Team:Paris Bettencourt/Design
Design
In 2016, Lam et al., described SNAPPS (Structurally Nanoengineered Antimicrobial Peptide Polymers). SNAPPs are a unique class of star-shaped antimicrobial peptides with activity against a range of Gram-negative pathogens.
| Box 1: What are SNAPPs? |
|---|
| SNAPPS are built from short chains of lysine and valine, 10-30 amino acids long. These peptide arms are conjugated at one end to a core of poly(amidoamine). Each core may be linked to 16-32 separate arms. The result is positively charged protein star, about 20 nm across. |
The goal of this project is to replace chemically synthesized SNAPPs with fully synthetic biological fusion proteins. We reasoned that the key features of SNAPP geometry could be reproduced by fusing short, positively charged AMPs to the N- and C- termini of highly multimeric proteins. In this way, the naturally multimeric protein subunits form the core of the structure and the fused AMPs extend outward to form the arms of the star-shape. We called the resulting proteins StarCores.
Architectural Principles of Star Cores
Following Shirbin (2018), we identified some key features needed for effective StarCores.
1) Arm number between 4-32. If each monomer of protein is fused to an AMP at both ends, then we want to use proteins that form 2- to 16-mer homomultimeric complexes.
2) Star size smaller than 30 nm. SNAPPs are most effective when they are dense, presenting many AMPs to the membrane.
3) Strong Positive Charge. Bacterial membranes are negatively charged and the SNAPP interaction and pore formation is charge-mediated.
Combinatorial Design of StarCores
Protein engineering is challenging. A good synthetic protein needs to express efficiently, fold tightly, avoid self-aggregation and then function with high activity. We knew it was unlikely that we would design a perfect StarCore on the first try. Therefore, we decided to design many StarCores and screen them systematically.
First, we selected 15 known AMPs to function as the arms of the StarCore. The purpose of these AMPs is to functionally substitute the lysine-valine chains in the original SNAPP design. Therefore, we chose AMPs with a high density of positive charges, reasoning that they would interact with bacterial membranes by a similar mechanism. Our second criterion for AMP selection was simply diversity. We chose natural AMPs from many organisms as well as synthetic AMPs.
| Peptide Name | Class | Source | Activity |
|---|---|---|---|
| Cg-Def | Defensin | Crassostrea gigas (Oyster) | Antibacterial, Antifungal |
| Alyteserin-2a | Alyteserin | Alytes obstetricans (Toad) | Antibacterial, Cytotoxicity |
| Pardaxin P-1 | Pardaxin | Pardachirus pavoninus (Fish) | Cytotoxic |
| ɑ-1-purothionin | Thionin | Triticum aestivum (Wheat) | Cytotoxic |
| Arenicin-1 | Arenicin | Arenicola marina (Lugworm) | Antibacterial, Antifungal, Cytotoxic |
| Ovispirin-1 | Sheep-derived | Synthetic | Antibacterial, Cytotoxic |
| V6 peptide | Cyclic cationic | Synthetic | Antibacterial |
| PolyVK-11 | Poly-VK | Synthetic | Antibacterial |
| PolyVK-12 | Poly-VK | Synthetic | Antibacterial |
| PolyVK-21 | Poly-VK | Synthetic | Antibacterial |
| Bactofencin A | IId | Lactobacillus | Antibacterial |
| Aureocin A5 | IId | Staphylococcus | Antibacterial |
| Bacteriocin E50-52 | IIa | Enterococcus | Antibacterial |
| Enterocin A | IIa | Enterococcus | Antibacterial |
| Pediocin PA-1 | IIa | Pediococcus | Antibacterial |
Next, we selected 14 proteins to use as the core of the StarCore. There are many potential multimeric proteins found in nature. We filtered them by a variety of criteria, favoring convenience and diversity (Box 2).
For each core protein, we used the PDB structure to determine if the N- or C-termini, or both, were free to accept protein fusions. We only constructed fusion proteins using ends that were determined to be free.
| Box 2: Criteria for Selecting Core Proteins |
|---|
|
- The monomer must be expressible in E. coli: - The monomer must self-assemble into a homomultimeric complex. - The complex must have a size between 4-25 nm. - The complex must have known structure deposited in the PDB. - The N- and/or C- termini of the monomers should be free. - The set of complexes should be of diverse shapes. |
Following these criteria, we selected the following proteins from RCSB Protein data bank.
