Project description

Concave - (R)evolutionizing protein targeting and degradation

Antibodies are among the most important tools not only for protein analytics, but also diagnostics and therapeutics. Antibodies evolved to be highly efficient and specific protein detectors and binders, playing an important role in the immune system of many vertebrates [1]. Yet antibody development suffers from multiple problems. It is not only expensive and time consuming [2], but also dependent on the use of animals [3, 4], which is why researchers have to rely on third parties for development. Furthermore, antibody based analytics, diagnostics and therapies struggle with targeting conserved mammalian proteins [5]. A new universal protein binder could be the solution to many of these problems.

This is where Repebodies come into play. Like antibodies, Repebodies are capable of detecting and binding specific proteins of interest (POI). Repebodies are derived from the antibody equivalent in jawless fish (‘Agnatha’), and ideal for expression in E. coli, greatly simplifying their production [6]. However, Repebody development is based on rational design and similar approaches, making it cost and time intensive. These challenges are only exacerbated by their compared novelty. With Concave we are working on a new way to dynamically create a vast library of variable Repebodies. Without any manual intervention, the system will also select for the Repebody best suited to the desired application. Additionally, we are exploring the possibility of applying Repebodies in targeted protein degradation (TPD). TPD uses a cell’s own degradation machinery to degrade a specific protein [7]. It is a potent method to determine a protein’s function in an organism or possibly even to combat disease.

In the following sections we will explore the theoretical background of every part employed in the Concave system, how we imagine the toolbox to function, and how far we have come with our research. Furthermore, we will give an overview over potential applications of Repebodies.

VLRs and Repebodies

Variable lymphocyte receptors (VLRs) play an essential role in the adaptive immune system of jawless vertebrates. VLRs are proteins with variable sequences that recognize and bind to antigens during the adaptive immune response. Each lymphocyte cell expresses one unique VLR sequence, formed by gene rearrangement during cell maturation. These proteins are the evolutionary counterpart to antibodies and T-cell receptors [8,9].

Structure of VLRs

VLRs are built from repeating units called leucine-rich repeats (LRRs). These are modular motifs commonly found in proteins that mediate protein-protein interactions [10]. The VLR structure (Figure 1) typically includes [11]:

• An N-terminal LRR (LRRNT) and a first non-variable LRR (LRR1).
• A variable number of variable LRRs (LRRVs), which determine antigen-binding specificity.
• A terminal variable LRR (LRRVe), a connecting peptide (CP), and a C-terminal LRR (LRRCT).
• A tail region that mediates membrane anchoring and multimerization of VLRs in the immune response.

VLRs fold into a crescent shape, with the concave antigen-binding surface presenting the variable amino acids in the LRRVs [9]. A peptide loop from the LRRCT region extends over the antigen-binding surface and can further influence binding specificity [12,13].

VLR diversity

Unlike antibodies, which are created through V(D)J (Variable-Diversity-Joining) recombination by combining gene fragments [14], VLRs achieve diversity through a different type of gene rearrangement. VLRs are expressed from just a single gene that is flanked by many cassette sequences, which encode variable LRRs. Each cassette can consist of one or multiple LRRVs of different amino acid composition. During development of the immune cell a random array of these cassettes is copied into the incomplete germ-line VLR gene, creating a vast pool of diverse antigen-binding receptors in a process similar to gene conversion or copy choice [15, 16, 17].

Why VLRs are interesting

VLRs represent a promising alternative to antibodies in research and diagnostics. The significant difference in structure and composition between VLRs and antibodies leads to substantial changes in their binding affinity to targets. Since VLRs originated in evolutionarily distant animals, they recognize molecular structures that antibodies often cannot [11,18]. For example, they can bind conserved mammalian proteins or complex sugar patterns (glycan structures) that antibodies do not usually attack, due to immune tolerance mechanisms in mammals [19]. This makes them a valuable alternative binding scaffold. Novel VLRs can also be discovered via immunization of jawless vertebrates, analogous to the selection of novel antibodies [18]. The soluble monomeric form of a VLR, without the multimerizing tail, can then be applied in research in the same manner as antibodies.

Why you should use Repebodies

For widespread use, proteins like VLRs or antibodies ideally need to be expressed in simple, scalable systems such as Escherichia coli. However, both proteins suffer from low yields and frequent misfolding when expressed in bacteria, which limits large-scale applications [20].

