Team:Freiburg/Description
Inspiration
It would be great if we could state that there are just a few problems to solve on our planet and therefore choosing our project was straightforward. Unfortunately, instead there are so many issues: climate change, food and energy shortage, pollution, and innumerable diseases…
Hard to decide on a project which would have the strongest impact on society! Inspired by previous iGEM teams, we realized that for us it was important to find a project that, despite having a specific application, would be modular, interdisciplinary and broad enough to fundamentally advance synthetic biology beyond a particular problem. Our team consisting of biology, molecular medicine, chemistry, engineering and informatics students, chose to tackle the toxicity of methicillin-resistant Staphylococcus aureus (MRSA), aware that the most important feature of our project would be the original way of solving this issue which should be beneficial also if applied in other contexts.
Thanks to the input we received from several experts and from reading the scientific literature, we discovered the fascinating world of D-amino acids, so far gone unnoticed within the iGEM community.
Description
Engineering proteins to enhance their activity or make them acquire new desired properties is a major goal of synthetic biology. Most approaches limit themselves to the 20 canonical L-amino acids. However their stereochemical counterparts, D-amino acids, harbor an immense potential. When assembled into peptides those cannot be recognized by the cellular machineries, thus evading proteolytic breakdown and immunological recognition.
This makes them perfect candidates for therapeutics. By establishing a multitude of tools we empower D-amino acids for synthetic biology. We demonstrated the potency of mirror-image phage display by identifying D-ligands towards a toxin of multiresistant Staphylococcus aureus. We created finDr, a software to perform this method in silico for any target enabling fast, cost-effective prediction of D-ligands. Alongside chemical synthesis, we implemented methods to synthesize, incorporate and detect D-amino acids in bacteria.
Antimicrobial resistance is a major threat to global health
The appearance of bacteria resistant against several antibiotics is a growing problem, which we are causing ourselves by the extensive use of antibiotics. (https://amr-review.org/home.html)
Fig. 1: If we do not find a solution against antimicrobial resistance, by 2050 every three seconds someone will die because of incurable bacterial infections.
Despite this deteriorating situation, many pharmaceutical companies have reduced their investment in the research of new antibiotics.
We recognize that the antibiotics crisis is a major challenge of our time and we decided to apply our mirror-image phage display strategy to peptide therapeutic identification on a target which is relevant in this context.
Staphylococcus aureus is one of the most notorious multiresistant pathogens and over the years a variety of strains have emerged that show resistance against every single known antibiotic (Vestergaard M. et al., Microbiol. Spectr. 2019). Infections with normal, antibiotic sensitive Staphylococci are very common in hospitals and cause severe diseases such as sepsis or toxic shock syndrome.
PSMα3 - a two-faced MRSA toxin
The phenol soluble modulin (PSM) family of toxins consists of 8 members which are increasingly expressed in the emerging strains of community-acquired MRSA, determining their enhanced virulence. The most toxic member of this family is PSMα3.
The mechanism of action of PSMα3 is just as two-faced as its molecular architecture. PSMα3 is a short helical protein of 22 amino acids, with a hydrophobic and a hydrophilic side opposing each other. Due to this amphipathic nature, PSMα3 can insert itself into biological membranes and form pores that lead to their disruption. The consequence is the lysis of cells and their organelles. The lytic properties of PSMα3 can help the bacteria to get rid of immune cells like T cells or neutrophils. Furthermore, PSMα3-secreting bacteria can even escape the phagosomes of macrophages thanks to this lytic mechanism. In addition to its lytic properties, PSMα3 is also a strong inflammatory stimulus acting upon the N-formyl peptide receptor 2 (FPR2) on neutrophils. Binding to FPR2 causes neutrophils to move towards the source of PSMα3 and to secrete inflammatory cytokines and reactive oxygen species, thereby aggravating the symptoms of the MRSA infection. (Cheung et al., PLoS Path, 2010). Furthermore, PSMα3 can form amyloid fibers at high concentrations, however their significance in disease is still under discussion (Tayeb-Fligelman et al., bioRxiv, 2019). Because of its potency and its diverse harmful effects on the human body, aggravating MRSA infection and promoting its spreading, we decided that targeting PSMα3 might prove a reasonable strategy to counteract MRSA infections. Therefore we performed a mirror-image phage display to identify a selective D-peptide ligand for PSMα3 and thereby demonstrating the potential of mirror-image phage display.
D-Amino Acids and Chemical Synthesis
Peptides are highly diverse, uncomplicated to develop and modify and can form specific interactions with many different target structures. This makes them great candidates for peptide therapeutics; however, they are very susceptible to proteolytic breakdown. Thus, their short half-life in vivo limits their application as therapeutics and remains an obstacle to overcome.
