Biophysical characterization of human-cell-expressed, full-length κI O18/O8, AL-09, λ6a, and Wil immunoglobulin light chains

Abstract

Immunoglobulin light chain (AL) amyloidosis involves the deposition of insoluble monoclonal AL protein fibrils in the extracellular space of different organs leading to dysfunction and death. Development of methods to efficiently express and purify AL proteins with acceptable standards of homogeneity and structural integrity has become critical to understand the in vitro and in vivo aspects of AL protein aggregation, and thus the disease progression. In this study, we report the biophysical characterization of His-tagged and untagged versions of AL full-length (FL) κI and λ6 subgroup proteins and their mutants expressed from the Expi293F human cell line. We used an array of biophysical and biochemical methods to analyze the structure and stability of the monomers, oligomerization states, and thermodynamic characteristics of the purified FL proteins and how they compare with the bacterially expressed FL proteins. Our results demonstrate that the tagged and untagged versions of FL proteins have comparable stability to proteins expressed in bacterial cells but exhibit multiple unfolding transitions and reversibility. Non-reducing SDS-PAGE and analytical ultracentrifugation analysis showed presence of monomers and dimers, with an insignificant amount of higher-order oligomers, in the purified fraction of all proteins. Overall, the FL proteins were expressed with sufficient yields for biophysical studies and can replace bacterial expression systems.

Keywords: AL amyloidosis; AL full-length proteins; Circular dichroism; Fluorescence; Mammalian cell protein expression.

Figure 1.

Mass spectrometry and amino acid sequences of the proteins studied. Panel A) shows the crystal structure of the immunoglobulin κI O18/O8 light chain dimer (PDB: 1B6D). Chains are colored in red and blue. Cysteine residues involved in disulfide bonds are shown as green spheres. The dimer contains four tryptophan residues which are shown as light red spheres. Structure was drawn using UCSF Chimera [42]. To ensure protein identity and correct disulfide linkage, the full-length light chain proteins were digested with trypsin and analyzed via LCMS. Panels B) and C) show a typical mass spectrum with assigned major fragments for H- κI O18/O8 FL and the corresponding fragment table with identified disulfide bonds. Tryptic fragments shown in green indicate that the peptide was identified successfully via MS/MS. Inset of B) shows UV-Vis spectra of an Ellman test [31] performed for H- κI O18/O8 FL and H- κI O18/O8 FL His (10 μM protein with 50 μM DTNB) that demonstrate the absence of free cysteine-residues in the protein solutions. Panel D) provides the amino acid sequences for all proteins studied. Residues without background belong to the variable domain, while amino acids highlighted in grey are part of the constant domain. Residues with green background indicate mutations in the protein with respect to the germline protein. Amino acids belonging to the expression constructs are highlighted in cyan. Peptides that were not found via MS/MS analysis are shown underlined. Fragment tables for all proteins can be found in the supplement.

Figure 2.

SDS-PAGE & western blot under non-reducing and reducing conditions. After purification, all proteins were analyzed via SDS-PAGE (panel A) & western blot (panel B). Non-reducing conditions (absence of β-mercaptoethanol) were used to assess the disulfide linkages in the protein preparations, while the subsequent western blot was used to assess oligomeric species present in the protein preparations. Horizontal green lines indicate the approximate positions of monomers, dimers, and trimers. For the sake of presentation, the gels were merged - the original images are provided in the supplement (Figure S2).

Figure 3.