To solve this problem, Lee, Sang-Chul et al. developed Repebodies, synthetic proteins inspired by the VLR scaffold but optimized for bacterial expression [21]. A LRRV consensus sequence was generated from published sequences on the UniProt and the NCBI database to create a template scaffold. The N-terminal part of the VLR template was replaced with a segment from another naturally occurring LRR protein, Internalin B from Listeria monocytogenes [22]. In particular, the LRRNT as well as LRR1 and one of the LRRVs were exchanged. This modification leads to an increase in solubility, proper folding and higher yields, when expressed in E. coli. Repebodies keep the diversity and specificity of VLRs, while being easier and cheaper to produce.

New Repebodies are created by introducing variability into the LRR regions through degenerate codons, which are codons that include one or more random bases. This results in codons that encode different amino acids, depending on which bases are introduced. By incorporating this into the Repebody nucleotide sequence during gene synthesis, random amino acids are introduced at the variable sites in the LRRVs. The resulting randomized Repebody library can then be screened (often using phage display) to select a Repebody that binds to the protein of interest (POI) [21]. Once selected, this Repebody can be fused to functional domains or with different effector proteins. This enables the use in assays, as one would use an antibody, e.g. with an attached alkaline phosphatase for use in ELISA [23].

Repebodies already have an advantage over antibodies in the design of novel binders for targets, by being animal-independent. Expression in E. coli also vastly simplifies production and purification. However, the diversity generated by artificial mutagenesis and selection is still limited in comparison to the diversity of naturally animal-generated VLRs and antibodies. Thus, while already promising, Repebody development has room for improvement in both diversity generation and selection throughput.

In our project, we focused on improving these limitations. We employed an in vivo LRRV mutation system by applying retron-based mutagenesis inside living cells to enhance diversity generation. To accelerate binder discovery, we use a high throughput selection system. Finally, to demonstrate the practical value of our engineered Repebodies, we applied them in standard lab assays and extended their application to targeted protein degradation in mammalian cells.

Retrons

A key challenge in synthetic biology that we are confronting is generating targeted, confined mutations, instead of relying on a global, uncontrolled and stochastic mutation rate. Some natural systems address this problem, of creating large amounts of variants of a single, functional gene. For example vertebrates create enormous diversity in their immune receptors through complex multi-enzyme systems, however for this reason among others, transposing this system into Escherichia coli or another prokaryotic cell would be too complex.

We want something simpler. A stand-alone system that exploits the benefits of E. coli as a genetically-tractable, fast growing, experimental workhorse.

Since early on in the project, we were motivated to design a platform to create confined variability within specific regions of a consensus sequence, mimicking the results of immune system processes, but using prokaryote-friendly machinery.

Conceptually this meant maintaining conserved structural sites, while allowing delimitated regions to vary between individual bacteria. For scaffolds such as Repebodies, composed of repeating modular motifs as long as a controllable system for site-specific mutagenesis can be found there is potential that by targeting a repeating sequence a scalable mechanism to generate large combinatorial libraries becomes possible.

Repebodies and ssDNA recombineering

Our initial approach was inspired by modular immune proteins like variable lymphocyte receptors (VLRs), and related Repebody-like structures. These repeating motifs are attractive because substituting individual repeat elements theoretically allows combinatorial tuning of binding pockets and recognition surfaces. However, we needed to figure out how to implement such targeted substitutions efficiently in E. coli. This led us to shortly explore ssDNA recombineering, a well-established technique for introducing mutations in prokaryotic genomes. In this system, short single-stranded DNA (ssDNA) oligonucleotides act as donor molecules that are aligned with transient ssDNA regions generated during replication, particularly on the lagging strand. With the help of single-strand annealing proteins (SSAPs) and strain-level optimizations such as mismatch repair knockouts or nuclease deletions, these oligos can integrate into the bacterial genome [24, 25]. While promising, the technology showed several limitations. Most notably, it required the repeated external delivery of donor oligos, typically by electroporation, followed by growth cycles, before new rounds of mutagenesis could occur. This makes it impractical for generating scalable libraries. Moreover, although sequencing studies have demonstrated the success for nanobody diversification using this approach, its dependency on resupplying donors limits its appeal for an in vivo system intended to produce diversity autonomously [26, 24].

