 88-PRO 209′ TYR 273-TYR 21′PRO 101-VAL 192 LEU 203-ILE 215′ LEU 54-TYR 239′ PRO 209-TYR 88′LEU 97-LEU 194 LEU 222-ILE 215′ VAL 155-TYR 265′ TYR 265-VAL 155′MET 96-LEU 196 ILE 215-LEU 222′PRO 57-PRO 200 MET 264-MET 264′ILE 166-MET 264 ILE 215-LEU 203′TRP 60-MET 300 LEU 203-ILE 205′PHE 95-LEU 320 VAL 296-ILE 208′VAL 150-VAL 321 ILE 215- ILE 215′TYR 168-PHE 183 TYR 224-ILE 215′ILE 135-ILE 184 ILE 215-TYR 224′TYR 139-PRO 185 ILE 208-VAL 296′ILE 171-PRO 185

VAL 71-VAL 192

TYR 68-LEU 194

VAL 61-LEU 196

VAL 155-MET 264

ILE 166-TYR 265

TYR 88-VAL 296

MET 92-MET 300

Extremophiles

1 3

Reports of hyperthermophilic multimeric proteins unfolding through two-state mechanism are known. Except instances of domain-swapped proteins, in multisubunit proteins, the individual subunits are likely to be brought together after independent folding of each polypeptide. Here, we found that for the full-length proteins, the curves fitted into a two-state model, showing cooperative fold-ing. However, the individual domains lacked co-operativity (Fig. 3b). Thus, we conclude that non-cooperatively fold-ing domains start behaving as cooperative folding partners when presented together in a setting that demands multi-meric assembly as native structure. The overall unfolding pattern of PfA is schematically presented in Fig. 6. Thus, this study provides an understanding about multidomain multimeric protein folding by deciphering details of indi-vidual domain folding vis-a-vis their association. Insights from this study will help in designing better and stable pro-teins for therapeutic and industrial applications.

Acknowledgments DKG and RT acknowledge CSIR-INDIA and AS acknowledges ICMR Govt. of INDIA for their Research Fellow-ship. BK acknowledges the financial and infrastructural support of IIT Delhi.

References

Alexandrov N (1993) Structural argument for N-terminal initiation of protein folding. Protein Sci 2:1989–1991

Auton M, Cruz MA, Moake J (2007) Conformational stability and domain unfolding of the Von Willebrand factor a domains. J Mol Biol 366:986–1000

Bansal S, Gnaneswari D, Mishra P, Kundu B (2010) Structural sta-bility and functional analysis of l-asparaginase from Pyrococcus furiosus. Biochem Mosc 75:375–381

Bansal S, Srivastava A, Mukherjee G, Pandey R, Verma AK, Mishra P, Kundu B (2012) Hyperthermophilic asparaginase mutants with enhanced substrate affinity and antineoplastic activity: structural insights on their mechanism of action. FASEB J 26:1161–1171

Cheng X, Gonzalez ML, Lee JC (1993) Energetics of intersubunit and intrasubunit interactions of Escherichia coli adenosine cyclic 3′,5′-phosphate receptor protein. Biochemistry 32:8130–8139

Dams T, Jaenicke R (1999) Stability and folding of dihydrofolate reductase from the hyperthermophilic bacterium Thermotoga maritima. Biochemistry 38:9169–9178

Di Venere A, Nicolai E, Rosato N, Rossi A, Finazzi Agrò A, Mei G (2011) Characterization of monomeric substates of ascorbate oxidase. FEBS J 278:1585–1593

Eftink MR, Ghiron CA (1976) Exposure of tryptophanyl residues in proteins. Quantitative determination by fluorescence quenching studies. Biochemistry 15:672–680

Galvagnion C, Smith MTJ, Broom A, Vassall KA, Meglei G, Gaspar JA, Stathopulos PB, Cheyne B, Meiering EM (2009) Folding and association of Thermophilic dimeric and trim-eric DsrEFH proteins: Tm0979 and Mth1491†. Biochemistry 48:2891–2906

Harris NJ, Findlay HE, Simms J, Liu X, Booth PJ (2014) Relative domain folding and stability of a membrane transport protein. J Mol Biol 426:1812–1825

He H-W, Zhang J, Zhou HM, Yan YB (2005) Conformational change in the C-terminal domain is responsible for the initiation of cre-atine kinase thermal aggregation. Biophys J 89:2650–2658

Humphrey W, Dalke A, Schulten K (1996) VMD: visual molecular dynamics. J Mol Graph 14:33–38

Kishore D, Kundu S, Kayastha AM (2012) Thermal, chemical and pH induced denaturation of a multimeric β-galactosidase reveals multiple unfolding pathways. PLoS One 7:e50380

Kotzia GA, Labrou NE (2009) Engineering thermal stabil-ity of l-asparaginase by in vitro directed evolution. FEBS J 276:1750–1761

Krissinel E, Henrick K (2007) Inference of macromolecular assem-blies from crystalline state. J Mol Biol 372:774–797

Lindorff-Larsen K, Piana S, Dror RO, Shaw DE (2011) How fast-folding proteins fold. Science 334:517–520

