Group Final Project

Bacteriophage Engineering Group Project Inputs_William & Mary Node Group ^{1 2}

Group Project_Protein Design 1

• Selected Goal: Increased stability (easiest)

• Brainstorm Session Questions:

  • Which tools/approaches from recitation you propose using (e.g., “Use Protein Language Models to do in silico mutagenesis, then AlphaFold-Multimer to check complexes.”)
    • We’ll attempt to run multi-environment/conditional modeling and simulation to down-select lysis stability approaches that show the greatest resilience across environments/conditions.
    • The team has selected a project focused on enhancing the stability of the Lysis Protein, a decision influenced by the group’s current experience level. The primary objective is to improve thermodynamic stability while concurrently preserving the native protein fold and maintaining functional integrity. The proposed methodology involves utilizing BLAST for identifying homologous sequences, followed by Clustal Omega to ascertain conserved residues susceptible to mutation intolerance. Subsequently, ESM2 will be employed to score candidate substitutions based on evolutionary plausibility. This will be succeeded by the application of ESM-Fold to predict and refine the integrity of the protein fold, as well as to optimize existing backbones. The results may then be further subjected to EvolvePro for accelerated directed evolution. Tools like Boltz-1 and ProteinMPNN offer a capability for redesigning solvent-exposed residues and optimizing the core packing of the protein. We can cross their performance for comparison. All selected variant candidates are slated for computational stress-testing under a range of environmental conditions that could potentially induce destabilization. Selected variant candidates that pass the stress test are prioritized for downstream experimental validation.
  • Why do you think those tools might help solve your chosen sub-problem?
    • The previous bullet point addresses tool functionality in our workflow, explaining why and how various tools will assist us in accomplishing our goal
  • Name one or two potential pitfalls (e.g., “We lack enough training data on phage–bacteria interactions.”).
    • One potential pitfall is that we may have insufficient in vitro quality and quantity of data to test the environmental constraints of interest. Thus wet-lab work would be needed to back-up the findings, in addition to follow-up
    • There are open questions regarding the validity of the stated research approach (i.e., if the approach makes sense relative to the larger goal of increased stability)
  • Include a schematic of your pipeline
    • See workflow schematic below
      • 

Group Project_Protein Design 2 ³

1. See results below

2. Notebook Inputs:

• Inputted L-Protein Mutants.csv file for analysis
  • Noted Encounter Issues and Resolutions:
    • Part 12 involved a step wherein the cell would be run and there would exist a prompt by which an upload of the “L-Protein Mutants - Sheet1.csv” was required. The current “L-Protein Mutants” file offered in the laboratory was listed as an excel file. We proceeded by saving the data into a .csv file that reflects the desired file name and uploading.

3. Utilized experimental data [4]

4. Experimental Correlation Results

• The results below indicate poor correlation between the experimental data and the scores from the notebook. The values indicate high levels of uncertainty between the experimental data and the notebook scores. These results indicate that:
  • 

