Molecular mechanisms of immune evasion by host protein glycosylation of a bacterial immunogen used in nucleic acid vaccines

This study reveals that host-derived N-glycosylation of the Mycobacterium tuberculosis Ag85B antigen in nucleic acid vaccines creates a physical and biochemical shield that occludes immune epitopes and recruits inhibitory Siglec-9 receptors, thereby explaining the vaccine's lack of efficacy and highlighting the need for glycosylation-aware immunogen engineering.

The Gist

The Big Idea: The "Imposter" Vaccine

Imagine you are trying to teach a security guard (your immune system) how to recognize a specific criminal (a bacteria like Mycobacterium tuberculosis).

Usually, you would show the guard a photo of the criminal's face. But in this new type of vaccine (called an nucleic acid vaccine, like mRNA vaccines), instead of giving the guard a photo, you give the guard a blueprint and tell them, "Go build a 3D model of this criminal right here in the workshop."

The problem? The workshop is inside a human cell. When the human cell builds the model, it accidentally paints the criminal with human makeup (sugar coats called glycans) that the real criminal never wears.

This paper discovers that this "human makeup" is actually a disguise. It hides the criminal's face, making it impossible for the security guard to recognize him. This explains why some new vaccines for bacteria haven't worked as well as hoped.

---

The Story of Ag85B: The Bacterial Target

The scientists focused on a specific part of the tuberculosis bacteria called Ag85B.

• In the wild: The bacteria wears a plain, rugged jacket. It has no sugar coats.
• In the vaccine: When the vaccine tells a human cell to build Ag85B, the cell treats it like a human protein. It slaps a thick, sticky layer of complex sugars (glycans) all over it.

The researchers asked: "Does this sugar coat help or hurt the vaccine?"

The Investigation: Putting on the "Sugar Shield"

The team used high-tech tools to look at the sugar-coated Ag85B. Here is what they found:

1. The "Fuzzy Blanket" Effect (Structural Change)

Imagine the bacterial protein is a statue. The human cell wraps it in a thick, fluffy, fuzzy blanket made of sugar.

• The Result: The blanket covers up the statue's features.
• The Science: Using computer simulations (like a digital wind tunnel), they saw that these sugar blankets take up a lot of space. They physically block the surface of the protein.

2. Hiding the "Wanted Posters" (Antibody Evasion)

Your body makes "wanted posters" called antibodies to catch the bacteria. These antibodies need to grab onto specific spots on the bacteria's surface to do their job.

• The Problem: The sugar blanket covers these grab-holds.
• The Analogy: It's like trying to put a sticker on a wall, but someone has painted a thick layer of frosting over the wall. The sticker won't stick.
• The Proof: The scientists tested this in a lab. When they tried to stick an antibody to the sugar-coated bacteria, it barely held on. When they tried it on the plain, non-sugar-coated bacteria, it stuck perfectly. The sugar coat reduced the antibody's ability to grab the bacteria by a massive amount.

3. Hiding the "ID Card" (T-Cell Evasion)

Your immune system also has "detectives" called T-cells. These detectives need to see a piece of the bacteria (an ID card) to know it's an enemy.

• The Problem: The sugar blanket covers the ID card.
• The Result: The detectives can't see the ID, so they don't get the alarm. The scientists found that the sugar coat blocked the T-cells from recognizing the bacteria, leading to a much weaker immune response.

4. The "Do Not Disturb" Sign (Immune Suppression)

This is the sneakiest part. The sugar coat isn't just a blanket; it's also a signal.

• The Analogy: The sugar coat has a specific pattern (sialic acid) that looks like a "Do Not Disturb" sign to the immune system.
• The Science: The immune system has sensors (called Siglec-9) that read this sign. When they see the sugar coat, they think, "Oh, this is a friendly human cell. Don't attack!"
• The Result: Instead of attacking the bacteria, the immune system actually gets told to calm down and stop fighting.