Luke KA, Higgins CL, Wittung-Stafshede P (2007) Thermodynamic stability and folding of proteins from hyperthermophilic organ-isms. FEBS J 274:4023–4033

Maity H, Mossing MC, Eftink MR (2005) Equilibrium unfolding of dimeric and engineered monomeric forms of λ Cro (F58 W) repressor and the effect of added salts: evidence for the forma-tion of folded monomer induced by sodium perchlorate. Arch Biochem Biophys 434:93–107

Mills IA, Flaugh SL, Kosinski-Collins MS, King JA (2007) Folding and stability of the isolated Greek key domains of the long-lived

Fig. 6 Schematic representation of unfolding pathway of PfA with increasing GdnCl as deduced from observations by various probes. I Negligible change in the structure up to 1.5 M. II Increase in the Stoke’s radius in SEC at 2 M, reduction in activity up to 17 %; N-domain intact, C-domain begins to unfold. III Further increase in the Stoke’s radius in SEC up to 3 M, activity reduced to 30 %;

unfolding initiation of N-domain, C-domain completely unfolded. IV Dimer dissociation at 3.25 M (peak II in SEC), loss of activity, partially unfolded N-domain. (V) Expansion of monomeric radius in SEC corresponding to gradual and complete loss of structure from 4 to 5 M

Extremophiles

1 3

human lens proteins γD-crystallin and γS-crystallin. Protein Sci 16:2427–2444

Oliver DC, Huang G, Nodel E, Pleasance S, Fernandez RC (2003) A conserved region within the Bordetella pertussis autotransporter BrkA is necessary for folding of its passenger domain. Mol Microbiol 47:1367–1383

Powers S, Robinson C, Robinson A (2007) Denaturation of an extremely stable hyperthermophilic protein occurs via a dimeric intermediate. Extremophiles 11:179–189

Rousseau F, Schymkowitz JWH, Wilkinson HR, Itzhaki LS (2001) Three-dimensional domain swapping in p13suc1 occurs in the unfolded state and is controlled by conserved proline residues. Proc Natl Acad Sci 98:5596–5601

Santoro MM, Bolen DW (1988) Unfolding free energy changes deter-mined by the linear extrapolation method, part 1. Unfolding of phenylmethanesulfonyl alpha-chymotrypsin using different denaturants. Biochemistry 27:8063–8068

Santra MK, Banerjee A, Rahaman O, Panda D (2005) Unfolding pathways of human serum albumin: evidence for sequential unfolding and folding of its three domains. Int J Biol Macromol 37:200–204

Spitzfaden C, Grant RP, Mardon HJ, Campbell ID (1997) Module-module interactions in the cell binding region of fibronectin: sta-bility, flexibility and specificity. J Mol Biol 265:565–579

Suhanovsky MM, Teschke CM (2013) An intramolecular chaperone inserted in bacteriophage P22 coat protein mediates its chaper-onin-independent folding. J Biol Chem 288:33772–33783

Takano K, Aoi A, Koga Y, Kanaya S (2013) Evolvability of Ther-mophilic proteins from Archaea and Bacteria. Biochemistry 52:4774–4780

Tomar R, Garg DK, Mishra R, Thakur AK, Kundu B (2013) N-terminal domain of Pyrococcus furiosus l-asparaginase

functions as a non-specific, stable, molecular chaperone. FEBS J 280:2688–2699

Tomar R, Sharma P, Srivastava A, Bansal S, Ashish Kundu B (2014) Structural and functional insights into an archaeal l-asparagi-nase obtained through the linker-less assembly of constituent domains. Acta Crystallogr Sect D 70:3187–3197

Wójciak P, Mazurkiewicz A, Bakalova A, Kuciel R (2003) Equilib-rium unfolding of dimeric human prostatic acid phosphatase involves an inactive monomeric intermediate. Int J Biol Macro-mol 32:43–54

Wriston JC Jr (1985) [79] Asparaginase. In: Alton M (ed) Methods in enzymology. Academic Press, San Diego, pp 608–618

Zhang Z, Chan HS (2010) Competition between native topol-ogy and nonnative interactions in simple and complex folding kinetics of natural and designed proteins. Proc Natl Acad Sci 107:2920–2925

Zhanhua C, Gan JGK, lei L, Sakharkar MK, Kangueane P (2005) Pro-tein subunit interfaces: heterodimers versus homodimers. Bioin-formation 1:28–39

Zhou P, Tian F, Shang Z (2009) 2D depiction of nonbonding interac-tions for protein complexes. J Comput Chem 30:940–951

Zhu L, Zhang XJ, Wang LY, Zhou JM, Perrett S (2003) Relationship between stability of folding intermediates and amyloid formation for the yeast prion Ure2p: a quantitative analysis of the effects of pH and buffer system. J Mol Biol 328:235–254

Zitzewitz JA, Bilsel O, Luo J, Jones BE, Matthews CR (1995) Probing the folding mechanism of a leucine zipper peptide by stopped-flow circular dichroism spectroscopy. Biochemistry 34:12812–12819