Differentially labeled flaviviral protease-cofactor complex for NMR spectroscopic applications

Abstract

Flaviviruses such as Dengue, Zika and West-Nile viruses have a positive strand RNA genome which is translated to a polyprotein inside the host cell. The viral polypeptide is matured to its constituents by the enzymatic action of NS2B-NS3 protease-cofactor complex. The flaviviral protease-cofactor complex attracted a lot of interest recently because of its potential for therapeutic intervention and the unique nature of catalysis where the peptide cofactor regulates the enzymatic activity. Obtaining the enzyme and cofactor differentially labeled with naturally abundant nuclei and NMR active nuclei respectively will be helpful in reducing the spectral complexity by making the enzyme invisible in a multidimensional NMR spectrum while only showing peaks from the cofactor. This will enable one to study the properties of the cofactor in isolation using NMR spectroscopy. Here, I have used a strategy for selectively labeling the cofactor within the complex with NMR active nuclei while peaks from the enzyme were rendered invisible. The protocol used here takes advantage of an ‘on-column unfolding’ step during the initial Ni-NTA chromatography to separate the enzyme and cofactor in unfolded conditions. The labeled cofactor was then allowed to fold in the presence of an unlabeled enzyme to obtain a differently labeled complex. We compared the ¹H-¹⁵N HSQC spectrum of the differently labeled, wild type and free cofactor to ensure that the cofactor attained the desired fold within the complex. The protocol is scalable, inexpensive and can be applied to other two-component enzyme systems where a peptide cofactor is essential for the folding of an enzyme.

1. Introduction

Flaviviruses are positive strand RNA viruses and pose a significant challenge to global healthcare by recently reported outbreaks especially in third world countries. Protease domain of flaviviral non-structural protein (NS3) and its NS2B cofactor (NS2B-NS3) complex is an attractive target for therapeutic intervention as the protease is responsible for the maturation of viral polyprotein at the early stages of infection [1]. Like many other viral cofactor protein systems, the flaviviral cofactor is essential for the both the structure and the activity of the enzyme. In the case of flavivirus like zika, dengue and West-Nile virus, the cofactor winds around the enzyme and the C-terminal domain of the cofactor attain close contact to the active site. N terminal half of the cofactor comprises residues that shares a beta sheet with the enzyme making it essential to maintain the solubility while the C terminal half is involved in the catalytic activity [2]. Several efforts were previously carried out using Nuclear Magnetic Resonance spectroscopy to understand the structure as well as the mechanism of the enzymatic activity, including solution NMR spectroscopy where 3D structure of the complex and with a peptide ligand was determined [3,4].

NMR spectroscopy is a valuable tool for studying the structure and function of the biomolecules in solution since it can obtain atomic-level structural and dynamic information with relative ease. Modern approaches in multidimensional NMR spectroscopy use heterologous expression of the protein in bacteria grown in media containing NMR active nuclei such as ¹³C and ¹⁵N [5]. These nuclei are incorporated into protein by de-novo amino acid synthesis and consequent ribosomal protein expression. The 2D plane formed by chemical shifts of the ¹H and ¹⁵N of amides in a protein constitutes the base spectrum of most multidimensional NMR experiments [6]. However, larger proteins can suffer from severe spectral complexity due to a greater number of amide groups or severe peak overlap in the center of the spectrum in the proton dimension (8–8.5 ppm) due to a higher degree of disorder. This complexity can be reduced by selective labeling of nuclei to make only residues of interest visible in the spectrum. There are several approaches for selectively labeling and unlabeling proteins for NMR spectroscopic studies; methyl labeling of amino acid side chains using precursors that intervene citric acid cycle is a recent addition for solution NMR spectroscopic studies of large proteins [7]. Selective labeling and unlabeling of amino acids are routinely used in solid-state NMR spectroscopy of large proteins as well [8,9]. In addition to reducing the spectral complexity, some of these labeling techniques also involve replacing the water in the growth medium with heavy water to reduce the number of NMR active ¹H nuclei thereby reducing the proton-proton dipolar relaxation, which is a major source of signal loss in multidimensional NMR spectroscopy [10].

