Individual Final Project

#MycoTint: Stress Responsive Mycelium

MycoTint is the conceptualization, genetic engineering, and biofabrication of reactive, living mycelium furniture capable of displaying visible color changes in response to mechanical stress. The objective is to engineer a mechanosensitive synthetic genetic circuit within the fungal host that actuates the expression of vivid, naked-eye-visible chromoproteins when specific points of the furniture are subjected to weight or impact. This document details the selection of the optimal biological chassis, the intricate design of the synthetic genetic circuit, fungal transformation protocols, specialized biofabrication and long-term viability maintenance strategies, practical feasibility assessments, alternative non-biological methods, and an analysis of precedent projects shaping the ELM landscape.

This project addresses this barrier by engineering a reactive mycelium composite that functions as an autonomous, visual mechanosensor. The overall objective is to genetically modify the high-density basidiomycete Ganoderma lucidum to express naked-eye-visible chromoproteins when subjected to mechanical stress. It can be utilized for load-bearing applications and receive real-time, non-destructive assessment of their structural integrity. This also doubles up as an interactive design feature that can be utilized in mycelium furniture and products where the user triggers the colour expression by sitting/interacting with it.

The different methods I see using to achieve this result are:

Mechanosensitive Genetic Circuits: Filamentous fungi like Ganoderma lucidum can be genetically engineered to hijack their innate cell wall integrity pathway, triggering the biological expression of visible anthozoan chromoproteins specifically when mechanical pressure is applied.

Bacterial Co-cultivation: Structural fungal mycelium can be grown alongside genetically engineered bacterial consortia that act as localized sensors, detecting environmental or physical cues to produce fluorescent reporters or color changes within the composite.

Engineered Melanin Synthesis: Fungal strains such as Aspergillus niger can be genetically modified to place melanin synthesis genes under the control of specific inducible promoters, causing the living material to display distinct color intensities when exposed to specific chemical triggers.

Embedding cell-free protein synthesis (CFPS) machinery and DNA encoding chromoproteins within hydrogels or polymers integrated throughout an inert mycelium scaffold. Upon mechanical stress, embedded microcapsules can release water or specific chemical triggers to rehydrate and activate the CFPS system, initiating the in vitro synthesis of visible pigments without the need to maintain living fungal cells.

The ideal organism must exhibit vigorous vegetative growth on low-cost lignocellulosic substrates, possess high mechanical compressive strength, resist environmental contamination, and, crucially, be highly amenable to genetic modification. Two white-rot basidiomycetes dominate the current landscape of mycelium biofabrication and serve as the primary candidates for this application: Pleurotus ostreatus (the oyster mushroom) and Ganoderma lucidum (the reishi mushroom).

Group Final Project

from Part D — Group Brainstorm on Bacteriophage Engineering

(Individual submission — solo student)

1. Project Goal

The primary goal of this project is to increase the structural stability of the MS2 bacteriophage lysis protein (L-protein), with a secondary goal of reducing its dependency on the host chaperone DnaJ, while preserving its capacity to lyse bacterial cells through membrane pore formation.

The MS2 L-protein is a 75-residue single-gene lysis toxin. Its architecture divides cleanly into two functional regions:

• Soluble N-terminal domain (residues 1–40): intrinsically disordered, interacts with DnaJ, and is responsible for chaperone-dependent folding and activation
• Transmembrane C-terminal domain (residues 41–75): forms a hydrophobic helix that inserts into the inner bacterial membrane, drives oligomerization into pore complexes, and executes lysis

A key E. coli resistance mechanism is a single point mutation in DnaJ (P330Q) that prevents it from interacting with the L-protein, blocking lysis. Engineering the L-protein to fold and function without DnaJ would directly circumvent this resistance route. Since the lytic activity resides in the transmembrane domain not the soluble domain that DnaJ binds. There is a credible path to separating folding assistance from lytic function through targeted mutagenesis of the N-terminal region.

The engineering strategy therefore focuses on three things simultaneously:

• Stabilizing the soluble domain so it folds autonomously without DnaJ
• Maintaining the transmembrane helix integrity for membrane insertion and pore formation
• Preserving the conserved L48–S49 dipeptide motif and neighboring residues that are essential for function

2. Computational Tools and Approaches

A multi-step computational pipeline combining sequence analysis, protein language model mutagenesis, and structural prediction will be used.

2.1 BLAST — Homolog Discovery

BLAST is used first to find homologous lysis proteins from related bacteriophages across sequence databases.

Purpose:

• Identify which positions across the protein are evolutionarily conserved vs. variable
• Collect natural sequence diversity for multiple sequence alignment
• Understand which parts of the L-protein have tolerated substitutions in nature, giving prior evidence that those positions can be mutated without destroying function

The BLAST results feed directly into the next step.

