2A Peptide Discovery

Systematic identification of viral 2A self-cleaving peptides and prokaryotic ribosome stalling sequences

Overview

This project identifies 2A peptides across protein sequence databases using profile hidden Markov models (HMMs). 2A peptides are short (~15-20 residue) cis-acting oligopeptides that cause ribosomal skipping during translation at conserved Gly-Pro (GP) dipeptides.

What are 2A Peptides?

2A peptides were first discovered in picornaviruses (foot-and-mouth disease virus) where they enable production of multiple proteins from a single open reading frame without requiring a stop codon. The ribosome “skips” peptide bond formation at the Gly-Pro junction, releasing the upstream protein while continuing translation of the downstream sequence.

NoteKey Features of 2A Peptides

• Conserved C-terminal motif: DxExNPGP
• Self-cleavage occurs between Gly and Pro
• No external protease required
• Widely used in biotechnology for multi-cistronic expression

Two Classes of 2A Peptides

We distinguish two classes based on conserved sequence features:

Class	N-terminal	Central Motif	C-terminal
Class 1	Leucine stretch	GDVE	NPGP
Class 2	Invariant Trp	EEGIE	PNPGP/PHPGP

Pipeline Overview

flowchart TB
    subgraph inputs[" "]
        direction TB
        Seeds["🧬 Seed Alignments<br/><small>Curated 2A sequences</small>"]
        style Seeds fill:#e1f5fe
    end

    subgraph databases[" "]
        direction TB
        UniProt[(UniProt)]
        RefProt[(RefProt)]
        UniParc[(UniParc)]
        MGnify[(MGnify)]
        IMGVR[(IMG/VR)]
        style UniProt fill:#fff3e0
        style RefProt fill:#fff3e0
        style UniParc fill:#fff3e0
        style MGnify fill:#fff3e0
        style IMGVR fill:#fff3e0
    end

    subgraph iteration1["Iteration 1"]
        direction TB
        HMM1["hmmbuild<br/><small>Seed HMMs</small>"]
        Search1["hmmsearch<br/><small>All databases</small>"]
        Filter1["Filter<br/><small>E-value, length</small>"]
        style HMM1 fill:#e8f5e9
        style Search1 fill:#e8f5e9
        style Filter1 fill:#e8f5e9
    end

    subgraph iteration2["Iteration 2"]
        direction TB
        HMM2["hmmbuild<br/><small>Refined HMMs</small>"]
        Search2["hmmsearch<br/><small>Expanded search</small>"]
        Curate["Auto-curate<br/><small>Motif validation</small>"]
        style HMM2 fill:#fce4ec
        style Search2 fill:#fce4ec
        style Curate fill:#fce4ec
    end

    subgraph outputs["Outputs"]
        direction TB
        FinalHMM["📊 Final Models<br/><small>Class 1 & 2 HMMs</small>"]
        Logos["🎨 Sequence Logos"]
        Report["📝 Analysis Report"]
        style FinalHMM fill:#f3e5f5
        style Logos fill:#f3e5f5
        style Report fill:#f3e5f5
    end

    Seeds --> HMM1
    HMM1 --> Search1
    databases --> Search1
    Search1 --> Filter1
    Filter1 --> HMM2
    HMM2 --> Search2
    databases --> Search2
    Search2 --> Curate
    Curate --> FinalHMM
    FinalHMM --> Logos
    FinalHMM --> Report

Figure 1: Complete workflow for 2A peptide and ribosomal stalling peptide discovery

Databases Searched

Eukaryotic Pipeline

Database	Description	Size
UniProt Swiss-Prot	Curated protein sequences	~570k
Reference Proteomes	Representative proteomes	Medium
UniParc	Non-redundant archive	Very large
MGnify	Metagenomic proteins	~300M

Prokaryotic Pipeline

Database	Description	Size
UniProt Bacteria	Swiss-Prot bacterial proteins	~500k
UniProt Archaea	Swiss-Prot archaeal proteins	~50k
NCBI Viral RefSeq	Viral protein sequences	~100MB
IMG/VR	Phage proteins	Millions
INPHARED	Curated phage genomes	~19k genomes

Key Analyses

Eukaryotic 2A Peptides

• Seed Search Results: Initial HMM search statistics
• Class Comparison: Differences between Class 1 and Class 2
• Sequence Logos: Visualization of conserved motifs

Prokaryotic Discovery

• Discovery Pipeline: Approach for finding novel stalling peptides
• GP Motif Analysis: Comprehensive GP context analysis
• Known Stalling Peptides: Validation against SecM, TnaC, etc.

Methods

• Pipeline Architecture: Snakemake workflow details
• HMM Approach: Profile HMM methodology
• Database Sources: Data provenance and URLs

Background

Ribosomal Stalling in Prokaryotes

Prokaryotes use related but distinct mechanisms for ribosomal stalling:

• SecM: Stalls at FXXXXWIXXXXGIRAGP to regulate secretion
• TnaC: Tryptophan-dependent stalling
• MifM: Regulates membrane protein insertion in B. subtilis

These peptides share the GP motif but have different upstream contexts and biological functions.

Biotechnology Applications

2A peptides are widely used for:

• Multi-gene expression from single vectors
• Stoichiometric co-expression of protein complexes
• CAR-T cell engineering
• Vaccine development

Resources

• Source Code: GitHub repository
• Seed Alignments: resources/seed-alignments/
• Final Models: results/models/final/

TipRunning the Pipeline

# Full pipeline
pixi run snakemake --cores 12

# Prokaryotic discovery only
pixi run snakemake prokaryotic_discovery --cores 12