Team:Slovenia/Coiled-coil polyhedra Results.html

Summary

Polypeptide production and isolation

Plasmid, coding for K2 was assembled from BioBricks (BBa_K245118, BBa_K245119, BBa_K245127) and transferred into a BL21(DE3) pLysS strain of E. coli. Bacteria were grown according to the standard procedures, protein expression was induced with IPTG and bacteria were harvested after 4 hours. Following cell lysis supernatant (Figure 1A, lane 1) and insoluble fraction (inclusion bodies; Figure 1A, lane 2) were analyzed for the production of K2 polypeptide. The protein was mainly present in form of inclusion bodies (Figure 1A, lane 2), which were further purified (Figure 1A, lane 3). Purified protein (Figure 1A, lane 4) was prepared by solubilizing inclusion bodies in 6 M GdnHCl and purification on nickel-NTA column under the denaturing conditions. Western blot revealed the presence of oligomers (Figure 1B).

Figure 1: Production and analysis of K2 polypeptide. A) SDS PAGE of the soluble fraction of bacterial cell lysate (lane 1), insoluble fraction (lane 2), washed inclusion bodies (lane 3), protein purified by chelating chromatography (lane 4). B) Analysis of K2 after slow chemical annealing by Western blot (line 2). Additional bands showing oligomers of K2 are visible. Line 1 represents standard proteins.

We wanted to experimentally test the design of P1 and P2 coiled-coil-forming segments with respect to their lack of secondary structure in the isolated form and formation of a coiled-coil heterodimer. CD spectra of individual P1 or P2 peptides show that each of them is disordered in solution while their mixture (P1+P2) shows a high level of α-helical content (Figure 2). Although we did not strictly prove that P1 and P2 peptides form a parallel and not antiparallel coiled-coil dimer our results demonstrate that we designed an adequate pair of coiled-coil forming segments suitable for the formation of more complex assemblies.

Figure 2: Far-UV CD spectra of peptides P1, P2 and mixture of P1 + P2 at 25° C. The shape of spectra indicated that individual peptides were not structured, while mixture of P1+P2 showed a high level of -helical folding. The peptide concentration was 0.1 mg/ml in 10 mM HEPES, pH 7.5.

We analyzed the secondary structure of a polypeptide K2 under the native conditions (Figure 3). CD spectra showed strong helical signal, confirming that the coiled-coil interactions occur also in the context of a longer polypeptide and in the presence of linker sequences.

Figure 3: CD spectrum of K2 reveals a high fraction of α-helical secondary structure. K2 polypeptide forms a precipitate under the native conditions due to the formation of multiple coiled-coil interactions.

Desired self-assembly includes interaction between several polypeptide chains. We presumed that the desired assemblies represent the energetic minimum, since all coiled-coil-forming potentials are satisfied. However with increasing number of polypeptide chains in the assembly the energetic differences between the correctly folded and misfolded structures decrease and kinetic limitations for formation of the most ordered structure may become important. In order to increase the fraction of correctly folded assemblies we wanted to remove the kinetic considerations by performing very slow annealing. In case of DNA origami this is usually done by a slow decrease of temperature over several hours or days. In our case temperature annealing did not give good results, presumably because the temperature-denatured state may not be completely unfolded and since both hydrophobic and electrostatic interactions take part in stabilization of coiled-coil assemblies, each with their different temperature dependence. Therefore we decided to use chemical annealing by decreasing the concentration of denaturing agent in solution towards the native conditions. For the efficient annealing we should keep the solution sufficiently long under the conditions, with equilibrium between the unfolded and folded structures. We determined the midpoint of transition of K2 from the CD analysis at different concentrations of the denaturing agent, which was around 4 M GdnHCl (Figure 4).

Figure 4: Secondary structure of the polypeptide at different concentrations of GdnHCl was determined by measuring the circular dichroism at 222nm on a CD spectrometer.

Usually slow refolding is performed by dialysis, which however may not be optimal if the difference in the concentration of both solutions is large, which may lead to the large concentration gradients at the edge of the dialysis tube. Therefore we invented a technique to perform chemical annealing over any desired length of time, by performing the dialysis against the continuously changing concentration of denaturing agent (Figure 5). From the analysis of the folding transition of K2 we determined that the optimal range should be between 5 M GdnHCl, where the large majority of the secondary structure is disrupted to 1 M GdnHCl, where almost all secondary structure is formed.

Figure 5: Scheme of the setup for slow chemical annealing of K2.

Buffer was slowly pumped into the beaker with mixing denaturant solution and protein samples in the dialysis bag, decreasing over 20 hours the conditions from the unfolding to the native conditions.

Continuous decrease of the denaturant in the dialysis solution was achieved by slowly pumping the buffer into the stirred dialysis solution, which initially contained 5 M GdnHCl, which was the same concentration as in the dialysis bag with K2 polypeptide. In this way the concentration difference between the solution inside and outside dialysis bag was minimal and programmable pump allowed setting the slope of the gradient to any desirable value, extending refolding from 12 hours towards potentially several days.

Concentration of the polypeptide represents an important factor in determining the type of self-assembled structures, either polygons, which require the assembly of eight chains or lattice, which requires over hundreds of molecules. At low polypeptide concentrations formation of polygons should be favored since they should form oligomers that are big enough to form a closed structure, satisfying all coiled-coil-forming potentials. Refolding procedure allowed us to perform the self-assembly in parallel using several different concentrations at the same time.

Experimental self-assembly of K2

K2 dissolved in 6 M GdnHCl at 10 mg/ml was diluted to 0.5 µg/ml, 5 µml and 50 µml in 5 M GdnHCl in 20 mM HEPES buffer pH 7. 100 ml of the solution was placed in a dialysis bag with membrane cutoff of 3.5 kDa, and placed in 2 l beaker containing a 200 ml of the 5 M GdnHCl in 20 mM HEPES buffer pH 7 (Figure 5A). Solution was stirred at room temperature with magnetic mixer at 1000 rpm. To the bottom of the dialysis solution in the vicinity of rotating magnet a solution with 20 mM HEPES pH 7 was added by a pump set to the flow rate of 1 ml/min. Within 20 hours the concentration of denaturing agent in the solution decreased to 1 M (Figure 5B). At this point, when the self-assembly was already completed the dialysis bag with protein solution was transferred to the solution containing 2 l or 20 mM HEPES pH 7 and dialyzed for 3 hours and the procedure was repeated twice, to remove the denaturant (Figure 5C).

Analysis of the self-assembled K2 structures by dynamic light scattering (DLS)

Sensitivity of DLS is rather low as it requires sample at concentrations above 0.2 mg/ml, therefore we had to concentrate the annealed polypeptide samples. DLS results (Figure 6) showed that solution contained two populations of K2 aggregates and both were present at appreciable amount. Hydrodynamic radius (R_H) of one population was 8 nm, and for the other 88 nm, indicating the presence of initial steps of larger assemblies in addition to small aggregates, compatible with the expected box.