HCF1 orchestrates O-GlcNAcylation and affinity-dependent transcription through extended molecular determinants and register-shifted binding

This study expands the known interactome of the transcriptional co-regulator HCF1 by identifying diverse canonical and non-canonical binding partners, elucidating how extended molecular determinants and register-shifted sequences govern affinity-dependent transcriptional activity and O-GlcNAcylation.

The Gist

The Story of HCF1: The Master Keyholder

Imagine your cell is a massive, bustling city. Inside this city, there are thousands of different jobs being done every second: building houses (making proteins), turning on streetlights (activating genes), and managing traffic (controlling cell growth).

For a long time, scientists knew about a very important character in this city named HCF1. They knew HCF1 was like a Master Keyholder or a Concierge. Its job was to stand at the door of the "Gene Library" and decide which keys (proteins) were allowed in to turn on the lights (start transcription).

For 30 years, scientists thought they knew exactly what the keys looked like. They believed HCF1 only accepted keys with a very specific shape: a pattern of letters like D-E-H-x-Y (think of it as a specific barcode). If a protein didn't have this exact barcode, the Concierge (HCF1) would ignore it.

But this new study reveals that HCF1 is much more complex and picky than we thought.

Here is what the researchers discovered, broken down into three main parts:

---

1. The "Barcode" Was Too Simple (The Screening)

The researchers decided to test this old rule. They built a giant library containing thousands of different "keys" (peptides) from the human body. Some had the old barcode (D-E-H-x-Y), and some had a slightly different one (N-Q-H-x-Y).

They threw all these keys at the Concierge (HCF1) to see which ones got a "VIP pass" to bind to him.

• The Surprise: They found that just because a key had the barcode didn't mean the Concierge would open the door. Many keys with the barcode were rejected.
• The Discovery: They found 41 new keys that HCF1 actually likes, which nobody knew about before. These include keys that help control cell growth and manage DNA.
• The Lesson: Having the barcode is necessary, but it's not enough. The Concierge is looking for something more specific.

2. The "Hidden Details" Matter (Deep Mutational Scanning)

To understand why some keys worked and others didn't, the researchers played a game of "What If?" They took a working key and changed one tiny letter at a time, like changing a single letter in a password.

They discovered that HCF1 isn't just looking at the main barcode (the "anchor" letters). It is also checking the surrounding neighborhood.

• The Analogy: Imagine the barcode is the main lock on a door. The researchers found that HCF1 also checks the doormat, the handle, and the frame around the door.
• The "x" Factor: In the middle of the barcode, there is a placeholder letter called "x." The study found that this "x" spot is very sensitive. If you put a bulky, heavy letter there, the door won't open. If you put a small, light letter there, it works perfectly.
• The Result: They created a new, super-accurate rulebook for what makes a key fit. This explains why some proteins that looked like they should work, actually didn't.

3. The "Two-Step" Keys (Non-Canonical Binders)

The biggest shock came when they found keys that didn't follow the rules at all. They found proteins like IRF1 (a protein that fights viruses) and SMCHD1 that bind to HCF1, but their barcodes were weird.

Instead of the standard D-H-x-Y pattern, these proteins had a D-H-x-x-Y pattern.

• The Analogy: Imagine the standard key has a gap of one tooth between the handle and the tip. These new keys had a gap of two teeth. It was like a "register-shifted" key.
• The Twist: Even though the gap was different, the key still fit! The Concierge (HCF1) stretched out to grab it.
• Why it matters: When the researchers tested this with the IRF1 protein, they found that the strength of this "weird" bond directly controlled how well IRF1 could fight viruses and stop cancer cells from growing. If they made the bond stronger, the cell stopped growing faster. If they broke the bond, the cell went wild.

4. The "Sugar Coating" Connection (O-GlcNAcylation)

Finally, the study connected HCF1 to a process called O-GlcNAcylation.