5. 5 Mutations

• A Heatmap of predicted mutations was given. The following mutations are chosen in coordinate form. While yellow indicates a high ratio, it shouldn’t be taken fully at face value.
  • Xavier Results
    • 
    • Chose the using the Predicted Effects Graph (allows for a first pass)
      • Soluble-Region: Y39 (L), F5 (S)
      • Transmembrane Region: N53, E61, T52 (All L)
  • Raphael Results
    • The following are mutations propositions based on the Predicted Effects Graph:
      • Soluble Region:
        • F5Q: Clusal Omega shows that the 5 position is optimal for change as it is frequently switched. ESM-2 deems this mutation at this position to have the most promising score (1.795244).
        • F5P: Clustal Omega shows 5 position as frequently changing. ESM-2 deems this mutation as third best for position 5 (1.596891).
      • Transmembrane Region:
        • E61L: Clustal Omega shows high variance at position 61. ESM-2 demonstrates a high score of 1.818098.
        • F47L: Clustal Omega shows high variance at this position, showing the protein is most likely to retain functionality after a mutation here. ESM-2 demonstrates a decent positive score.
      • Free:
        • F5S, K50L, E61L: F5S mutation is supported by the Clustal Omega results as it is shown across multiple species. The K50L and E61L mutations are not supported as much by Clustal Omega but show high results in ESM-2.
  • Jason Results
    • 
    • Soluble Region:
      • F5_R: Potentially reduces protein aggregation risk, increases solubility via Arginine inclusion
      • S9_Q: Potential glutamine inclusion could lead to more stable hydrogen bonds
      • C29_R: Current cysteine prone to potential misfolding. Arginine could improve solubility, potentially preventing clumping or misshapen configurations
    • Transmembrane Region:
      • L44_I: Increased isoleucine density might increase thermodynamic stability
      • A62_V: Valine hydrophobicity might help protein orientation and improve lipid piercing
  • Nana Results
    • Soluble Region:
      • Position 5 : F -> Q ( 1.79524445533752)
      • Position 17: N -> R (1.32365107536315)
    • Transmembrane Region:
      • Position : 40 V -> L ( 1.79524445533752)
      • Position 50 K -> L (2.56146419048309)
      • Position 65 R -> L (1.0260357856750488)
    • Mutants:
      • Mutant 1: METRQPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
      • Mutant 2: METRFPQQSQQTPASTRRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
      • Mutant 3: METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYLLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT
      • Mutant 4: METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLEAVIRTVTTLQQLLT
      • Mutant 5: METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVILTVTTLQQLLT

6. Generated Multimeric Assemblies
   • Xavier Results
     • Oligomer: METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLLVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFLLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTLQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLLAVIRTVTTLQQLLT
     • Chain Separated Based on:

       • Chain A: F5S: position 5, F → S

       • Chain B: Y39L: position 39, Y → L

       • Chain C: T52L: position 52, T → L

       • Chain D: N53L: position 53, N → L

       • Chain E: E61L: position 61, E → L

       • METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT

       • METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLLVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT

       • METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFLLQLLLSLLEAVIRTVTTLQQLLT

       • METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTLQLLLSLLEAVIRTVTTLQQLLT

       • METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLLAVIRTVTTLQQLLT

     • Sequence Coverage Map
       • 
     • Predictions
       • 
     • Display 3D Structure
       • 
     • Plots
       • 
   • Raphael Results
     • F5Q:
       • 
       • 
     • F5P:
       • 
       • 
     • E61L:
       • 
       • 
     • F47L:
       • 
       • 
     • F5S, K50L, E61L:
       • 
       • 
         • RESULTS: Mutation 5 (F5S, K50L, E61L) produced very promising results. It seems that the soluble region cannot be improved (most likely due to its disordered nature which prevents it from forming a tangible final structure). But when the K at position 61 is changed into an L, the transmembrane region improves greatly. The stats for this rank are the following: pLDDT=62.4 pTM=0.565 ipTM=0.558.
           • Multimeric Assembly:
             • METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLLAVIRTVTTLQQLLT:METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLLAVIRTVTTLQQLLT:METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLLAVIRTVTTLQQLLT:METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLLAVIRTVTTLQQLLT:METRSPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLLAVIRTVTTLQQLLT
   • Jason Results
     • Original Wild-Type Sequence 3D visualization and Sequence Coverage Map
       • 
       • 
     • Multimeric Assembly based on mutations listed in 5.
       • METRRPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQQQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPRRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFIAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLVEVIRTVTTLQQLLT
         • Multimeric Assembly Sequence 3D visualization and Sequence Coverage Map
           • 
             • Unfortunately it appears that my introduced mutations did not improve the lysis abilities of the L protein based on the results above
   • Nana Results
     • Multimeric Assembly: METRQPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTRRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYLLIFLAIFLSKFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSLFTNQLLLSLLEAVIRTVTTLQQLLT:METRFPQQSQQTPASTNRRRPFKHEDYPCRRQQRSSTLYVLIFLAIFLSKFTNQLLLSLLEAVILTVTTLQQLLT

All of Jason’s supporting prompts for this work listed below