The usage of D-amino acids is an elegant way to evade the poor bioavailability of peptide drugs. Due to their mirrored stereochemical orientation, the peptide bonds between D-amino acids cannot be broken by natural proteases making D-peptides resistant to proteolytic degradation. This leads, in turn, to low antigen presentation of D-peptide fragments, which makes them immunologically inert, another favorable property for therapeutics.
Although biological expression systems are reliable, fast and easy to implement, they have a crucial limitation - they fail to express proteins made entirely of non-canonical amino acids. This limitation applies especially for D-amino acids, which are intrinsically foreign to the translational machinery of all organisms on this planet. Overcoming this limitation given by the homochirality of biological systems, we established a chemical method to synthesize D-PSMα3 in our lab. This was based on solid phase peptide synthesis (SPPS). The method can be carried out by hand or automated.
In SPPS, unlike in biological systems, the peptides are elongated beginning from the C-terminus in sequential reaction cycles. The peptide is anchored on a chemically unreactive and insoluble polystyrol resin which is equipped with a bifunctional linker coupled to the carboxy-group of the first amino acid. After completion of synthesis the resin can be cleaved at this residue to obtain the crude product.
Mirror-Image Phage Display
Nowadays, various peptide-selection approaches are available for the identification of putative ligands. Three of the most commonly used ones are the ribosome display, in vitro compartmentalisation and phage display. We chose phage display because it is the best established method and because it does not depend on the rather unstable RNA as used in ribosome display or on high amounts of purified proteins, as would be the case for in vitro compartmentalisation. Furthermore the potential of phage display has been underlined by the selection of the Nobel prize committee in 2018 for chemistry. Phage display relies on a random library of peptides that are fused to the coat proteins of bacteriophages via short flexible linkers. This ensures the exposure of the peptide on the phage surface. Billions of individual phage clones can be screened against a target protein. Sequential rounds of selection and enrichment of target-binding phage clones lead to the identification of peptide ligands with high affinity to the target protein.
Fig. 2: Description missing.
However any in vivo system is limited by the biological functions the system provides. Because the phage amplification takes place in E. coli , it is difficult to incorporate non-canonical amino acids in a phage peptide library.
As an extension of the classical phage display, mirror-image phage display (MIPD) was developed to identify not only L- but also D-peptide ligands. MIPD is based on the principle that an occurring interaction between a D-peptide ligand and an L-target protein would also exist between the mirror-image L-peptide ligand and the D-version of the target protein.
Consequently, in order to identify a pair of interacting ligand/protein, MIPD does not use the naturally occurring L-form of a target protein and a library of D-peptide ligands, but rather their mirror image: a chemically synthesized D-target protein of choice and a phage library of L-peptide ligands. Once an L-peptide ligand is identified, the mirror plane can be inverted, and, as a result, the D-peptide ligands bind to the naturally occurring L-target.
MIPD is a versatile method that can be applied to a wide variety of targets from different fields of research. We aimed to utilize the power of this innovative method to identify therapeutic D-peptide ligands to tackle the current crisis of antibiotic resistance. However, the workflow used here can be easily extended.
Modelling and software
MIPD is based on random libraries for the identification of D-peptides that bind specifically to a chosen target. As a consequence of this de novo approach, the resulting peptide ligands have to be characterized extensively as there is nothing known about their properties apart from their amino acid sequence and the fact that they bind to the target molecule. However, the mechanism of this binding and its location on the target molecules surface remain to be elucidated. We approached this with a computational model of PSMα3 and its ligands. In order to model the mirror image D-version of PSMα3, based on its L-structure, we inverted the X-coordinate of every atom in the protein structure, thereby mirroring the stereochemical orientation of the protein. To obtain a faithful model of our protein, we performed molecular dynamics simulations of PSMα3 using GROMACS. Our model can be used to predict the functional properties of the ligands from MIPD and their effect on PSMα3 cytotoxicity.
We wanted to expand our computational work, beyond the mere modelling of our wetlab MIPD results. Therefore, we created the software tools to perform an in silico equivalent of MIPD enabling the virtual screening of D-peptide ligands against any target of choice. This high-throughput approach overcomes the limitations of wet-lab MIPD. We exploited the possibilities of parallelization in computational simulations and implemented the principles of Darwinian evolution in a genetic algorithm for the de novo creation and optimization of D-peptides for various purposes.
Incorporation of D-amino acids into proteins
The incorporation of non-canonical amino acids into proteins has been established in vitro and in vivo demonstrating improvements of several desired characteristics, such as brightness or photostability of fluorescent proteins. While in vitro there have been many studies on the acceptance of D-AAs by ribosomes, very little has been done so far to specifically in vivo incorporate D-AAs into proteins. Ma and colleagues were the first to show that the in vivo incorporation of a single D-Phenylalanine at residue 66 inside the chromophore of the superfolder green fluorescent protein (sfGFP) resulted in a red-shift for excitation and emission maxima and an improvement of the thermal stability of the protein.