The rationale for retrons

From this point, we asked ourselves: Could an in vivo system generate its own donor DNA continuously, without the need for external oligos? This question brought us to retrons, which are naturally occurring elements found in certain bacteria. Retrons consist of two components: a non-coding RNA and a retron-specific reverse transcriptase (RT). The non-coding RNA folds into a secondary structure that self-primes reverse transcription. The retron RT then transcribes a portion of this RNA into DNA, forming a unique branched RNA–DNA hybrid, known as msDNA. Importantly, the DNA arm of this hybrid can serve as a donor molecule for recombineering, just like externally supplied single-stranded (ss)DNA [27, 28].

We realized that if we replaced the msd sequence with our own designed sequence, retrons could provide a continuous intracellular supply of ssDNA donors. This would confine diversity to our regions of interest while avoiding the practical limitations of electroporation-based ssDNA recombineering. In principle, a single retron donor could interact with multiple copies of our highly similar LRRV modules. By using n different retrons, we generate a combinatorial variability space of n⁴ different Repebodies, since our Repebody has four retron-accessible repeats [27, 28].

Designing our retron system

When designing our retron architecture, we evaluated recent retron tools described in the literature, including sophisticated multiplex retron libraries [29, 30]. Although these approaches offered high throughput, at this stage we focused on testing two donor cassettes in parallel to validate the system before scaling it up further. We chose to work with the plasmid pMS366, which was deposited to Addgene (Plasmid #182131) and is described as a streamlined retron-recombineering vector [30]. Several features made it attractive: It encodes CspRecT, a specialized recombinase that requires lower expression levels than the more established SSAP lambda red beta recombinase, thereby reducing cellular stress. Its recombinase expression is controlled by the pBAD promoter, which is inducible and less metabolically burdensome than stronger constitutive promoters, yielding faster bacterial growth. It features a temperature-sensitive origin of replication (ori) that allows “curing” of the plasmid when cultures are shifted to high temperatures. This design provides an elegant way to switch off mutagenesis once diversity has been generated, so that further propagation occurs in a more stabilized genetic state.

Cloning strategy

One important technical consideration was where to place our custom retron RNA sequence. In pMS366, a multiple cloning site downstream of the RT and CspRecT gene offers accessibility. We deliberately chose this location, because positioning the retron RNA in close proximity to the origin of replication reduces competition between donor products and their own genetic template sequence. This could improve the efficiency of integration into our chosen target gene rather than into the retron locus itself [31]. This arrangement also helps to avoid a significant potential pitfall. In our case, Repebody modules are highly homologous sequences. Without careful design, donor DNA could recombine with other donor cassettes, creating a catastrophic “homogenization” feedback loop in which all designed cassettes converge into the same sequence. By placing our retron in a position where it is more insulated from competing recombination, we reduce the risk of generating uniform populations rather than diverse ones.

Conclusion - Why retrons matter for our project

Retrons elegantly solve our initial problem of generating defined, scalable, and confined variability within a bacterial system. Unlike classical mutagenesis, which is random, and ssDNA recombineering, which is logistically intensive, retrons offer a continuous endogenous source of donor DNA. This enables programmability: the variability can be confined to specific regions of our target sequence, thus avoiding genome-wide mutagenesis. For our project, this means that large libraries of Repebody variants can be generated autonomously inside E. coli, leveraging their natural growth dynamics to explore a vast combinatorial search space. With the benefits of switchable expression and curable plasmids, retrons provide both flexibility and control. This makes them a central part of our long-term strategy for engineering controllable self-creating binding protein libraries.

Selection

BACTH

The Bacterial Adenylate Cyclase Two Hybrid (BACTH) system was developed in 1998 by Karimova et al. [32]. It is based on the reconstitution of the activity of a split adenylate cyclase and is used to detect protein-protein interactions. The BACTH system is capable of detecting the interaction between two cytoplasmic or inner membrane bound proteins [33].

