Design, Screening, and Testing of Non-Rational Peptide Libraries with Antimicrobial Activity: In Silico and Experimental Approaches

Abstract

One of the challenges of modern biotechnology is to find new routes to mitigate the resistance to conventional antibiotics. Antimicrobial peptides (AMPs) are an alternative type of biomolecules, naturally present in a wide variety of organisms, with the capacity to overcome the current microorganism resistance threat. Here, we reviewed our recent efforts to develop a new library of non-rationally produced AMPs that relies on bacterial genome inherent diversity and compared it with rationally designed libraries. Our approach is based on a four-stage workflow process that incorporates the interplay of recent developments in four major emerging technologies: artificial intelligence, molecular dynamics, surface-display in microorganisms, and microfluidics. Implementing this framework is challenging because to obtain reliable results, the in silico algorithms to search for candidate AMPs need to overcome issues of the state-of-the-art approaches that limit the possibilities for multi-space data distribution analyses in extremely large databases. We expect to tackle this challenge by using a recently developed classification algorithm based on deep learning models that rely on convolutional layers and gated recurrent units. This will be complemented by carefully tailored molecular dynamics simulations to elucidate specific interactions with lipid bilayers. Candidate AMPs will be recombinantly-expressed on the surface of microorganisms for further screening via different droplet-based microfluidic-based strategies to identify AMPs with the desired lytic abilities. We believe that the proposed approach opens opportunities for searching and screening bioactive peptides for other applications.

Keywords: antibiotic resistance; antimicrobial peptides; deep learning; library screening; microfluidics; molecular dynamics; non-rational; rational.

Figure 1

Antimicrobial Peptides (AMPs) discovery framework. Rational design steps: (I) Deep learning techniques identify sequences with potential antimicrobial activity, (II) membrane-disruption capabilities of selected sequences are analyzed via molecular dynamics (MD), (III) the host cell is modified, and sequences are inserted, finally (IV) antimicrobial activity is corroborated by a microfluidic system. Non-rational design steps: (I) Random sequences are expressed on host cells through cell surface display, (II) modified microorganisms are analyzed by a microfluidics system to obtain AMPs candidates, and (III) DNA is extracted, sequenced, and cloned (Created with BioRender).

Figure 2

Representation of a traditional membrane-protein system used in molecular dynamics. The interaction among the components is modeled through force fields that account for variations in key parameters and impose restrictions on the accessible states (Created with BioRender).

Figure 3

Recurrent neural networks used for peptide property prediction. A representation of each of the amino acids is input to one unit of the recurrent network. The last unit’s output contains information considering the output of the previous ones, therefore considering the whole sequence. On the left a long-short term memory networks (LSTM) unit representation and on the right a gated recurrent units (GRU) unit representation. (Created with BioRender).

Figure 4

Scheme of the different microbial cell surface display methods. The phage display method is shown on the left, while the bacterial display process is presented in the center, and the yeast display is presented on the right. In all cases, through molecular biology tools it is possible to express the protein fragments of interest quite robustly for further biomolecular interaction analysis and screening (Created with BioRender).

Figure 5

Schematic representation of a classic molecular dynamics simulation (MD) process. Initially, a preparation stage is required in which the system is assembled. Subsequently, the position restraints are turned off to run the MD simulation, and finally, the data of the trajectories are obtained and analyzed (Created with BioRender).

Figure 6

Potential mean force curve calculated from an umbrella sampling simulation across a phospholipid bilayer model where the zero represents the center of mass (COM) of a 6 nm thick membrane (Created with BioRender).

Figure 7

Droplet-based microfluidics screening strategies. (A) Schematic of screening performance used by Yaginuma where cells and peptide secreting yeast are encapsulated in a droplet system. Droplets with fluorescence due to the secretion of LacZ are sorted and recovered for yeast culture. Finally, peptides secreted are sequenced to obtain functional candidates. (B) Schematic of screening performance used by Guo where a droplet library is generated using a microfluidic approach. The antimicrobial activity analysis is carried out injecting the droplets in a microfluidic platform where single microbial cells are added and incubated. Finally, the droplet-encapsulated compounds are screened and sorted based on the microbial growth arrest (Created with BioRender).

Figure 8

Membrane-based approaches for bioactive compounds screening. (A) Surface plasmon resonance (SPR) interaction analysis carried out by Hall et al. and Šakanovič et al. where a lipid planar bilayer is generated in the surface of a sensor chip in which the peptide membrane interactions are measured by surface plasmon resonance [231,232]; (B) Membrane-based screening using giant unilamellar vesicles (GUVs) as described by Nahas et al. where GUVs are generated by the octanol-assisted liposomes assembly method (OLA) technique and immobilized in a chamber where the peptides are injected in order to analyze the membranolytic activity by the leak of a dye previously encapsulated into the GUVs lumen. (C) Vesicle screening platform used by Kuhn where the GUVs are immobilized onto a glass-bottom and a delivery of the tetracycline is carried out to analyze the drug permeation using a total internal reflection fluorescence (TIRF) microscopy due to the red fluorescence complex generated by Europium-tetracycline binding.(Created with BioRender).

Figure 9

Schematic of combinatory microarray screening performance used by Zhao where possible permutations of the combinatorial peptide library are obtained using different combinatorial flow patterns of amino acids applied to a blank microdisk array. (Created with BioRender).

Figure 10

Schematic of combinatory microarray screening performance used by Du et al. where droplets formed in a Nanowell chip by the interaction of cell suspension and oil are exposed for changes in media and schedules dosage-dependent drug assays to perform stimulation–response studies based on the fluorometric analysis in a high-throughput manner. (Created with BioRender).