• The Analogy: Think of O-GlcNAcylation as sprinkling sugar on proteins. This "sugar" acts like a sticker that tells the protein, "Go to work! Stay stable! Turn on the lights!"
• The Discovery: The researchers found that HCF1 acts like a delivery truck. It picks up its "VIP keys" (the proteins it binds to) and brings them right next to the Sugar Sprinkler (an enzyme called OGT).
• The Result: Because HCF1 brings these proteins close to the sprinkler, they get covered in sugar. This sugar coating helps them do their jobs better. When the researchers blocked HCF1, the proteins stopped getting their sugar coating, and their jobs suffered.

---

The Big Picture: Why Should We Care?

This paper changes how we see HCF1 in three ways:

1. It's a Bigger Club: HCF1 interacts with way more proteins than we thought, including many that control cancer and cell growth.
2. It's Picky: It doesn't just look at the main barcode; it checks the whole shape of the key. This helps us predict which proteins are actually important in diseases.
3. It's a Cancer Target: Since HCF1 helps cancer cells grow (by bringing them sugar-coating and turning on growth genes), and since different cancers rely on it differently, blocking HCF1 could be a new way to treat cancer.

In short: HCF1 is not just a simple doorstop. It is a sophisticated matchmaker that carefully selects partners, checks their details, and then introduces them to a "sugar-coating" machine to supercharge their ability to control life and death in our cells.

Technical Summary

1. Problem Statement

Host Cell Factor 1 (HCF1) is a transcriptional co-regulator discovered over 30 years ago, known to interact with proteins containing a consensus D/EHxY (or N/QHxY) sequence via its Kelch domain. Despite its established role in viral infection, cell cycle regulation, metabolism, and as a cancer dependency (particularly in MYC-driven cancers), several critical gaps remain:

• Incomplete Interactome: The full diversity of HCF1-binding partners is unknown, and it is unclear why only a subset of proteins containing the consensus sequence actually bind HCF1.
• Molecular Determinants: The specific molecular rules governing binding affinity beyond the core "anchor" residues (D/E, H, Y) are poorly defined.
• Mechanistic Output: How binding affinity translates to downstream transcriptional activity and how HCF1 orchestrates post-translational modifications (specifically O-GlcNAcylation) remain elusive.

2. Methodology

The authors employed a multi-faceted approach combining high-throughput screening, deep mutational scanning (DMS), structural modeling, and functional assays:

• Proteome-Wide Peptide Display: They constructed bacterial surface display libraries containing 347 peptides with the D/EHxY consensus and 259 with the N/QHxY consensus (derived from intrinsically disordered regions of the human proteome). These were screened against the HCF1 Kelch domain fused to EGFP using Flow Cytometry (FACS) and deep sequencing to determine enrichment scores.
• Deep Mutational Scanning (DMS): To map molecular determinants beyond the consensus, they created 11-mer peptide libraries based on four known binders (KANSL3, DIDO1, SETD1A, UL48). Every position (from H-5 to Y+3) was mutated to all 20 amino acids to quantify the impact on binding affinity.
• Structural Modeling: AlphaFold3 was used to model HCF1-peptide complexes to visualize solvent accessibility and residue contacts, guiding the interpretation of DMS data.
• Biophysical Validation: Fluorescence Polarization (FP) assays were used to measure dissociation constants ($K_d$) for wild-type and mutant peptides.
• Functional Assays:
  • Peptide Pulldowns: Validated binding of candidate peptides to endogenous HCF1 in cell lysates.
  • RNA-seq & Cell Cycle Analysis: Investigated the transcriptional and phenotypic consequences of HCF1 binding using IRF1 variants (Wild-type, Non-binding, and High-affinity grafted) in MDA-MB-231 cells.
  • O-GlcNAc Co-IP + Mass Spectrometry: Used a competitive peptide strategy (expressing a high-affinity HCF1 binder) to disrupt endogenous HCF1 interactions, followed by anti-O-GlcNAc immunoprecipitation and proteomics to identify substrates dependent on HCF1 for O-GlcNAcylation.

3. Key Contributions & Results

A. Expansion of the HCF1 Interactome

• Subset Specificity: The screen revealed that only 66 out of 347 D/EHxY peptides are bona fide binders. This indicates that the consensus sequence is necessary but insufficient; additional context is required.
• Novel Binders: The study identified 41 previously uncharacterized HCF1 binders, including transcription factors (e.g., MXD3, CREB3L1) and chromatin regulators.
• Non-Canonical Motifs: The authors discovered a register-shifted, non-canonical motif: DHxxY (two residues between H and Y). They identified SMCHD1 and IRF1 as binders using this motif. AlphaFold3 modeling confirmed that the anchor residues (D, H, Y) align similarly to canonical binders, but the intermediate "x" region is extended.