The Conclusion: Why This Matters

The scientists concluded that when we use these fancy new vaccines for bacteria, we might be accidentally turning the bacteria into a "wolf in sheep's clothing." The human cell adds a sugar coat that:

1. Hides the bacteria's face from antibodies.
2. Blocks the ID card from T-cells.
3. Tricks the immune system into thinking the bacteria is friendly.

The Solution?
The paper suggests that in the future, when we design vaccines for bacteria, we need to be "glycosylation-aware." We might need to edit the blueprint so the human cell doesn't add the sugar coat, or we need to remove the specific spots where the sugar gets attached. This way, the bacteria looks exactly like the real criminal, and the immune system can finally catch it.

Summary in One Sentence

This paper explains that some new vaccines fail against bacteria because the human body accidentally covers the bacterial target in a "sugar disguise" that hides it from the immune system and tricks it into standing down.

Technical Summary

1. Problem Statement

Nucleic acid vaccines (DNA and mRNA) function by introducing genetic sequences into host cells to drive in situ expression of pathogen antigens. While highly effective for viral pathogens (which naturally express proteins within host cells), this platform faces a critical challenge when applied to non-viral pathogens (e.g., bacteria, parasites).

• The Mismatch: Bacterial antigens are not naturally glycosylated. However, when expressed in mammalian cells (the host for nucleic acid vaccines), they undergo host-derived post-translational modifications (PTMs), specifically N-linked glycosylation.
• The Consequence: These non-native glycans may structurally remodel the antigen, masking critical B-cell and T-cell epitopes or engaging immunoregulatory receptors (e.g., Siglecs), thereby dampening the immune response.
• Case Study: Despite being a primary target for Tuberculosis (TB) vaccines, nucleic acid vaccines targeting the Mycobacterium tuberculosis Antigen 85B (Ag85B) have failed in clinical trials. The authors hypothesize that host-imposed glycosylation is the mechanism behind this lack of protective efficacy.

2. Methodology

The study employed a multi-disciplinary approach combining structural biology, mass spectrometry, computational modeling, and immunology:

• Protein Expression: Ag85B was expressed in two systems:
  1. Native-like: E. coli BL21 (non-glycosylated).
  2. Host-mimic: Human Expi293 cells (glycosylated).
• Glycomics and Glycoproteomics:
  • Site Occupancy: Used EndoH/PNGase F digestion with heavy water ($H_2^{18}O$) labeling and LC-MS/MS to quantify glycosylation at specific sequons.
  • Glycan Profiling: Released N-glycans were analyzed via LC-MS to determine composition (high-mannose, hybrid, complex), sialylation, and fucosylation.
  • Glycopeptidomics: Site-specific glycan structures were identified using stepped high-energy collisional dissociation (sHCD) and the pGlyco3 search engine.
• Structural Modeling & Molecular Dynamics (MD):
  • The crystal structure of native Ag85B (PDB: 1F0N) was modified with site-specific glycans identified via glycoproteomics using GLYCAM-Web.
  • 1-µs MD simulations were performed to assess conformational flexibility and surface accessibility.
  • Accessibility Analysis: Calculated Solvent-Accessible Surface Area (SASA) using a 1.4 Å probe (water) and Antibody-Accessible Surface Area (ASA) using a 10 Å probe (antibody).
• Immunological Assays:
  • Antibody Binding: Competitive ELISA and Biolayer Interferometry (BLI) to measure affinity ($K_D$) of a monoclonal antibody (mAb) against glycosylated vs. non-glycosylated Ag85B.
  • Epitope Mapping: Hydrogen-Deuterium Exchange Mass Spectrometry (HDX-MS) to identify the mAb epitope.
  • T-Cell Proliferation: Ex vivo stimulation of human PBMCs from BCG-vaccinated donors to measure CD4+ T-cell proliferation.
  • Receptor Binding: Lectin blotting to assess binding to Siglec-9 (an inhibitory lectin) and the effect of sialidase treatment.

