De novo acyl carrier proteins display structure-independent modification and sequence novelty

This study demonstrates that a hybrid computational-experimental approach using the ALGO-CP algorithm can generate diverse, soluble acyl carrier protein variants that lack the canonical helical structure yet retain full post-translational modifiability, revealing that key ACP functions can operate independently of their traditional structural fold.

The Gist

Imagine a cell as a bustling, high-tech factory. Inside this factory, there are tiny, essential workers called Acyl Carrier Proteins (ACPs). Think of these ACPs as molecular delivery trucks. Their job is to pick up raw materials (fatty acids), drive them to different assembly lines (enzymes), and drop them off to build vital products like fats, vitamins, and even medicines.

For these trucks to work, they need a specific "hook" (a chemical tag) attached to them. Without this hook, they are just empty shells. Once the hook is attached, they become "holo-ACPs," ready to grab cargo and deliver it.

The Big Question

Scientists have known about these trucks for a long time. They usually look like a specific shape: a tight bundle of four springs (helices). For years, engineers tried to build new trucks by tweaking the existing blueprints or swapping parts between different models. But they were stuck in a small corner of the design world. They wondered: Is that specific springy shape absolutely necessary for the truck to work, or can we build a delivery vehicle that looks completely different but still does the job?

The New Tool: ALGO-CP

To answer this, the researchers built a new computer program called ALGO-CP. Think of this program as a creative chef who knows the rules of cooking (chemistry) but isn't bound by traditional recipes (evolution).

Instead of just copying old recipes, the chef mixes two approaches:

1. The Historian: Looks at thousands of old recipes to see what ingredients are always used (like the essential hook).
2. The Innovator: Looks at the properties of ingredients (taste, texture, weight) and tries to invent new combinations that have never been seen before, as long as they make sense chemically.

By mixing these two approaches, the computer generated thousands of brand-new "truck designs" that had never existed in nature.

The Experiment: Building the Unlikely

The team picked a few of these weird, computer-generated designs and tried to build them in a lab. They were looking for two things:

1. Solubility: Did the truck fall apart in the water?
2. Modifiability: Could the factory attach the essential "hook" to it?

The Surprise:
They found two designs, ALGO-055 and ALGO-059, that worked! But here is the twist:

• The Shape: When they looked at these proteins under a microscope (using a technique called Circular Dichroism), they realized these trucks did not have the classic springy shape. They were floppy, messy, and unstructured—like a pile of spaghetti rather than a neat bundle of springs.
• The Function: Despite looking like a mess, the factory enzymes could still attach the essential hook to them. Even better, once the truck grabbed its cargo (a fatty acid), the spaghetti snapped into a neat, springy shape.

The Analogy: The Origami Truck

Imagine you have a piece of paper.

• Natural ACPs: You fold the paper into a specific, rigid crane shape. It's sturdy and works perfectly.
• ALGO-055/059: You leave the paper crumpled in a ball. It looks useless. But, if you attach a heavy weight (the cargo) to it, the paper magically unfolds and folds itself into a working crane.

The researchers found that the "cargo" acts like a molecular chaperone, forcing the messy protein to organize itself into a working shape.

Why Does This Matter?

This discovery changes how we think about biology and engineering:

1. Structure isn't everything: You don't need the perfect, rigid shape to start the process. The "hook" can be attached to a floppy, disordered protein.
2. Evolutionary Insight: It suggests that in the ancient past, these delivery trucks might have started as messy, floppy blobs. Only later, as the factory got more complex and needed to be more precise, did they evolve into the rigid, springy shapes we see today. The cargo itself might have been the force that taught them how to fold.
3. Future Engineering: This gives scientists a new playground. We can now design proteins that are totally different from anything nature has made, opening doors to new medicines and bio-factories that work in ways we never imagined.

The Bottom Line

The team proved that you can build a functional biological machine from scratch using a computer, even if it looks nothing like the original. Sometimes, the "messy" designs work just as well as the "perfect" ones, provided they have the right ingredients to grab onto their cargo. It's a reminder that in the world of biology, function can often emerge from chaos.

Technical Summary

1. Problem Statement

Acyl Carrier Proteins (ACPs) are essential, dynamic proteins that shuttle metabolic substrates in primary and secondary metabolism. While naturally occurring ACPs share a conserved four-helix bundle structure, they exhibit significant sequence diversity (often <30% identity). Previous engineering efforts have been limited to rational mutagenesis or "helix swaps" between natural homologues, exploring only a narrow fraction of the potential ACP design space.

The central challenge addressed by this work is: Can ACPs be designed de novo with sequences that deviate significantly from natural homologues, potentially lacking the canonical $\alpha$-helical fold, yet still retain the critical ability to undergo post-translational modification (PTM) and function? Current computational design tools often rely heavily on phylogenetic constraints, limiting the discovery of novel, non-canonical functional proteins.

2. Methodology

The authors developed a hybrid computational-experimental approach centered on a bespoke sequence-generation algorithm called ALGO-CP.