Several strategies have been reported for selectively labeling a domain of interest in a large protein or a multi-protein complex. Sortase and intein-mediated enzymatic ligation are used for labeled domains with unlabeled ones in large proteins [11,12]. Other nature-inspired ligation strategies are reviewed elsewhere [13]. For large multi-protein complexes, ‘LEGO-NMR’ is used for sequentially expressing labeled and unlabeled counterparts aided by different promoters [13,14]. NS2B-NS3 protease-cofactor complex consists of 230 amino acids: 177 amino acids from the enzyme and 53 amino acids from the cofactor. While 3D structural characterization and several binding studies were done in the protein using NMR spectroscopy, there ¹H-¹⁵N HSQC spectrum suffers from considerable overlap due to disordered regions in the C-terminal half of the cofactor as well as the loops present in the enzyme. For example, there are around 164 peaks in the center in the 8–9 ppm range in the proton dimension akin to a highly disordered protein [15] and probably due to the dynamic nature of the protein. Reducing the number of peaks in the NS2B-NS3 complex by selectively unlabeling the peaks from the large enzyme will help study the properties of the cofactor in isolation. Selective labeling the cofactor with NMR active nuclei while rendering the rest of the enzyme ‘invisible’ can be advantageous in studying the cofactor in NMR experiments as well as in biochemical investigations such as binding to the inhibitors or substrate molecules. Previous efforts on this front have been made by Yang et al., where they used a co-expression strategy of both cofactor and enzyme in two different plasmids and their subsequent expression in differently labeled media [16]. Here we propose an alternative strategy with an ‘on-column unfolding’ step to separate the protein and cofactor and make a complex in vitro using dialysis. This strategy avoids additional cloning of cofactor and enzyme into two separate vectors for expression and eliminates the risk of leaky expression of a second protein in minimal media. Additionally, the risk of contamination of the labeled media due to incomplete washing during the co-expression step can also be minimized in the presented approach.

2. Principle

The enzyme-cofactor complex is non covalently formed and stabilized by secondary structural elements between them to make a functional complex. During the Ni-NTA chromatography, an additional wash step in presence of an unfolding agent (6M Guanidine HCl), separates the cofactor from the complex while the enzyme remains attached to the Ni-NTA column by virtue of the N terminal 6xHis tag. This ‘on column’ unfolding of the complex and subsequent elution of cofactor and the enzyme by unfolding buffer and unfolding buffer-imidazole mixture respectively. The labeled cofactor and unlabeled enzyme are later mixed in guanidine HCl and dialyzed against refolding buffer to obtain selectively labeled cofactor.

3. Materials and methods

3.1. Protein expression and purification

a. Plasmid used for the study

pET-Duet vector expressing the unlinked complex [17] was used for the study expressing NS2B-NS3 protien. Genes encoding Zika virus NS3 protease (1–177) and NS2B cofactor (45–92) were cloned into a pET-Duet vector under separate T7 promotors and terminators. The NS3 polypeptide had a 6xHis tag attached to the N terminus followed by a thrombin cleavage site. Vector map is given in SF-1. For expressing the cofactor alone, the gene encoding the cofactor NS2B was cloned into a pET29b + vector with an N-terminal hexahistidine tag followed by a short linker (DYDIPTT) and a tobacco etch virus (TEV) protease cleavage site (ENLYFQG).