Adenylate cyclase catalyses the conversion of ATP to cyclic AMP (cAMP), an important second messenger not only in bacteria, but also in mammalian cells [34]. The BACTH system uses parts of the adenylate cyclase toxin from the bacterium Bordetella pertussis. In its natural context, after being secreted by B. pertussis, the toxin invades the mammalian host cell through a tunnel-forming hemolytic domain at its C-terminus. In the host, it binds to the mammalian protein calmodulin. The resulting conformational changes activate the catalytic properties of the adenylate cyclase. This leads to elevated intracellular cAMP levels in the invaded cell and disrupts many cellular signaling pathways [35].

The BACTH system uses only the catalytic domain and the calmodulin-binding site of the adenylate cyclase. The hemolytic domain of the toxin is excluded. It was discovered that the catalytic and calmodulin binding sites, also called T25 and T18 respectively, can be separated into two inactive subunits, inactivating the adenylate cyclase. If both subunits are put in close proximity the enzymatic activity is restored. The BACTH system takes advantage of this by fusing the two subunits to proteins of interest. If the proteins, also called ‘prey’ and ‘bait’, interact, functional complementation occurs and the adenylate cyclase synthesizes cAMP from ATP [33].

In E. coli, cAMP can bind to the cAMP receptor protein (CRP), also known as catabolite activator protein (CAP). The cAMP-CRP complex can bind to specific DNA sequences and act as a transcription activator [34]. Among other processes, the cAMP-CRP complex can also activate the lactose and maltose operons, allowing for these sugars to be metabolized [33]. Therefore, one can monitor protein-protein interactions by assessing the ability of engineered E. coli cells, lacking their native adenylate cyclase, to metabolize sugars such as lactose or maltose.

Protein interaction - reporter gene activation

In order to test whether two proteins interact, they are each fused to one of the subunits. These fusion proteins are introduced into an E. coli strain that lacks its endogenous adenylate cyclase. Such bacteria are unable to use lactose or maltose as an energy source, since cAMP is needed to express the necessary genes. If the proteins fused to the adenylate cyclase subunits interact, cAMP is produced, enabling the expression of proteins under control of the lactose or maltose operon. Different readout assays can be used to detect the activation of these operons [33].

Colony agar assays

The bacteria can be grown on MacConkey agar. MacConkey agar contains the pH-indicator neutral-red, which can detect pH-changes, caused by acidic byproducts of the lactose metabolism [36]. Pink-colored colonies indicate cAMP production and therefore that the proteins fused to the adenylate cyclase subunits are interacting, while off-white colonies suggest the opposite [33, 36]. An alternative to MacConkey agar is X-Gal. X-Gal can be cleaved by the beta-galactosidase, encoded by the lac-operon, into galactose and 4-chloro-3-bromoindigo [37], which can self-dimerize to form a blue dye. Blue colonies indicate an interaction between the proteins fused to the adenylate cyclase subunits [33].

Growth-based assays

Another readout confirming targeted protein interaction is the activation of the lactose or maltose operon. When bacteria are grown on minimal M63 agar supplemented only with lactose or maltose, survival depends on restored adenylate cyclase activity. If the fusion proteins interact, cAMP production enables operon expression, which allows the cells to metabolize the sugar and grow. In contrast, bacteria expressing non-interacting fusion proteins fail to generate cAMP and cannot survive under these conditions [33].

Applicability

For the Concave toolbox we need a reliable and high-throughput selection system, capable of screening large libraries of various Repebodies, for the one with the desired target protein. This system would also need to function in life bacteria, as this would make sequencing of the selected Repebody significantly easier. The BACTH system might be suitable to be applied in the Concave toolbox. It functions in vivo, which means that the Repebodies in selected clones can easily be sequenced and produced. A growth based assay could especially be well suited to select single functional Repebody-target combinations from a large library. Finally, it may also be possible to adapt it to be high-throughput, using simple automation solutions. Because of these characteristics, we decided to test the BACTH system on its applicability in at the Concave toolbox (see results, engineering cycle pages).

