Engineered OAA lectins as selective and sensitive high mannose glycan targeting tools

This study utilizes phage display and structural analysis to engineer Oscillatoria agardhii agglutinin (OAA) lectin variants with high selectivity and affinity for specific high mannose N-glycans, demonstrating their utility as precise profiling tools and tunable antiviral agents.

The Gist

Imagine the human body is a bustling city, and the proteins on the surface of our cells (and viruses) are like buildings. These buildings are covered in a sticky, sugary coating called glycans. Think of these glycans as different types of "ID badges" or "uniforms" that tell other cells who is who.

Some of these badges are made entirely of a specific sugar called mannose. Scientists call these "High Mannose" badges. The problem is, these badges come in many slightly different sizes: some have 5 mannose sugars (M5), some have 6 (M6), up to 9 (M9). It's like having a family of twins where the only difference is the number of buttons on their shirts.

The Problem:
Scientists have a tool called a lectin (specifically one called OAA) that acts like a security guard. This guard is good at spotting the "mannose family," but it's a bit clumsy. It grabs anyone with a mannose badge, whether they have 5 buttons or 9. It can't tell the difference between the specific sizes. This makes it hard to study specific diseases or catch specific viruses that wear a particular size badge.

The Solution: Engineering a Better Guard
The researchers in this paper decided to take this clumsy security guard and use a technique called Phage Display to "evolve" it into a super-specialized detective.

Here is how they did it, using a simple analogy:

1. The "Giant Library" of Mutations

Imagine you have a master key (the OAA lectin) that opens many doors. You want to make a new key that only opens the door with exactly 5 buttons (M5).
Instead of trying to guess the right shape, the scientists created a giant library of millions of slightly broken keys. They randomly changed the teeth on the keys (mutated the protein) to see which ones worked best.

2. The "Try-On" Contest

They put these millions of mutant keys in a room with a pile of "5-button badges" (M5 glycans).

• The bad keys fell off.
• The keys that stuck a little bit stayed.
• They washed away the weak ones and kept the best ones.
• They repeated this process four times, like a game of "survival of the fittest," until they found the keys that stuck the tightest and most specifically to the 5-button badges.

3. The Discovery: Two New Super-Tools

From this contest, they found two amazing new versions of the security guard:

• The "Sniper" (Variant V4):
  This new guard is incredibly picky. It looks at the 5-button badge and says, "Yes, you're the one!" But if it sees a 6-button badge, it says, "Nope, wrong size," and walks away.

  • How it works: The scientists found that by changing just a few specific "teeth" on the key, they physically blocked the space where a 6-button badge would fit. It's like putting a wall in the keyhole so only the 5-button key can turn.
  • The Magic: When they made this guard "double-sided" (giving it two hands instead of one), its pickiness went from "pretty good" to "absolutely perfect." It became 200 times better at ignoring the wrong badges.

• The "Super Glue" (Variant PM6):
  This guard is the opposite. It isn't picky about the size; it just wants to stick harder to any mannose badge.

  • How it works: By shifting a small part of the guard's structure (like moving a shelf inside a closet), it created a better fit for all sizes of mannose badges.
  • The Magic: When made double-sided, this guard became 26 times stickier than the original. It's like upgrading from a piece of tape to industrial-strength super glue.

4. Putting Them to Work

The scientists tested these new tools in two real-world scenarios:

• Sorting the Mail (Glycan Profiling):
  They used the "Sniper" (V4) to sort through a mixed bag of proteins (RNase B). The original guard grabbed everything. The new Sniper grabbed only the proteins with the 5-button badge, leaving the rest behind. This is huge for scientists who need to study specific types of proteins without the noise of the others.

• Stopping the Virus (Antiviral Defense):
  They tested the "Super Glue" (PM6) against the SARS-CoV-2 virus. Viruses wear mannose badges to hide from our immune system.

  • The original guard could stop the virus, but it took a lot of it.
  • The "Super Glue" version stopped the virus 4 times better than the original, even though it's smaller than other known antiviral tools.
  • Interestingly, the "Sniper" (V4) failed to stop the virus. Why? Because the virus is a shapeshifter; it wears a mix of 5, 6, 7, and 8-button badges. The Sniper was too picky to catch the virus, while the "Super Glue" caught them all.

The Big Takeaway

This paper shows that we can take a natural protein that is a bit "sloppy" and, through smart engineering, turn it into two distinct, powerful tools:

1. A precision scalpel that can cut out only the exact sugar structure you want to study.
2. A heavy-duty magnet that can grab onto viruses or cancer cells with incredible strength.

It proves that by understanding the tiny details of how proteins fit together, we can redesign nature's tools to be much more effective at fighting disease and understanding biology.

Technical Summary

1. Problem Statement

High mannose N-glycans (HMGs) are critical biomarkers for protein folding, cell signaling, and disease states (e.g., cancer, viral infections). HMGs share a common pentamannose core (Man5GlcNAc2, or M5), with variations (M6–M9) arising from the addition of mannose residues to specific arms (D1, D2, D3).

• The Challenge: There is a scarcity of tools capable of distinguishing between specific HMG structures (e.g., M5 vs. M6) in a biological context. Existing lectins often bind promiscuously to the entire HMG class or lack the sensitivity to resolve specific glycoforms.
• The Goal: To engineer a lectin scaffold capable of high-affinity, structure-specific binding to individual HMGs (specifically M5) while maintaining stability and evolvability.

2. Methodology

The authors utilized the Oscillatoria agardhii agglutinin (OAA), a stable 14 kDa cyanobacterial lectin with a unique $\beta$-barrel topology and two binding sites, as a scaffold.