2.2 Clustal Omega — Multiple Sequence Alignment (MSA)

Homologous sequences retrieved from BLAST are aligned using Clustal Omega.

Purpose:

• Map fully conserved positions (* in the alignment) — these must not be mutated
• Identify partially conserved positions (:) where only similar-chemistry substitutions are tolerated
• Confirm that the L48–S49 motif and surrounding residues are conserved, protecting them from mutagenesis

A key finding from the MSA of MS2 L-protein homologs is that all conserved positions cluster in the soluble domain (residues 1–40), specifically at positions 21, 25, 28–29, 33, 35–37, and 40. This is biologically meaningful these positions likely form the DnaJ-binding epitope and the structural core of the soluble domain. The transmembrane region (41–75) is less conserved, making it more accessible for hydrophobicity-enhancing substitutions.

2.3 ESM Protein Language Models — In Silico Deep Mutational Scan

The ESM2 protein language model is used to generate a log-likelihood ratio (LLR) score for every possible single-point substitution at every position in the L-protein.

Purpose:

• Produce a mutation heatmap across the full 75-residue sequence
• Identify substitutions the model predicts as tolerated or stabilizing (positive LLR) vs. harmful (negative LLR)
• Guide rational mutation selection rather than random or intuition-based choices

Importantly, LLR scores reflect evolutionary plausibility and structural stability — they do not directly predict lytic function. Cross-referencing against the experimental lysis dataset (Chamakura et al., 2017) is therefore essential to exclude mutations that score well computationally but have been shown to abolish lysis in the wet lab.

2.4 ESMFold — Structure Prediction for Candidate Mutants

Promising mutations identified from the ESM scan are input into ESMFold to predict the 3D structure of the mutant L-protein monomer.

Purpose:

• Assess predicted confidence (pLDDT) of the mutant structure vs. wild-type
• Confirm the transmembrane helix remains intact in the TM-domain mutants
• Identify mutations that significantly distort the backbone and discard them

A known limitation here is that ESMFold, like most structure predictors, performs less well on small intrinsically disordered proteins like the L-protein soluble domain. Low pLDDT scores in the N-terminal region may reflect genuine disorder rather than bad mutations — this ambiguity is a recognized pitfall of the approach.

2.5 AlphaFold Multimer — Oligomerization and DnaJ Interaction

AlphaFold Multimer is used for two separate runs per mutant:

Run A — 8-mer pore assembly: Eight copies of the mutant L-protein are submitted as separate chains to test whether the protein retains the capacity to oligomerize into the cylinder-like transmembrane pore that drives lysis.

Run B — DnaJ co-fold: The mutant L-protein is submitted alongside the DnaJ sequence to assess whether soluble-domain mutations reduce the predicted interaction interface between the two proteins.

A key insight from the reference implementation is that all five designed mutants, as well as a known experimentally validated lytic mutant (R30Q), returned very low pLDDT scores (<50) and low-confidence PAE plots for inter-chain contacts. This confirms a systematic limitation of AlphaFold for this class of small membrane-disrupting proteins — low confidence does not rule out functional lysis activity. All five mutants remain viable candidates for wet lab validation.

3. Proposed Engineering Pipeline

Wild-type L-protein sequence (75 aa)
         ↓
    BLAST search
    (find homologous lysis proteins)
         ↓
  Clustal Omega MSA
  (conserved vs. mutable positions)
         ↓
  ESM2 deep mutational scan
  (LLR heatmap across all positions)
         ↓
  Cross-reference with experimental
  lysis dataset (Chamakura et al.)
         ↓
  Select candidate mutations
  ┌──────────────────────────────┐
  │ 2 in soluble domain (1–40)  │
  │ 2 in TM domain (41–75)      │
  │ 1 anywhere (highest LLR)    │
  └──────────────────────────────┘
         ↓
  ESMFold — monomer structure
  (pLDDT check, helix integrity)
         ↓
  AlphaFold Multimer
  ┌──────────────┬──────────────┐
  │ 8-mer pore   │ DnaJ co-fold │
  │ (retained?)  │ (disrupted?) │
  └──────────────┴──────────────┘
         ↓
  Final 5 mutant candidates
  → Submit for synthesis (Twist)

4. Chosen Mutations

Based on the pipeline above, the following five mutations were selected:

All five avoid the fully conserved positions (21, 25, 28–29, 33, 35–37, 40) and the three mutations that appeared on both the ESM heatmap and the experimental sheet with a lysis score of zero.

