On Dec 20, 2021, Capraz et al. from spanning institutions such as the University of Natural Resources and Life Sciences (BOKU) in Austria, published a paper in the journal eLife, titled "Structure-guided glyco-engineering of ACE2 for improved potency as soluble SARS-CoV-2 decoy receptor". The paper details a rational, structure-guided approach to enhance the neutralizing capability of recombinant soluble human angiotensin-converting enzyme 2 (ACE2). The core finding demonstrates that selectively removing specific N-glycans from ACE2 drastically improves its binding affinity to the SARS-CoV-2 Spike protein, thereby creating a superior decoy receptor with therapeutic promise.

Overview

Viral entry for Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) is fundamentally dependent on the interaction between its trimeric Spike glycoprotein and the host cell receptor, ACE2. While recombinant soluble ACE2 has proven effective as a "decoy receptor"—a competitive inhibitor that binds the Spike protein and prevents it from initiating infection—there has been a persistent drive to improve its affinity and efficacy. This endeavor is especially critical given the emergence of viral variants.

The interaction is profoundly influenced by glycosylation. Both Spike and ACE2 are heavily glycosylated proteins; these Extensive Glycan shields play a dual role, often protecting the protein surface but also critically mediating or modulating protein-protein recognition events. Previous research provided conflicting computational insights into the precise influence of individual ACE2 N-glycans on Spike binding, making the task of rational engineering challenging. This provided a compelling case for a deep-dive, integrated approach combining computational modeling and experimental validation to definitively map which ACE2 glycans contribute to the interaction and, more importantly, whether their effect is favorable or detrimental to binding affinity. The overarching goal was to apply principles of Synthetic Glycobiology to optimize the molecular architecture of ACE2 for maximal potency as a therapeutic agent.

Research Results

• Computational Glycan Modeling and Dynamics

The study began with a meticulous in silico strategy, establishing fully glycosylated, atomistic 3D models of the trimeric Spike protein complexed with dimeric hACE2. The researchers first elucidated the entire glycome of their rshACE2 construct, including sites previously underreported, and then structurally modeled the complex and performed extensive all-atom molecular dynamics (MD) simulations. This computational approach was essential for exploring the inherently flexible and dynamic nature of the glycans, which is often difficult to capture using static crystal or cryo-EM structures alone.

The MD simulations tracked the conformational distribution and dynamic effects of the seven ACE2 N-glycans on the interaction interface. By analyzing atomic contact and hydrogen bond formation over the simulation time, the team identified the glycans at N90 and N322 as the most prominent interaction determinants. These two glycans exhibited the highest propensity for hydrogen bond formation with the Spike protein, suggesting they were indeed structurally relevant to the binding interface. This sophisticated modeling provided the necessary structural and dynamic context to move from simple observation to rational prediction of functional impact.

Fig.1 3D structural model of the glycosylated Spike-hACE2 complex. (Capraz, et al., 2021)

• Unveiling Entropic and Steric Barriers

Following the identification of the key glycans, the team used the dynamic data to assess the precise mechanism of their influence. A critical finding was derived from analyzing the conformational freedom of the ACE2 glycans upon binding to the Spike protein. The density maps of the unbound ACE2 glycans showed extensive, continuous coverage of the interface area. However, upon formation of the Spike-ACE2 complex, this conformational freedom was dramatically reduced, particularly for the N90 and N322 glycans.

This reduction in conformational entropy upon complex formation carries a significant energetic penalty. The researchers hypothesized that the highly dynamic N90 and N322 glycans, in addition to potentially imposing minor steric restraints, primarily impede the binding process through an unfavorable entropic loss. If the glycan is highly flexible when unbound but rigidly constrained when bound, the associated free energy cost opposes complex formation, ultimately lowering the binding affinity. This counterintuitive finding challenged some previous computational predictions that suggested a favorable interaction at these sites and firmly established the potential for de-glycosylation as a potent glyco-engineering strategy.

• Targeted Glyco-Engineering for Enhanced Decoy Potency

Utilizing site-directed mutagenesis, the team engineered ACE2-Fc fusion constructs where the N-glycosylation sites N90 and N322 were individually ablated. Specifically, the N322 residue was mutated to glutamine (N322Q) to prevent glycosylation without disrupting the underlying protein structure, which can be an issue with other mutation strategies.

The resulting mutant proteins were expressed and purified, and their binding kinetics were assessed. As predicted by the MD simulations, the single glycan removal at N322 led to a measurable enhancement in Spike binding affinity. The study confirmed that the engineered, de-glycosylated ACE2 variants demonstrated significantly improved virus neutralization capacity in vitro. This synergistic result, bridging advanced computational modeling with precise biochemical validation, culminated in the conclusion that simultaneously removing all accessible N-glycans from the recombinant soluble human ACE2 yields a dramatically superior SARS-CoV-2 decoy receptor, substantially increasing its binding strength and therapeutic promise.

Fig.2 Binding of Spike and RBD to glyco-engineered ACE2 variants. (Capraz, et al., 2021)

Conclusion

By rigorously integrating molecular dynamics simulations with targeted glyco-engineering, the authors have provided definitive, experimentally validated proof that specific N-glycans on ACE2, far from stabilizing the interaction, actually impose entropic and steric penalties that reduce the receptor's affinity for the Spike protein. The resulting optimized, de-glycosylated ACE2 variant offers a significant leap in potency over the native form. This rational design strategy of eliminating unfavorable glycan contributions paves a high-impact pathway for developing highly effective, broad-spectrum decoy receptors, potentially accelerating the development of novel therapeutic proteins with enhanced functionality against infectious diseases where glycan modulation is a critical factor.