• Library Design & Phage Display:
  • A combinatorial library of OAA variants was generated by mutating 21 residues located in three loops (Loop 1, 2, and 3) surrounding the M5 binding pocket.
  • To isolate mutations affecting a single binding site, the second binding site was rendered inactive via three alanine mutations (W10A, R95A, E123A).
  • The library was displayed on M13 bacteriophage and subjected to four rounds of enrichment against streptavidin beads coated with biotinylated M5 (Man5GlcNAc2).
• Screening & Selection:
  • Enriched phage populations were analyzed via amplicon sequencing to identify sequence motifs.
  • Top variants (V1–V10) were expressed in E. coli, purified, and screened using Bio-Layer Interferometry (BLI) against a panel of HMGs (M5–M9).
• Structural & Mutational Analysis:
  • The most promising variant (V4) was crystallized in its bivalent form (V4V4) bound to M5 to determine the structural basis of selectivity.
  • A systematic mutational dissection (25 point-mutation variants) was performed to isolate the specific contributions of individual mutations to affinity and selectivity.
• Functional Validation:
  • Engineered variants were tested for binding to glycoproteins (RNase B) and the SARS-CoV-2 spike protein.
  • Viral inhibition assays (focus reduction neutralization) were conducted to assess antiviral potential.

3. Key Contributions

• Development of a Selective M5 Tool: The creation of V4, a variant with high selectivity for the M5 core over extended HMGs (M6–M9), a feat rarely achieved with natural lectins.
• Discovery of an Affinity-Enhanced Tool: The identification of PM6, a variant with significantly improved affinity for all HMGs (including M9), demonstrating that distal loop mutations can enhance binding without altering direct contact residues.
• Mechanistic Insight into Co-Dependent Mutations: The study reveals that selectivity is not achieved by single mutations but through a network of co-dependent mutations that remodel the binding pocket. Specifically, the paper elucidates how mutations in conserved residues (Q31, N75, R28) and loop shifts create a steric occlusion of the M6 binding site.
• Avidity Engineering: The successful transfer of monovalent mutations to a bivalent scaffold (V4V4 and PM6PM6), demonstrating that bivalency can amplify selectivity by orders of magnitude (e.g., increasing M5 preference over M6 from 7-fold to 205-fold).

4. Key Results

A. Identification of Variant V4 (M5 Selective)

• Binding Profile: V4 showed a $K_D$ of ~2.2 $\mu$M for M5 but a ~60-fold reduction in affinity for M6–M8 and no detectable binding to M9.
• Selectivity Mechanism: Crystallography of V4V4 revealed:
  • Loop 1 Shift: A 2 Å inward shift of Loop 1 driven by the S27Y mutation.
  • Salt Bridge Formation: The N75E mutation forms a salt bridge with R28, causing a rotamer flip of R28.
  • Steric Occlusion: The Q31R mutation releases constraints, allowing the R28 flip. This repositioning, combined with the S27Y shift, shrinks the binding pocket, physically blocking the D1 arm extension found in M6 and larger HMGs.
• Mutational Dependencies: Single mutations (e.g., N75E or Q31R alone) were often deleterious or non-selective. Selectivity required the specific combination of S27Y, Q31R, and N75E (and S86Y) to stabilize the new conformation.

B. Identification of Variant PM6 (High Affinity)

• Binding Profile: PM6 (containing S27Y and S86Y) exhibited an ~8-fold improvement in affinity for M9 compared to wild-type (WT) and sub-100 nM affinity for M7.
• Mechanism: The tyrosine mutations (S27Y, S86Y) shift Loop 1 inward, optimizing the pre-existing hydrogen bond network with the glycan core rather than creating new contacts. This "buttressing" effect stabilizes the ligand.

C. Bivalent Amplification

• V4V4: Restoring bivalency to the selective variant resulted in a 205-fold preference for M5 over M6 (compared to 7-fold for monovalent V4) and a $K_D$ of 18 nM for M5.
• PM6PM6: The bivalent affinity-enhanced variant showed a $K_D$ of 22 nM for M9, a 26-fold improvement over WT, establishing it as a potent, unbiased HMG binder.

D. Biological Applications

• Glycoform Separation: In pulldown assays with RNase B (a mixture of M5–M9), V4 selectively captured >96% M5 glycoforms, effectively separating them from the mixture.
• Antiviral Activity:
  • PM6PM6 demonstrated potent inhibition of SARS-CoV-2 with an $IC_{50}$ of 1.7 nM, comparable to the tetrameric lectin BOA (0.136 nM) despite having only two binding sites.
  • V4V4 failed to inhibit SARS-CoV-2, highlighting that viral inhibition requires broad HMG recognition due to the heterogeneity of glycans on the viral spike protein.

5. Significance

• Precision Glycobiology: This work provides the first highly specific, engineered tool to detect and isolate the M5 glycoform, which is a critical marker for serum protein clearance and stability.
• Engineering Paradigm: The study challenges the notion that lectin engineering requires stepwise evolution. It demonstrates that combinatorial libraries can access "loss-of-function" pathways that are rescued by co-mutations, allowing for the creation of phenotypes (like extreme selectivity) that are inaccessible via rational design or natural selection.
• Therapeutic Potential: The engineered variants offer a new class of tunable antivirals. The ability to modulate affinity and selectivity allows for the design of agents that either broadly neutralize viruses (via high-affinity bivalent binding) or specifically target glycan-dependent disease mechanisms.
• Scaffold Versatility: The successful engineering of the OAA $\beta$-barrel confirms its utility as a robust platform for developing carbohydrate-binding proteins (CBPs) with tailored specificities, paving the way for future designer glycan probes and therapeutics.