Development of the VersaTile technique and a microdroplet assay for the ultrahigh-throughput engineering of lysins as antibacterial agents

The discovery and development of antibiotics have shaped modern medicine. However, our health care system is increasingly threatened by antimicrobial resistance. Despite warnings since the very beginning, overuse and misuse of antibiotics have led to the emergence and spread of antimicrobial resistance against every known antibiotic. Moreover, many pharmaceutical companies have abandoned the search and development of new antibiotics due to often disappointing and financially unsustainable screening campaigns.

This combination of the ever increasing antimicrobial resistance and the failure to develop new antibiotics, forced the World Health Organization (WHO) in 2014 to the bold statement that the world is on the brink of a “post-antibiotic era” if no action is undertaken. Currently, the Centers for Disease Control and Prevention are even stating that we are already living in a post-antibiotic era. Reports are indicating that if no new treatments are found, the situation could lead to the death of 10 million people each year worldwide, the disappearance of important agriculture races and a cumulative worldwide economic loss between $ 2.1 and $ 124.5 trillion by 2050. The urgent need to find new antibiotics has led to the revival of old treatments and has flared up the search for novel approaches.

Lysins are peptidoglycan hydrolases encoded by bacteriophages. Since 2001, lysins have been intensively studied for their use as antibacterials. Their mode-of-action is based on induction of osmotic lysis upon active peptidoglycan degradation. The first lysin is currently evaluated in a Phase III clinical trial against bacteremia and endocarditis caused by Gram-positive multidrug-resistant Staphylococcus aureus strains. However, the WHO list of critical priority pathogens for which novel antibiotics are urgently needed is dominated by Gram-negative pathogens. Unfortunately, most natural lysins show minor or no antibacterial activity against Gram-negative bacteria due to the presence of an outer membrane that acts as a shield. By fusing an outer membrane permeabilizing peptide to a lysin, the bulky lysin is able to pass through the outer membrane and to kill Gram-negative  bacteria through peptidoglycan degradation. The modular composition of lysins that may comprise modules that bind to the bacterial cell wall and modules that degrade the peptidoglycan, allows to customize the properties of lysins by recombining modules of different lysins along with diverse peptides. This property is called the modularity principle.

Yet, the full potential of the modularity principle remained inaccessible due to the lack of a method that allows combinatorial shuffling of non-homologous modules at a large scale. Therefore, we have created VersaTile, a new DNA assembly method. This method is based on the intrinsic ability of type IIs restriction enzymes to cut outside of their recognition sequence. Combined with the careful selection of position tags, the method allows to do combinatorial DNA shuffling at an unprecedented scale. Nanopore sequencing demonstrated that VersaTile is unbiased towards assembling large libraries of engineered lysins. To cope with the high complexity of the assembled libraries, we implemented an iterative hit-to-lead development process using deepwell plate protein expression and a parallel growth inhibitory assay. A subset of variants was analyzed and design rules were extracted from the initial top hits. The VersaTile technique was used again to construct a smaller library based on the implementation of these design rules. Analysis of this new library led to more and improved lysin variants that were eventually screened under more stringent conditions to identify a lead variant. This iterative approach led to the discovery of a new engineered lysin 1D10 that proved to be active in human serum against multidrug-resistant Acinetobacter baumannii strains. In addition, lysin 1D10 slows down bacterial growth and biofilm formation in an ex vivo pig skin explant burn wound model.

As lysin 1D10 was the only active engineered lysin that came out of the previously mentioned screen, we aimed to get more insight into its unique characteristics. By using time-lapse microscopy a new, dual mode-of-action was revealed. Instead of explosive lysis as is seen with other (engineered) lysins, lysin 1D10 makes local holes, often at the septum or poles, through which the intracellular content leaks out. Via elucidating the role of each module in lysin 1D10, it was shown that (i) a linker is necessary for the antibacterial activity, (ii) a CBD is not essential but improves the antibacterial activity, (iii) the EAD is responsible for the high thermoresistance of 1D10, (iv) the dual mode-of-action of lysin 1D10 is due to the partially retained antibacterial activity of Cecropin A and the cell wall degradation of the EAD. Finally, it was shown that 1D10 and its derivatives do not trigger cytotoxicity on a HaCaT epithelial cell line.

With the deepwell plate assay, we were only able to screen 4 % out of a library of approximately 10,000 variants, which is only the tip of the iceberg. Therefore, a new screening assay had be developed to fully unlock the potential of the VersaTile platform. We developed an assay based on the ultimate miniaturization of a bioreactor: a microdroplet. Using water-in-oil droplets generated by microfluidic chips, we built a microfluidic assay that is capable of screening libraries of potentially millions of engineered lysins. The assay starts with the encapsulation of a library of engineered lysins, together with a rolling circle based amplification mixture, in water-in-oil droplets. After amplifying the DNA, an in vitro transcription/translation mixture will be added to express the engineered lysin. Thereafter, the bacterium of interest will be added together with a substrate to detect esterase activity. If the bacteria are lysed by the engineered lysin, they will release intracellular esterases. These esterases will generate a fluorescent signal that we will be used by fluorescence activated droplet sorting in order to separate the active variants out of the pool of negative ones. Finally, after DNA extraction, purification and debranching, the sequence of the engineered lysins will be determined using Nanopore sequencing. We proved that the microfluidic assay is functional by screening a positive control, the native lysin LysRODI, and a library of 15,000 variants against Staphylococcus aureus SA9.