MACS: Modular Activated Cancer Suppressor

Single-mRNA-Encoded Multi-Domain Protein System with Acidic Microenvironment-Responsive Logic Gating

Selman Ali Dokumacı

Independent Computational Biophysics Project · 2026

Current Stage: Computational Discovery (In Silico). Experimental validation pending.

Contents

The Selectivity Barrier in Solid Oncology

Highly potent modalities (ADCs, T-cell engagers, cytokine fusions) often fail due to severe dose-limiting toxicities in normal organs expressing baseline target levels. Single-target constitutively active biologics are vulnerable to target downregulation, antigen loss, and stromal barriers in NSCLC and PDAC.

MACS addresses this by ensuring the biotherapeutic remains shielded in healthy circulation (pH 7.4) and activates selectively within the acidic tumor milieu (pH <= 6.8).

System Architecture

System architecture components
Component	Definition
Chain Definition	947 amino acids encoded sequentially.
Module A (1-587)	Effector and structural scaffold components.
Module B (588-947)	pH-responsive shielding module.
Translational Advantage	Single mRNA construct delivery potential.

Design Hypothesis & Scope

The architecture leverages specific domain pairings designed to conditionally occlude functional interfaces.

Note: While designed as a generalized platform, computational validation currently applies only to the specific prototype construct modeled here. Broad modularity remains a design hypothesis requiring diverse empirical testing.

Mechanistic Hypothesis: Acidosis-Driven Quaternary Gating

Proposed states by pH condition
Condition	Proposed state
pH 7.4 — Proposed quiescent state	Histidine residues neutral (HIE). Intermodular latch disengaged. Module B flexible, extended. Effector interface proposed to remain shielded.
pH 6.8 — Proposed activated state	Protonation of calibrated histidine clusters (HIP). Module B compaction (+29.5 internal contacts). Candidate unmasking of effector binding register.

The simulations support a distributed, multiresidue pH-response model rather than a single-residue mechanism. Strongest convergence: B293 alternative microenvironments, B324 pKa-coupled remodeling.

Caveat: This is formulated as a structural hypothesis supported by all-atom MD simulations. Biological activation requires experimental testing.

Computational Evidence Stack

Conformational & MD
- All-Atom MD via GROMACS 2026.1
- 300 ns per pH condition (600 ns cumulative)
- Fixed-protonation approximation (specific residues modeled)
- Conformational descriptors
- Contact topology
Dynamics & Networks
- Elastic Network Models (ANM/GNM)
- 20-mode mobility profiles
- Dynamic Cross-Correlation Matrices (DCCM) with Louvain clustering
- Network analysis
- Sensor exposure mapping
Energetics & Titration
- Ensemble PROPKA 3.4 (101-frame ensembles)
- PB-MMPBSA: 1001-frame endpoint binding energy

Note: All analyses derive from fixed-protonation classical MD. Frame-level uncertainties are within-trajectory distributions, not independent replicate-level statistics.

Key Structural Findings

Module B Compaction

Figure 1 — Trajectory-averaged conformations (pH 6.8 Module A light blue, Module B blue; pH 7.4 Module A dark gray, Module B orange; histidines purple, hotspots yellow)

Table 1 — Module B conformational descriptors by pH condition
Metric	pH 6.8	pH 7.4	Observation
Internal Contacts	778.09 ± 13.89	748.63 ± 14.51	+29.46 contacts
Ca Rg (nm)	2.6284 ± 0.0159	2.6481 ± 0.0186	Slight contraction
Asphericity (nm²)	1.4226 ± 0.0717	1.6061 ± 0.1141	More spherical at pH 6.8
Kappa-squared	0.0989 ± 0.0035	0.0625 ± 0.0070	Anisotropy change

Source: moduleB_full_vs_last100ns.tsv (Data from Module B, last 100 ns)

Caveat: Single primary trajectory per condition. Values describe within-trajectory distributions, not replicate-level uncertainty.

Visual Evidence

Conformational Configurations & Register States

Figure 2 — Representative interface configuration — pH 6.8

Figure 3 — Representative interface configuration — pH 7.4

Representative structures selected from the analyzed MD trajectories. Conformational differences are computational observations and do not constitute experimental validation.

Figure 4 — Representative interface-register state — pH 6.8

Figure 5 — Representative interface-register state — pH 7.4

Residue Network Topology & Correlation Mapping

Figure 6 — Protonation Topology: Histidine network overlay showing pH-dependent reorganization of protonation-sensitive residue clusters.

Figure 7 — Allosteric Relay: Dynamic cross-correlation difference map (pH 6.8 minus pH 7.4). Reveals condition-dependent changes in residue-residue correlated motion.

Limitations & Planned Validation

Current Limitations

Most analyses originate from one 300 ns trajectory per pH condition
Fixed-protonation classical MD approximates pH effects through assigned protonation states
PROPKA is an empirical structure-based estimator, not constant-pH MD
ANM is a coarse-grained normal-mode approach
PB-MMPBSA excludes configurational entropy
No experimental validation exists in the current dataset

Next Steps

Planned independent MD replicas: 4 per condition (8 total; 2.4–4.8 μs cumulative)
Construct synthesis and expression in HEK293 / E. coli
pH-dependent CD spectroscopy (folding stability)
SPR/BLI binding kinetics (pH 6.8 vs 7.4)
Selectivity in spheroid models (NSCLC & PDAC)

Important Note on Scope

The following claims are NOT supported by current data: that the pH-switch is experimentally proven; that a single histidine controls the entire mechanism; that Module B completely closes at pH 6.8; that PB-MMPBSA difference constitutes replicate-level statistical proof.

Research Documentation

Technical documentation, evidence summaries, and methodological records supporting the MACS computational research program.

Year	Title	Document type	Access
2026	Mechanistic Evidence Summary	Technical Report	Read
2026	Mechanistic Evidence Matrix	Technical Documentation	Read
2026	Methodological Decision Log	Technical Documentation	Read
2026	MDP Consistency Report	Technical Documentation	Read