Structures of annexin A2-PS DNA complexes show dominance of hydrophobic interactions in phosphorothioate binding

Control of the Assembly and Disassembly of Spherical Nucleic Acids Is Critical for Enhanced Gene Silencing

Spherical nucleic acids─nanospheres with nucleic acids on their corona─have emerged as a promising class of nanocarriers, aiming to address the shortcomings of traditional nucleic therapeutics, namely, their poor stability, biodistribution, and cellular entry. By conjugating hydrophobic monomers to a growing nucleic acid strand in a sequence-defined manner, our group has developed self-assembled spherical nucleic acids (SaSNAs), for unaided, enhanced gene silencing. By virtue of their self-assembled nature, SaSNAs can disassemble under certain conditions in contrast to covalent or gold nanoparticle SNAs. Gene silencing involves multiple steps including cellular uptake, endosomal escape, and therapeutic cargo release. Whether assembly vs disassembly is advantageous to any of these steps has not been previously studied. In this work, we modify the DNA and hydrophobic portions of SaSNAs and examine their effects on stability, cellular uptake, and gene silencing. When the linkages between the hydrophobic units are changed from phosphate to phosphorothioate, we find that the SaSNAs disassemble better in endosomal conditions and exhibit more efficacious silencing, despite having cellular uptake similar to that of their phosphate counterparts. Thus, disassembly in the endolysosomal compartments is advantageous, facilitating the release of the nucleic acid cargo and the interactions between the hydrophobic units and endosomal lipids. We also find that SaSNAs partially disassemble in serum to bind albumin; the disassembled, albumin-bound strands are less efficient at cellular uptake and gene silencing than their assembled counterparts, which can engage scavenger receptors for internalization. When the DNA portion is cross-linked by G-quadruplex formation, disassembly decreases and cellular uptake significantly increases. However, this does not translate to greater gene silencing, again illustrating the need for disassembly of the SaSNAs when they are in the endosome. This work showcases the advantages of the dual nature of SaSNAs for gene silencing, requiring extracellular assembly and disassembly inside the cell compartments.

AQcalc: A web server that identifies weak molecular interactions in protein structures

Weak molecular interactions play an important role in protein structure and function. Computational tools that identify weak molecular interactions are, therefore, valuable for the study of proteins. Here, we present AQcalc, a web server (https://aqcalcbiocomputing.com/) that can be used to identify anion-quadrupole (AQ) interactions, which are weak interactions involving aromatic residue (Trp, Tyr, and Phe) ring edges and anions (Asp, Glu, and phosphate ion) both within proteins and at their interfaces (protein-protein, protein-nucleic acids, and protein-lipid bilayer). AQcalc identifies AQ interactions as well as clusters involving AQ, cation-π, and salt bridges, among others. Utilizing AQcalc we analyzed weak interactions in protein models, even in the absence of experimental structures, to understand the contributions of weak interactions to deleterious structural changes, including those associated with oncogenic and germline disease variants. We identified several deleterious variants with disrupted AQ interactions (comparable in frequency to cation-π disruptions). Amyloid fibrils utilize AQ to bury anions at frequencies that far exceed those observed for globular proteins. AQ interactions were detected three and five times more frequently than the hydrogen-bonded AQ (HBAQ) in fibril structures and protein-lipid bilayer interfaces, respectively. By contrast, AQ and HBAQ interactions were detected with similar frequencies in globular proteins. Collectively, these findings suggest AQcalc will be effective in facilitating fine structural analysis. As other web utilities designed to identify protein residue interaction networks do not report AQ interactions, wide use of AQcalc will enrich our understanding of residue interaction networks and facilitate hypothesis testing by identifying and experimentally characterizing these comparably weak but important interactions.