To enable D-amino acid incorporation into proteins produced in living cells, we have created a modification of an orthogonal translation system used in genetic code expansion. Our approach is based on the polyspecific aminoacyl-tRNA synthetase from Methanocaldococcus jannaschii which we speculate could incorporate a multitude of D-amino acids into proteins in response to the amber stop codon.
Fig. 3: To improve yield and specificity, we have combined two strategies: a) the use of a genetically engineered bacterial strain where all amber stop codons have been mutated and Release Factor 1 has been deleted (Lajoie et al., 2012); and b) a compartmentalization strategy recently developed for mammalian cells, whereby physical proximity and encapsulation of the orthogonal translation machinery is obtained using the MCP-MS2 loops system and droplet-forming proteins (Reinkemeier et al. Science, 2019). Despite not having directly demonstrated the incorporation of D-Phe into the model protein sfGFP via mass spectrometry, we show that our engineered strain leads to higher GFP fluorescence compared to the strain without compartmentalization as well as to the wild type MG1655 wild type strain.
Since our work centers around D-amino acids, we wanted to provide a way to produce D-amino acids by using naturally occurring amino acid -racemases. For this, we decided to work with the exemplary PLP-independent aspartate racemase (BBa_K3009012).
Amino acid racemases exist in many organisms (Yoshimura et. al, J.Biosci. Bioeng. 2003) and convert L- to D-amino acids for different purposes, like building the bacterial cell wall or using D-amino acids as neurotransmitters. Racemases are especially well-characterized in bacteria and archaea, and can be divided into two classes: Those dependent on the cofactor pyridoxal phosphate (PLP), and those that are independent of co-factors (Ahn et. al, Organic & Biomolecular Chemistry, 2018) . Previously biology paid little attention to this topic, that is why we wanted to transform these enzymes into a tool for D-amino acid production. With a racemase biobrick, it should be possible to make D-amino acids in any lab, without the need to order them at great cost.
Our main concerns were that the racemase in question would not interfere with the cellular processes of E. coli , and be as easy to use as possible. We therefore picked the PLP-independent aspartate-racemase from the archeon Pyrococcus horikoshii. (Kawakami et al., amino acids 2015) In the course of our experiments, we wanted to first design a plasmid which is able to express the enzyme in E. coli , and then detect the L-to-D conversion of L-Aspartate supplied to the bacterial liquid cultures. To detect the D-aspartate, the method of choice was based on our D-amino acid oxidase-subproject, joining both tools into an interconnected workflow. By introducing racemases as tools to provide D-amino acids through conversion in E. coli , we wanted to encourage teams working with D-amino acids to explore new possibilities to use them in an in vivo setting. Producing pure D-amino acids with the help of cells would also be a great application, as we have experienced that chemically synthesized D-amino acids can be quite expensive. This way, we want to make D-amino acids accessible to everyone who wants to dive deeper into this field.
Based on the fact that D-peptides cannot be degraded by proteases, we concluded that they would also be resistant to degradation by the proteasome, a multimeric protein complex for breaking down proteins. In addition to degrading superfluous and misfolded proteins, the proteasome has also functions in the regulation of NFκB signalling (Mitchell,S Wiley Interdiscip Rev Syst Biol Med. 2016). Accordingly, the proteasome is a target of cancer therapy, especially for cancer types that depend on deregulated NFκB signalling for example multiple myeloma(Sun CY Biosci Rep. 2017). There are clinically approved proteasome inhibitors like Bortezomib and there is a lot of research on the field of peptidomimetic proteasome inhibitors. However, all of them share the same disadvantage that their function relies on highly reactive chemical structures for inhibition. Off-target reactions lead to dose-limiting, severe side-effects of these drugs. As an alternative to non-optimal drug candidates, we had the idea to test D-peptides as proteasome inhibitors. Firstly, D-peptides possess good qualities for therapeutics in general as they are stable(Milton et al., Science 1992), non-immunogenic and can be administered orally. Secondly we hypothesized that D-peptides would not be released from the proteasome due to incomplete digestion, leading to an obstruction and inhibition of the organelle. In order to test this hypothesis we performed proteasome activity assays in cell lysates comparing the effect of L- and D-peptides.
When we decided to work with PSMα3, not just the synthesis of the D-form posed a problem but also the expression of the natural L-form. Because of its lytic properties an organism expressing it would pose a potential safety risk. Additionally, due to the antimicrobial activity of PSMα3, the expression in E. coli would be difficult 1. These problems necessitated the chemical synthesis of L-PSMα3. To enable future iGEM teams to safely express toxic proteins in E. coli and produce them in sufficient yields we thought about an expression system based on split inteins. Split inteins are naturally occuring proteins and when they bind together they reconstitute to a full intein that autocatalytically splices out of the precursor proteins and attaches the flanking regions together by forming a new peptide bond.