Analytical ultracentrifugation, DLS, and SEC. Panel A) shows the sedimentation velocity results for all FL proteins. Plots on the left show fitted scans and the corresponding residuals for H- λ6a FL and H- λ6a FL His, while the resulting sedimentation coefficient distributions and there corresponding molecular weight (kDa) are shown on the right side. Plots with the AUC scans of the other proteins can be found in the supplement (Figure S3). A table with detailed sedimentation coefficient (s) with corresponding molecular weights (kDa) of different species is shown in table S3. Panel B) shows the intensity weighed particle size distribution for all proteins, obtained via DLS. Estimated hydrodynamic radii and major populations are indicated in the graphs. The corresponding autocorrelation functions for each protein are shown in the supplement (Figure S5). Panel C) shows SEC chromatograms for the FL proteins. Retention times of marker proteins (Biorad SEC Standard consisting of thyroglobulin (670 kDa), γ-globulin (158 kDa), ovalbumin (44 kDa), myoglobulin (17 kDa) and vitamin B12 (1.35 kDa)) are indicated in the top panel for reference (see also Figure S6).

Figure 4.

Far-UV CD and fluorescence emission spectra as a function of temperature for all studied FL proteins. Panel A) shows far-UV CD spectra of all FL proteins taken at temperatures ranging between 10 and 80 °C (color gradient) and re-equilibrated at 10 °C after refolding (○). All spectra were converted to mean ellipticity per amino acid residue (Θ^MRW). Panel B) shows intrinsic protein fluorescence emission spectra at the same temperatures and after re-equilibration at 10 °C (○). The excitation wavelength was 280 nm. Bar chart in B) shows the wavelength of the maximum intensity for each protein at 10 °C before and after unfolding.

Figure 5.

Thermal denaturation and renaturation of the FL proteins followed by far-UV CD. CD signal and the corresponding dynode voltage (panel insets) were recorded at 216 nm as a function of temperature. The scan rate was 1.0 °C/min, and the protein concentration was 10 μM. Fits are shown as red and blue lines. Fit residuals from a 2-state fit are shown at the bottom of each panel. Obtained transition midpoints (T_M) are plotted in Figure 7 and reported with the corresponding enthalpies in the supplement (Table S2a).

Figure 6.

Thermal denaturation followed by fluorescence. For all FL proteins, fluorescence emission spectra were recorded between 278 and 440 nm as a function of temperature, using an excitation wavelength of 280 nm. The scan rate was estimated to be 0.88 °C/min; the protein concentration was 2 μM. All spectra were corrected for the signal of the corresponding buffer. Each panel shows the light scattering on top, obtained from plotting the emission at 280 nm. Main plots show the thermal unfolding followed at 360 nm (●) and the corresponding spectral emission maximum (λ_max, ●). After thermal unfolding, each sample was re-equilibrated at 10 °C and another fluorescence spectrum was recorded. Values for light scattering, F_360nm and λ_max, from this spectrum are indicated by (□). Reversibility of unfolding (●) was tested further for each protein by performing a second unfolding experiment on previously unfolded protein after 3 days at 4 °C. The scan rate for this subsequent unfolding experiment was 2.0 °C/min. Due to the more or less apparent three-state denaturation, all unfolding traces were fitted with a 3-state model in which the intermediate state baseline was assumed to be the average of the native and denatured state baselines. Fit residuals for F_360nm, λ_max, and the reversibility test are shown at the bottom of each plot. Transition midpoints obtained from each fit are plotted in Figure 7 and are reported in the supplement (Table S2b).

Figure 7.

Transition midpoints from CD and fluorescence experiments. Panel A) shows the T_M-values from the un- and refolding of all FL proteins: Top left – CD (unfolding & refolding; 2-state fits), Top right– Fluorescence F_360nm, bottom left – Fluorescence λ_max and bottom right - Fluorescence F_360nm three days after the first thermal scan. All fluorescence thermal scans were fitted with a 3-state model resulting in T_M1 and T_M2. Panel B) shows correlations between the obtained T_M values. For reference, a unity line (slope = 1) is shown. Top panels show the correlation between the CD T_M-values and those from both F_360nm thermal unfolding experiments (left) and λ_max (right). Bottom panels correlate T_M-values obtained by fluorescence. Left panel shows Initial F_360nm unfolding vs reversibility. Right panel shows F_360nm unfolding vs λ_max.