B. Elucidation of Extended Molecular Determinants

• Beyond Anchors: DMS revealed that residues outside the core D/EHxY motif significantly dictate binding affinity.
  • Flanking Residues: Positions H-4, H-1, Y+1, and Y+2 show strong selectivity. For example, bulky hydrophobic residues are favored at H-4, while proline is strongly disfavored at Y+2.
  • Intermediate 'x' Residue: In canonical binders, small residues are often preferred in the 'x' position to avoid steric hindrance.
  • Context Dependence: The tolerance for mutations varies by peptide background. For instance, the H-1 position prefers Aspartate (D) over Glutamate (E) in some contexts, and the presence of E at H-1 can impose coupled constraints on other positions (epistasis).
• Predictive Modeling: Using Position-Specific Scoring Matrices (PSSM) derived from DMS, the authors successfully predicted binding scores for natural peptides. They demonstrated that "sub-optimal" non-anchor residues in the E2F1 peptide explained its weak binding, and mutating these residues enhanced E2F1 binding to levels comparable to high-affinity binders like KANSL3.

C. Affinity-Dependent Transcriptional Regulation

• IRF1 Case Study: The authors investigated the non-canonical DHxxY motif in the transcription factor IRF1.
  • Binding Affinity: The IRF1 DHxxY peptide binds HCF1 with a $K_d$ of ~4.8 $\mu$M, weaker than canonical binders (e.g., KANSL3 at ~0.1 $\mu$M) but sufficient for interaction.
  • Functional Impact: Replacing the native IRF1 motif with a high-affinity KANSL3-derived motif (IRF1-HB) resulted in:
    • Enhanced anti-proliferative effects and G1 cell cycle arrest.
    • A broader transcriptional response, including upregulation of interferon signaling genes and TNF$\alpha$/NF-$\kappa$B signaling pathways, which were not fully activated by the wild-type IRF1.
  • Conclusion: HCF1 binding affinity acts as a tunable rheostat for the transcriptional output of its partners.

D. HCF1 as a Scaffold for O-GlcNAcylation

• Mechanism: The study confirms that HCF1 acts as a recruitment scaffold for the O-GlcNAc transferase (OGT).
• Global Effect: Upon disrupting HCF1 interactions (via competitor peptide expression), 39 proteins showed reduced O-GlcNAcylation.
• Specificity: The majority of these substrates (17/39) contained HCF1-binding motifs, and 36/39 were nuclear proteins. This suggests HCF1 directs OGT to specific nuclear targets, including histone modifiers and transcription regulators, thereby linking HCF1 binding directly to the O-GlcNAc modification landscape.

4. Significance

• Redefining Binding Rules: The paper moves beyond the simple "consensus sequence" model, establishing that HCF1 binding is governed by an extended code involving flanking residues and specific intermediate constraints. This explains why many consensus-containing proteins do not bind HCF1.
• Discovery of Non-Canonical Motifs: The identification of the DHxxY register-shifted motif expands the known repertoire of HCF1 interactors and suggests a mechanism for recognizing diverse protein structures.
• Affinity-Function Coupling: The study provides a mechanistic link between binding affinity and biological output, showing that subtle changes in the HCF1-binding interface can drastically alter transcriptional programs (e.g., NF-$\kappa$B activation).
• Therapeutic Implications: By characterizing HCF1 as a master regulator of O-GlcNAcylation and a cancer dependency, the work reinforces the Kelch domain as a viable therapeutic target. Understanding the precise molecular determinants allows for the design of more specific inhibitors that could disrupt oncogenic HCF1 interactions (e.g., with MYC) while sparing essential functions.

In summary, this work provides a comprehensive, quantitative, and structural framework for understanding how HCF1 selects its partners, how binding affinity dictates function, and how it orchestrates a critical post-translational modification network in the nucleus.