b. Selective labeling strategy of the NS2B-NS3 complex

In order to achieve selective labeling of NS2B with 15N labeling (or any other selective labeling), the protein was expressed in 1L 2xM9 minimal media with 1 g/L¹⁵NH₄Cl as nitrogen source [18]. BL21(DE3) cells transformed with pET-Duet plasmid (plasmid details in 1.a) expressing both NS2B and NS3 were grown in the prescribed media, the expression was induced by 1 mM IPTG and the cells were pelleted down using centrifugation at 9300g for 30 min. The cells were resuspended in lysis buffer and lysed by sonication (20 mM Tris HCl pH 7.4, 500 mM NaCl, 30 mM Imidazole and 1.4 mM β-mercaptoethanol). Cell lysate was cleared by centrifugation at 19700g for an hour. For selectively labeling the protein, the cell lysate was passed through a His-trap FF column (Ni-NTA chromatography) pre-equilibrated with lysis buffer. The wash step was continued until the A₂₈₀ reached ~0.01 and then wash buffer containing 6M Guanidine HCl was applied onto the column at 0.5 ml/min. At this stage of purification, the non-covalently bound cofactor is expected to dissociate from the complex due to the presence of unfolding agent (6M Guanidine-HCl). The flowrate was increased after sufficient contact time with the guanidine-HCl was attained (30 min) and the flowthrough containing the labeled cofactor was collected until the absorbance reached near zero. Labeled NS3 enzyme was then collected using the elution buffer containing 6M Guanidine HCl. Now that the enzyme and cofactor was separated in Guanidine HCl containing buffer, same was done for unlabeled enzyme cofactor complex to separate the enzyme and cofactor. The labeled cofactor was then mixed with unlabeled enzyme in 1.5:1 ratio and the guanidine-HCl was removed by dialyzing against a 2L gel filtration buffer (20 mM Tris HCl pH 7.4, 200 mM NaCl, 1 mM EDTA, 1.4 mM β-mercaptoethanol). To ensure that the protein-cofactor complex is formed after the dialysis, the Ni-NTA was repeated and the complex was eluted as before. The scheme of purification is given in Fig. 1. The Hisx6 tag was removed by incubating the eluted protein with thrombin at 10 units/ml of eluted protein and the cleaved protein was purified by size exclusion chromatography in Superdex 75 size exclusion column.

Fig. 1.

Scheme used for purifying the selectively labeled complex. NS2B-NS3 labeled with NMR active compounds is shown in color where the NS3 protease is represented in green and NS2B cofactor in red. Unlabeled complex is represented in grey color. Structure of unlinked zika viral NS2B-NS3 complex used here is 5GJ4. The second Ni-NTA was followed by a size exclusion chromatography (not represented in figure) before the data acquisition. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

c. Purification of uniformly labeled or unlabeled complex

Expression of protein in BL21(DE3) cells as well as the lysis was done as described in methods section 1.b. Clear lysate after centrifugation was passed through His-Trap FF column and the wash was continued until the absorbance at 280 nm reached ~0.01. The protein was eluted in same buffer with 300 mM imidazole. Eluted protein was then added with thrombin to remove the 6xHis-tag and proceeded with gel filtration as mentioned previously. LB medium was used to obtain the unlabeled protein where 2xM9 media containing ¹⁵N labeled NH₄Cl at 1 g/L was used for obtaining uniformly ¹⁵N labeled protein.

d. Purification of free cofactor

Steps involved in the expression and initial purification of the free cofactor were same as the complex. The N-terminal His-tag was removed using TEV-protease at 1 mg protease for 50 mg of eluted protein.

3.2. Enzyme assay

Bz-nle-KRR-amino methyl coumarin (Cayman Chemical, USA, Cat. #CAY27710) was used as a substrate to measure the enzymatic activity of unlabeled and selectively labeled complexes. The assay was conducted by incubating 100 nM enzyme with 20 μM substrate in protease buffer (50 mM Tris, pH 8.2, 10 % glycerol, 10 % dimethyl sulfoxide, and 0.01 % Triton X-100) at 25 °C for up to 1 h. Protease activity was quantified by measuring the fluorescence emission of free amino methyl coumarin at 440 nm, using an excitation wavelength of 340 nm, at 5-min intervals. Reactions were performed in triplicate, and standard deviations were calculated to assess variability. Control reactions were conducted by substituting the enzyme with protease buffer.

3.3. NMR spectroscopy

¹⁵N-¹H heteronuclear single quantum coherence (HSQC) NMR spectrum was used for analyzing the extent of labeling the complex. The experiment records the correlation between the amide proton and nitrogen (¹⁵N) in the peptide backbone to make a 2D spectrum with proton and nitrogen chemical shifts in the X and Y axis respectively. NMR samples were prepared from the pooled and concentrated fractions eluted from gel filtration chromatography. Final buffer contained 10 % deuterium oxide as a lock solvent and the NMR sample was loaded onto a 5 mm NMR tube for acquiring the spectrum. All NMR experiments were carried out at 298 K on a Bruker 700 MHz spectrometer equipped with a room temperature triple resonance single-axis gradient TXI probe and Avance-Neo console. The concentration of the protein varied from 500 μM to 900 μM in various preparations. For the temperature titration, HSQC experiment was repeated with similar experimental parameters over a range of temperatures (15–35 °C).