• Algorithm Design (ALGO-CP):
  • Input: A multiple sequence alignment (MSA) of 2,167 E. coli AcpP homologues.
  • Dual-Strategy Approach: The algorithm combines two sub-routes:
    1. Route-AC (Evolutionary): Samples amino acids based strictly on observed frequencies in the MSA (conservative).
    2. Route-AP (Physicochemical): Samples amino acids based on matching local physicochemical properties (isoelectric point, Kyte–Doolittle hydropathy, van der Waals volume) to the MSA block, allowing for rare amino acids not seen in nature but chemically compatible.
  • Tuning: A weighting coefficient ($r$) blends these routes ($r=0$ is purely evolutionary; $r=1$ is purely physicochemical). The authors identified $r=0.60$ as the optimal balance for generating diverse yet plausible sequences.
• Experimental Workflow:
  • Selection: From 5,000 generated sequences, candidates were screened for solubility and PTM capability.
  • PTM Pathway: Recombinant enzymes were used to convert apo-ACPs to holo-ACPs (via 4'-phosphopantetheinyl transferases: BsSfp or EcAcpS) and subsequently to acyl-ACPs (via fatty acyl-ACP synthetase: VhAasS).
  • Characterization:
    • Mass Spectrometry (LC/ESI-MS): To confirm PTM (mass shifts of +340 Da for holo, +182 Da for acyl).
    • Circular Dichroism (CD) Spectroscopy: To assess secondary structure ($\alpha$-helical content) in apo, holo, and acyl states.
    • Molecular Dynamics (MD) Simulations: 1 $\mu$s simulations using GROMACS to analyze stability and conformational dynamics of the designed proteins compared to wild-type E. coli AcpP (EcAcpP).
    • AlphaFold3: Used for initial structural prediction.

3. Key Contributions

• Development of ALGO-CP: A tunable algorithm that successfully explores the "dark matter" of protein sequence space by relaxing evolutionary constraints while maintaining physicochemical viability.
• Discovery of Structure-Independent Function: The identification of de novo ACPs that are fully functional (modifiable) despite lacking the canonical $\alpha$-helical fold in their apo/holo states.
• Chimera Generation: Successful creation of chimeric variants from the initial de novo designs, further expanding the functional sequence space.
• Insight into ACP Evolution: Evidence suggesting that the canonical ACP fold may have evolved as a consequence of cargo engagement (acylation) rather than being a strict prerequisite for the initial PTM event.

4. Key Results

• Successful De Novo Design: Two sequences, ALGO-055 and ALGO-059, were expressed as soluble proteins. Both underwent complete PTM:
  • Apo $\to$ Holo: Successfully modified by EcAcpS (and partially by BsSfp).
  • Holo $\to$ Acyl: Successfully loaded with a lauroyl (C12) chain by VhAasS (>99% efficiency).
• Chimeric Variants: Two chimeras, chALGO-012 and chALGO-024, derived from ALGO-055/059, also exhibited full modifiability and solubility.
• Structural Paradox (The "Surprise"):
  • In Silico: AlphaFold3 and MD simulations predicted these variants would adopt a canonical 4-helix bundle fold.
  • In Vitro (CD Spectroscopy): Contrary to predictions, ALGO-055 and ALGO-059 lacked the canonical $\alpha$-helical signature in their apo and holo states. They appeared as minimally structured, non-helical ensembles.
  • Structural Recovery: Upon acylation (binding the C12 chain), both variants showed a marked increase in helicity, indicating that the acyl cargo acts as a structural chaperone, inducing folding.
• Sequence Analysis:
  • The successful variants contained numerous "rare" amino acid variations (present in <1% of natural homologues) and non-conservative substitutions (high Grantham distances).
  • Despite these deviations, they preserved key acidic "hotspots" required for electrostatic interactions with modifying enzymes, explaining their functional success.
• MD Simulation Insights: Simulations revealed that the 4'-PP prosthetic group induces backbone destabilization in the holo state, which is partially reversed upon acylation. ALGO-055 showed superior stability compared to ALGO-059, which exhibited significant helical loss.

5. Significance

• Redefining Structure-Function Relationships: This work challenges the dogma that the canonical ACP fold is strictly necessary for PTM. It demonstrates that sequence novelty and specific electrostatic features can drive function even in the absence of a pre-formed tertiary structure.
• Evolutionary Implications: The findings suggest a potential evolutionary trajectory where early ACPs might have existed as disordered, modifiable precursors, with the stable $\alpha$-helical fold emerging later to enhance substrate sequestration and partner selectivity.
• Engineering Platform: ALGO-CP provides a robust framework for generating novel protein lineages. It can be adapted to design ACPs for non-natural substrates, fungal polyketide biosynthesis, or mitochondrial engineering.
• Limitations of Prediction Tools: The study highlights a critical gap in current structural prediction tools (like AlphaFold), which tend to over-predict $\alpha$-helicity in de novo sequences, likely due to training biases toward known folded structures.

In conclusion, the authors successfully expanded the ACP design space beyond natural limits, proving that functional ACPs can be generated with sequences that defy canonical structural expectations, provided key interaction motifs are preserved.