3. Key Results

A. Structural and Biochemical Characterization

• Microheterogeneous Glycosylation: Ag85B expressed in Expi293 cells is glycosylated at four canonical sequons: N52, N224, N234, and N280.
• Glycan Composition: The glycans are predominantly complex type (72%), highly fucosylated (91%), and significantly sialylated (46%).
• Site-Specific Diversity:
  • N52: Predominantly high-mannose.
  • N224 & N280: Predominantly complex, fucosylated structures.
  • N234: A mix of oligomannose, hybrid, and complex structures.

B. Structural Impact (MD Simulations)

• Surface Occlusion: Glycosylation significantly reduced the solvent-accessible surface area (SASA) of the protein (107 nm² for glycosylated vs. 116 nm² for non-glycosylated).
• Epitope Shielding: The glycans occupy substantial conformational space, drastically reducing antibody-accessible surface area (ASA) for residues near glycosylation sites.
  • Antibody Epitopes: Sites overlapping or adjacent to glycans (e.g., N276-Y287, W288-G306) showed 81–99% loss of accessibility.
  • T-Cell Epitopes: Known MHC-restricted peptides (e.g., F121-P138, K220-L228, I245-F253) showed significant accessibility loss (up to 100%) when proximal to glycosylation sites.

C. Functional Immunological Impact

• Reduced Antibody Binding:
  • ELISA: Glycosylated Ag85B failed to compete effectively for mAb binding compared to the non-glycosylated form.
  • BLI: The dissociation constant ($K_D$) for glycosylated Ag85B was 298 µM, compared to 54 nM for the non-glycosylated form (a ~5,500-fold reduction in affinity).
  • HDX-MS: Confirmed that the mAb epitope (containing Gln105 and Ala112) is physically shielded by the glycan cloud in the glycosylated model.
• Impaired T-Cell Response:
  • Glycosylated Ag85B induced a significantly reduced CD4+ T-cell proliferation compared to the non-glycosylated form in human PBMCs.
• Immunosuppressive Signaling:
  • Glycosylated Ag85B bound to Siglec-9 (an inhibitory receptor).
  • This binding was abolished by sialidase treatment, confirming that terminal sialic acids mediate the interaction, potentially actively suppressing the immune response.

4. Key Contributions

1. Mechanistic Explanation for Vaccine Failure: Provides the first structural and biochemical evidence explaining why nucleic acid vaccines targeting bacterial antigens (specifically Ag85B for TB) may fail: the host modifies the antigen, rendering it "invisible" or "tolerogenic" to the immune system.
2. Dual Mechanism of Evasion: Identifies a two-pronged evasion strategy:
   • Passive: Physical shielding of B-cell and T-cell epitopes by bulky glycans.
   • Active: Engagement of inhibitory Siglec receptors via sialic acid moieties.
3. Design Principles for Nucleic Acid Vaccines: Establishes that for non-viral pathogens, antigen engineering must include glycosylation-aware design. This involves:
   • Removing vulnerable N-glycosylation sequons (N-X-S/T) via mutagenesis.
   • Ensuring antigen trafficking avoids glycosylation pathways if native-like structure is required.

5. Significance

This study fundamentally shifts the paradigm for developing nucleic acid vaccines against non-viral pathogens. It demonstrates that the "host" is not a neutral factory but an active modifier that can structurally and immunologically alter the vaccine antigen.

• Clinical Relevance: Explains the failure of previous TB nucleic acid vaccine trials and offers a clear path forward for redesigning these candidates.
• Broader Impact: The findings apply to any nucleic acid vaccine targeting bacteria or parasites (e.g., malaria, leishmaniasis). Future vaccine design must prioritize epitope integrity by preventing aberrant host glycosylation, ensuring the immune system recognizes the pathogen as it naturally exists, not as a modified host protein.