| NAME | Ferritin | MS2 | Qbeta | Small HSP |
|---|---|---|---|---|
| PDB reference | 4XGS | 7msf | 5vlz | 1SHS |
| Antimicrobial peptide fused to | N-terminal | N-terminal C-terminal | C-terminal | C-terminal |
| Image | 4XGS | 7MSF | 5vlz | 1SHS |
| NAME | ConcavalinA | Pyruvate dehydrogenase | Designed protein lyase cage | Isoaspartyl Dipeptidase |
|---|---|---|---|---|
| PDB reference | 2g4i | 1EAA | 4QCC | 5LP3 |
| Antimicrobial peptide fused to | N-terminal C-terminal | N-terminal | N-terminal | N-terminal |
| Image | 2g4i | 1EAA | 4QCC | 5LP3 |
| NAME | GTP_Cyclo-hydrolase | FucU isomerase | Pizza_3 | Holo- synthase |
|---|---|---|---|---|
| PDB reference | 1GTP | 2WCV | 3WW8 | 5xu7 |
| Antimicrobial peptide fused to | N-terminal | C-terminal | N-terminal C-terminal | N-terminal C-terminal |
| Image | 1GTP | 2WCV | 3WW8 | 5xu7 |
| NAME | Pizza_2 | RADA self-assembling peptide |
|---|---|---|
| PDB reference | 3WW7 | – |
| Antimicrobial peptide fused to | N-terminal C-terminal | N-terminal C-terminal |
| Image | 3WW7 | – |
In total, we designed 210 StarCore proteins as the combinatorial fusion of 15 AMPs to 14 cores.
Results
Construction of the StarCore Fusion Proteins
StarCore proteins were constructed as translational fusions using GoldenGate assembly (see methods). Of 210 designed proteins, we were able to successfully clone 131 constructs (Table 3).
We expected the cloning of our constructs to be difficult because they encode highly toxic products. Our constructs were designed to be expressed using the T7 polymerase, which is absent in cloning strains of E. coli. Nevertheless, some of the constructs may give leaky expression, killing the host.
Expression of the StarCore Fusion Proteins
StarCore fusion proteins were expressed using the myTXTL Sigma 70 Master Mix Kit, generously provided by our team sponsor, Arbor Biosciences. As an expression vector, we used pACYCDuet-1 from Novagen. This vector is widely used for protein production in strains of E. coli that express T7 polymerase such as BL21 (DE3). It contains a T7 promoter upstream of a strong RBS and a lac operator, allowing IPTG-controllable protein expression.
To express from the T7 promoter, it was necessary to first produce T7 polymerase in the cell-free extract. For this purpose, we used the plasmid P70a-T7rnap, supplied by the manufacturer. We also included 100 uM IPTG in the master mix, to relieve lac repression.
Cell-free extracts were assayed for the presence of StarCore proteins by a variety of methods, described in the Characterization page. Unfortunately, none of these assays produced evidence of successful protein expression.
Commercial Sourcing of StarCore Fusion Proteins
Fortunately, we had arranged an alternate source for StarCore proteins. The Bioneer company generously offered to sponsor us by giving us access to their ExiProgen™automated protein synthesis platform. This is a fully automated system that takes synthetic DNA as input and performs cell-free expression and protein purification.
Like us, Bioneer found most of the StarCore constructs to be difficult to clone, express and purify. However, thanks to their efforts we were able to obtain 11 StarCore proteins at high yield.
Methods
T7 Expression Vector
StarCores were cloned into the pACYCDuet-1 Vector (Novagen). The cloning strategy made use of the manufacturer-suggested start codon, downstream of a T7 promoter. The final construct omitted protein purification tags included in the original vector.
For cloning, the plasmid was first linearized by digestion with AvrII and NcoI. Then we performed PCR on the linearized vector with Golden Gate cloning sites added as primer tags.
DNA Synthesis
Coding sequences for AMPs and core proteins were generously synthesized by team sponsor IDT (Coralville, USA). Codon usage was optimized by the supplier for expression in E. coli. Sequences were free of BioBrick restriction enzyme cut sites, as well as common type IIS restriction sites.
Golden Gate cloning sites were added to the ends of each coding sequence by PCR with tagged primers.
Golden Gate Cloning
We created translational fusions of the AMP peptides to the core proteins using standard Golden Gate cloning. Golden Gate Assembly Mix was supplied by NEB (#E1600S). Cloning reactions made use of the the BsaI enzyme and 4 custom cloning overhangs.
The AMP, core and plasmid backbones were assembled in one step.
In cases where the N-terminus of the core protein could not be tagged, the core sequence was amplified with GG1 and the N-terminal AMP was omitted. The C-terminus was treated similarly. No additional amino acids were added to the coding sequences of core termini deemed unsuitable for tagging.
Transformation and Construct Validation
Golden Gate asseumbly products were transformed into E. coli DH5a. Because this strain lacks T7 polymerase, it does not express the potentially antimicrobial StarCore peptides.
Successful clones were verified by colony PCR.
paris-bettencourt-2018@cri-paris.org 