4. Results

4.1. Protein expression and purification

On-column unfolding strategy described here for obtaining differently labeled protein-cofactor complex in zika virus NS2B-NS3 involved chromatographic processes such as Ni-NTA chromatography and gel filtration chromatography. From the Ni-NTA column, more than 20 mg protein was obtained for the unlabeled and uniformly labeled protein. 20 mg NS2B cofactor was obtained from 1L of culture when expressed in pET29b + vector.

4.2. Selectively labeling the protease-cofactor complex

The selectivity labeled complex was made by following the protocol described in methods section. Elution of proteins during a second round of Ni-NTA chromatography after dialysis helps in the separation of selectively labeled complex as unbound constituents enzyme comes out in wash instead of elution. Final yield of the complex with selective labeling is presented in supplementary data (ST1). The yield was reduced by half during the process of making the selectively labeled complex. SDS-PAGE of the complex is shown in SI Fig. 2 and gel filtration profile of the fractions of the complex in comparison with the uniformly labeled complex is given in Fig. 2. A.

Fig. 2.

A) Size exclusion chromatography of unlabeled and selectively labeled NS2B-NS3 complex. Size exclusion chromatography showing the structural homogeneity of the selectively labeled complexes on a Superdex75 analytical gel filtration column as indicated in the legend. The curve of elution volume against the log of molecular weight markers used is plotted in doted dashed line. Molecular weight of each marker is indicated next to the data points in kilodaltons. Void volume of the column is 7.92 mL. B) Enzymatic activity of the unlabeled and selectively labeled complexes. Activity of the three complexes in 2.A on bz-nle-KRR-amino methyl coumarin as a measure of fluorescence of free AMC emission at 440 nm in 1 h. Control reaction contains the substrate without the enzyme in protease buffer. The data is fitted with an exponential function.

4.3. Enzyme assay

The time-dependent enzymatic cleavage assay demonstrated that all three enzyme variants cleaved the substrate, albeit at different rates. The uniformly labeled enzyme reached saturation within 30 min, whereas the complex with the ¹⁵N-labeled NS2B cofactor required 45 min for complete substrate cleavage. In contrast, the complex with the ¹⁵N-labeled NS3 enzyme exhibited significantly reduced activity, achieving only approximately 70 % substrate cleavage compared to the other two in 1 h of reaction (Fig. 2B).

4.4. NMR spectroscopy

¹H-¹⁵N protein HSQC spectrum was used to analyze the peaks of the protein where all proton-nitrogen correlations of the amide groups of backbone as well as the side chains were plotted in a 2D contour plot. The HSQC spectrum of the uniformly labeled protein is given in Fig. 3A (black spectrum) where all peaks from the protease and the cofactor amide groups are visible. While the peaks from the free cofactor remained in the center of the spectrum indicating the secondary structural features are absent in the free cofactor compared to the selectively labeled complex. A spectrum in which the cofactor peaks are invisible and only enzyme peaks are visible is provided in the supplementary material (SI F-3). HSQC spectrum of the selectively labeled complex under various temperature are plotted in Fig. 4. Cofactor peaks did not dissociate from the complex in all temperatures tested as seen in the overlay indicating the stability of the complex.

Fig. 3. ¹H-¹⁵N HSQC NMR spectrum of the selectively labeled complex.

A) Overlay of spectra between the uniformly labeled complex (black) with selectively labeled cofactor (red) in the left. 2D NMR spectrum of the selectively labeled cofactor in complex (red) with uniformly labeled free NS2B expressed and purified in the absence of the enzyme (blue) is given in the right. Green data represent negative peaks, either from the folded peaks of arginine side chains or from the backbone amides B). ¹H(left) and ¹⁵N(right) correlation of chemical shifts of various residues in the uniformly and selectively labeled protein. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4. Temperature titration of selectively labeled cofactor:

A). 15N-1H HSQC spectra recorded at various temperature. The temperature is indicated by color of the legend towards the right. B). Residues indicated as cofactor peaks in A is highlighted in spheres. The enzyme and cofactor are colored pale green and orange respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

