Design Rules for the Sequestration of Viruses into Polypeptide Complex Coacervates

Abstract

Encapsulation is a strategy that has been used to facilitate the delivery and increase the stability of proteins and viruses. Here, we investigate the encapsulation of viruses via complex coacervation, which is a liquid-liquid phase separation resulting from the complexation of oppositely charged polymers. In particular, we utilized polypeptide-based coacervates and explored the effects of peptide chemistry, chain length, charge patterning, and hydrophobicity to better understand the effects of the coacervating polypeptides on virus incorporation. Our study utilized two nonenveloped viruses, porcine parvovirus (PPV) and human rhinovirus (HRV). PPV has a higher charge density than HRV, and they both appear to be relatively hydrophobic. These viruses were compared to characterize how the charge, hydrophobicity, and patterning of chemistry on the surface of the virus capsid affects encapsulation. Consistent with the electrostatic nature of complex coacervation, our results suggest that electrostatic effects associated with the net charge of both the virus and polypeptide dominated the potential for incorporating the virus into a coacervate, with clustering of charges also playing a significant role. Additionally, the hydrophobicity of a virus appears to determine the degree to which increasing the hydrophobicity of the coacervating peptides can enhance virus uptake. Nonintuitive trends in uptake were observed with regard to both charge patterning and polypeptide chain length, with these parameters having a significant effect on the range of coacervate compositions over which virus incorporation was observed. These results provide insights into biophysical mechanisms, where sequence effects can control the uptake of proteins or viruses into biological condensates and provide insights for use in formulation strategies.

Figure 1

Coacervate design using sequence-defined block-co-polypeptides. (a) Schematic depiction of virus-containing coacervate formulation. (b) Example depiction of the variations in polypeptide charge density and hydrophobicity of the cationic, lysine (K)-containing polymers. Charge blockiness is defined by the parameter τ, while hydrophobicity is indicated by the color of the gray blocks as the neutral amino acid spacers go from glycine (G) to alanine (A) to leucine (L). (c) Experimental design matrix to study the effect of polypeptide characteristics such as charge patterning and hydrophobicity on virus encapsulation. (d) Plot of the charge ratio (K+/E−) as a function of the total cationic charge fraction from the polypeptides present in the system. The dashed line represents a guide for the eye. Some images made with BioRender.com.

Figure 2

Trends in virus encapsulation as a function of polypeptide chain length. (a–c) PPV and (d–f) HRV titers in the supernatant and coacervate phases as a function of charge stoichiometry for coacervates formed from poly(lysine) and poly(glutamate) of chain length N = 48, 400, and 800. Yellow boxes highlight the range of charge fractions over which ln (P) > 0. (g,h) Corresponding partition coefficient ln (P) for (g) PPV and (h) HRV as a function of charge stoichiometry showing the shift toward lower charge fraction with increasing chain length. Coloring of the background in the plots indicates when ln (P) > 0, meaning that the virus preferred the coacervate phase and ln (P) < 0 represents the virus preferring the supernatant phase. The data are the average of three encapsulation experiments with the error bars shown as the standard deviation of the replicate measurements. Open symbols indicate that the data were measured at the limit of detection of the MTT assay.

Figure 3

Trends in virus encapsulation as a function of polypeptide charge density. Plot of ln (P) as a function of charge stoichiometry for (a,c) PPV and (b,d) HRV in coacervates made from charge-patterned polypeptides (N = 48) of (a,b) K_nG_n with E₄₈ and (c,d) K₄₈ with E_nG_n. Coloring of the background in the plots indicates when ln (P) > 0, meaning that the virus preferred the coacervate phase, and ln (P) < 0 represents the virus preferring the supernatant phase. The data are the average of three encapsulation experiments with the error bars shown as the standard deviation of the replicate measurements. Open symbols indicate that the data were measured at the limit of detection.

Figure 4

Virus surface charge. Structural depiction of the location of (a,b) charged residues and (c,d) resulting electrostatic potential on the (a,c) PPV (PDB: 1K3V(66)) and (b,d) HRV (PDB: 4RHV(67)) capsid. The protonation state was calculated at pH 8 using the PROPKA method on the Poisson–Boltzmann server considering lysine (K), arginine (R), glutamate (E), and aspartate (D) as the ionizable residues. Images were generated using ChimeraX 1.3. Radial distribution function [g(r)] with respect to each charged residue on the (e) PPV and (f) HRV subunits. Only the charged residues that appear on the capsid surface and are solvent accessible were considered in this calculation.

Figure 5

Trends in virus encapsulation as a function of polypeptide hydrophobicity. (a) Visual depiction of the Eisenberg hydrophobicity scale and the hydrophobic amino acids used. Plot of ln (P) as a function of charge stoichiometry for (b) PPV and (c) HRV in coacervates made from charge-patterned polypeptides (N = 48) with increasing hydrophobicity (G < A < L). Coloring of the background in the plots indicates when ln (P) > 0, meaning that the virus preferred the coacervate phase, and ln (P) < 0 represents the virus preferring the supernatant phase. The data are the average of three encapsulation experiments with the error bars shown as the standard deviation of the replicate measurements.

Figure 6

Virus surface hydrophobicity. Structural depiction of the location of hydrophobic residues on the (a) PPV (PDB: 1K3V(66)) and (b) HRV (PDB: 4RHV(67)) capsid. Images were generated using ChimeraX 1.3. Radial distribution function [g(r)] is given with respect to each hydrophobic residue on the (e) PPV and (f) HRV subunits. Only the residues that appear on the capsid surface and are solvent accessible were considered in this calculation.