5. Expected Outcomes

The engineered variants are expected to produce:

• Increased intrinsic structural stability in the soluble domain, particularly for Y39L and S9Q, reducing dependence on DnaJ for folding
• Improved membrane insertion kinetics for K50L, N53L, and T52L, by replacing polar/charged residues with leucine in the hydrophobic TM helix, potentially producing faster or more efficient lysis
• Retention of the pore-forming oligomeric assembly, since the transmembrane domain is not disrupted at the conserved functional core
• A DnaJ-independent folding pathway in the best-case scenario for the soluble-domain mutants, enabling the phage to overcome the P330Q DnaJ resistance mutation in E. coli

6. Potential Pitfalls

6.1 Limited training data for phage proteins ESM2 and ESMFold are trained predominantly on globular, well-characterized proteins. Short transmembrane phage toxins like the MS2 L-protein are under-represented in training data. This likely reduces prediction accuracy and may explain why even experimentally validated lytic mutants return low pLDDT and PAE scores from AlphaFold Multimer.

6.2 LLR scores predict stability, not function The ESM heatmap captures evolutionary plausibility and structural fitness, not lytic activity. Three mutations that had high LLR scores were found to abolish lysis completely in the experimental dataset. This confirms that computational stability predictions must always be cross-referenced against functional data — a lesson that the broader field of computational protein design is still learning to internalize.

6.3 Risk of over-stabilization Mutations that rigidify the soluble domain too much could prevent the conformational changes needed for membrane insertion or DnaJ dissociation. A protein that is too stable may be non-functional even if it folds correctly.

6.4 Poor annotation of amurin-class proteins Single-gene lysis proteins (amurins) are a poorly annotated class. Homolog discovery via BLAST retrieves relatively few high-quality sequences, which limits the power of the MSA for identifying truly conserved vs. mutable positions.

6.5 Host protease sensitivity New surface-exposed residues created by the soluble-domain mutations may accidentally introduce protease cleavage sites, reducing the effective concentration of functional L-protein inside infected bacteria and blunting lytic efficacy.

7. Literature Summaries

MS2 Lysis of E. coli Depends on Host Chaperone DnaJ (Chamakura et al., 2017) This study demonstrates that the L-protein requires the host chaperone DnaJ for efficient lysis. A single missense mutation (P330Q) in DnaJ’s C-terminal domain blocks L-mediated lysis at 30°C, establishing the mechanistic basis of the resistance strategy this project aims to overcome. Genetic suppressor screening found that truncated L-proteins lacking the basic N-terminal domain can bypass DnaJ entirely, directly motivating the idea of engineering the soluble domain to achieve chaperone independence.

Mutational Analysis of the MS2 Lysis Protein L (Chamakura & Young, 2018) Comprehensive random mutagenesis of all 75 residues showed that most loss-of-function mutations cluster in the C-terminal half, particularly around the conserved L48–S49 dipeptide. Many inactivating mutations were conservative substitutions that still allowed protein accumulation and membrane association, suggesting that lysis depends on specific protein–protein interactions rather than nonspecific membrane disruption. This explains why ESM structural scores are insufficient predictors of lytic activity — function is more sensitive than stability.

In Vitro Characterization of the Phage Lysis Protein MS2-L (Arulandu et al., 2023) This study shows that MS2-L assembles into high-order oligomeric complexes (≥10 monomers) after insertion into lipid nanodiscs, driven primarily by the transmembrane domain. DnaJ interacts with the N-terminal domain but is not required for membrane insertion or oligomerization itself, suggesting its role is primarily as a folding or stability partner. This supports the feasibility of engineering DnaJ-independent variants — if the TM domain can self-insert and oligomerize, then eliminating DnaJ dependence through N-terminal modifications should not impair pore formation.

Phage Therapy: From Biological Mechanisms to Future Directions (Gordillo Altamirano & Barr, 2023) This review surveys therapeutic phage applications and engineering strategies. It highlights that phage resistance — including via host factor mutations — remains a central challenge, and that engineered phages with modified lysis proteins represent a promising avenue for overcoming bacterial adaptation. The L-protein engineering effort directly addresses one of the most common and fastest-arising resistance mechanisms identified in clinical phage therapy.

8. Future Wet Lab Validation Steps

If promising computational mutants are identified, the following experimental steps would be required before drawing biological conclusions:

• Chemical synthesis of the mutant L-protein gene via Twist Bioscience
• Cloning into an expression plasmid using Gibson Assembly
• Expression in wild-type and DnaJ-mutant (P330Q) E. coli strains
• Plaque assays to measure lysis activity and compare to wild-type L-protein
• Western blot to confirm protein accumulation levels are not affected by the mutations
• Thermal shift assays (DSF) to directly measure whether the soluble-domain mutants show higher melting temperatures, confirming computational stability predictions