5. Discussion

Purification of a selectively labeled cofactor in flaviviral enzyme-cofactor complex is presented here. The selective labeling is achieved by introducing an on-column unfolding step during the washing of the Ni-NTA chromatography. Significant loss of protein was observed during the dialysis for the refolding of the complex due to precipitation of the NS3 during dialysis (supplementary information ST1). Loss of protein during the refolding is the major drawback of the approach, however the yields were sufficient to get final sample for NMR spectroscopic experiments, which usually demands hundreds of micromolar concentrations. Care should be taken to keep the total NS3 concentration in the mixture below 1 mg/ml to avoid rapid precipitation. It is promising to see that the complex formed shows marked structural difference from its constituent counterparts as free cofactor elutes as a disordered protein and the free NS3 enzyme appears as a soluble aggregate in the size exclusion profile. The size exclusion chromatography can therefore play as a pivotal step in ensuring the structural homogeneity of the selectively labeled complex.

¹H-¹⁵N protein HSQC spectrum was used to confirm the conservation of structure in the selectively labeled complex. The overlay in Fig. 3A indicates that the cofactor attained a fold similar to a functional complex. Most of the peaks were distributed around the center of the amide region since the cofactor remains without much secondary structure even in the complex. However, a few peaks away from the center match well with the cofactor from the uniformly labeled spectrum indicating the fold of the cofactor is preserved (Fig. 3A). A comparison of the cofactor in the complex and in the free form is shown in Fig. 3b, where the cofactor peaks show less dispersion 1H in chemical shifts comparable to an intrinsically disordered protein. The degree of disorder of the cofactor within the complex is indicating by several overlapping peaks. Around 61 peaks are present in the cofactor which indicates all peaks from the cofactor can be seen albeit their low signal intensity. It has to be noted only 54 peak assignments are available in the BMRB database indicating that missing peaks from the complex can be recovered by selectively labeling the protein [3]. However, the number of peaks may be exaggerated due to the dynamic nature of the cofactor [16,19].

After confirming the proper folding of the enzyme, functional characterization of the labeled enzyme was conducted. An assay using the bz-nle-KRR-AMC substrate revealed that the wild-type enzyme exhibited the highest rate of activity, while the 15N-labeled enzyme showed the lowest activity. It remains unclear whether this reduced activity is due to partial misfolding of the protein or a lack of stability compared to the native enzyme. These results indicates that the method presented here can be used to produce stable functional complex of unlinked zika virus protease. In order to showcase the stability of the selectively labeled complex formed by the method described here, a series of HSQC spectra of the complex was acquired with selectively labeled cofactor over a range of temperature from 5 to 35 °C (Fig. 4A). It is noteworthy that the thermal denaturation studies done previously showed that the complex behaves like an unfolded protein upon thermal denaturation [20]. Appearance of new peaks at higher temperature may indicate the presence of a minor conformation. Moreover, the residue I30 showed dramatic changes to the change in temperature another peak while A42 lost intensity significantly at higher temperature. These two residues are located at the folded region of the cofactor (Fig. 4B) [3]. Several viral nonstructural proteins were reported to have peptide cofactors similar to the NS2B-NS3 complex and the labeling strategy described here can be extended further into structural or functional characterization of these proteins. Several studies involving the binding of small molecules for allosterically targeting the enzymatic activity is has been reported in recent past and selective labeling can help in identifying the effect of these molecules in the cofactor in isolation [21]. Even though the application discussed here is apparently limited to NMR spectroscopic studies, such selectively labeled complexes can be useful for other bio-physical applications where site specific labeling of a residue/residues either from cofactor or the enzyme can help in biophysical or functional characterization of the enzymes.

In conclusion, the labeling strategy described here can be used for the selective characterization of flaviviral two-component protease complexes with NMR spectroscopy. The techniques described here do not involve additional media components or precursors for selective labeling and can be scaled up by increasing the volume of the bacterial growth culture. No additional chromatographic preparations are necessary since only Ni-NTA columns and size exclusion matrices for the purification process. This is the first time the ‘on-column’ unfolding is carried out to separate a heterodimer followed by reconstitution to achieve the natively folded complex with differential labeling. Though the method is now applied only on flaviviral protease complex, the same can be extended for multiprotein complexes suffering from severe signal overlap or spectral complexity in multidimensional NMR spectroscopy [22]. Moreover, site-specific labeling of the cofactor amino acids can be a valuable tool in fluorescence and other biophysical studies.

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pep.2025.106684.

Data availability

Data will be made available on request.