Cell-Surface presentation

Magnetic selection

We are also developing the magnetic selection system based on a publication by Aubrey et al. [38] The magnetic E. coli cells were designed by overexpressing a fusion construct of mCherry and ferritin. Ferritin from Pyroccocus Furiosus is a homo-24-meric protein central to iron storage. Ferritins exist in all domains of life and are mainly used to prevent hydroxyl radical formation by shielding the iron from hydrogen peroxide and storing it in the less reactive Fe(III) form [40]. The subunits assemble into a hollow spherical structure, the center of which is used to store a ferrihydrite-like mineral core [39]. Each subunit contains an active centre which can oxidize Fe²⁺ to Fe³⁺, using oxygen as the electron acceptor. By fusing each of the ferritin subunits to mCherry and overexpressing the resulting construct, the proteins aggregate and begin to degrade, leading to the formation of magnetic iron oxide inclusion bodies. [38]

To implement a novel selection approach we plan to modify the magnetic E. coli described in this article by attaching the target antigen on their surface via a SpyTag/SpyCatcher-system. In turn, another strain of bacteria, expressing our randomized Repebodies on their surface, would be able to bind the target antigen, depending on their Repebodies binding affinity (see Fig. X below).

A mixture of magnetic bacteria presenting the target antigen and bacteria with a functional Repebody can then be pulled out of the culture using a magnet.

The Repebody presenting strain is then further selected from the mixture using antibiotic resistance. To select the optimal antigen-binding Repebody, we further assess single colonies with our previously established assays.

Cellulose-Gel selection

Developed in correspondence with Sebastian Cocioba, we planned this system as an alternative to magnetic selection in case the magnetic force exerted on the bacteria was too weak to efficiently separate them. In this system, the target antigen is fused to a cellulose binding domain (CBD) from Clostridium Thermocellum, which then binds it to the fibres of a cellulose gel. This cellulose gel can be produced from dissolved and reprecipitated cellulose from Schweizer’s reagent (tetraammine copper(II) hydroxide) [41]. After thorough washing, the gel can be treated with the CBD-target fusion construct and used as the stationary phase in cell chromatography, where the cells use the Repebody presented on their surface to bind to the gel. This method has the added advantage over magnetic selection, that the selection isn’t just qualitative. In this method, the stronger the binding strength, the slower the bacteria move through the column, meaning that later fractions bind stronger to the target antigen than previous fractions.

Targeted protein degradation

Historical Background of Targeted Protein Degradation

The development of therapeutic agents has traditionally focused on regulating protein activity. However, the efficacy of these approaches is highly limited to proteins with accessible enzymatic or binding domains. In 1999, the establishment of Targeted Protein Degradation (TPD) shifted the focus toward event-driven pharmacology. Instead of inhibiting protein function, TPD aims to leverage the cell’s native quality control machinery to eliminate specific disease-associated proteins [42]. The mechanism underlying this concept is based on proteostasis. The term encompasses the network that governs protein synthesis, folding, transport, and degradation, which is essential for cellular vitality. In eukaryotic cells, there are two major interconnected pathways for degradation: the ubiquitin-proteasome system (UPS) and the autophagy-lysosomal pathway (ALP) [43].

Ubiquitin–Proteasome System–Based Approaches

The ubiquitin-proteasome system (UPS) mainly degrades short-lived, soluble proteins. This process is dictated by a sequential enzymatic cascade involving E1 (activating), E2 (conjugating), and E3 (ligase) enzymes, which culminates as polyubiquitination (Figure 11)[42]. Early targeted protein degradation (TPD) approaches focused on utilizing the UPS system. First published methods like PROteolysis TArgeting Chimeras (PROTACs) are synthetic, heterobifunctional molecules composed of three parts: a ligand that binds to the target protein (protein of interest, POI), a ligand recruiting an E3 ubiquitin ligase, and a flexible linker that connects the two. By bringing the POI and the E3 ligase together, PROTACs induce ubiquitination and subsequent proteasomal degradation of the target protein. Their modular design allows different E3 ligases to be linked to diverse targets that lack traditional druggable sites. Moreover, they have proven to be highly effective in overcoming drug resistance caused by target mutations or overexpression. These features provide a new spectrum in drug discovery [45].

Despite their therapeutic promises, they have high molecular weights (700-1000 Da), which can limit oral bioavailability and cell permeability [46]. To resolve these issues, molecular glues (MGs) were developed. They stabilize the interaction between a ubiquitin ligase and a protein of interest (POI) without a linker (Figure 12) [47]. Although MGs offer a solution to the size problem, issues exist in designing them. This is because the binding interface between the two proteins was not pre-existing as in PROTACs, but emerges only when the glue is bound. Therefore, the binding energy arises from a ternary complex instead of binary binding, making the interface or induced structural rearrangement hard to predict computationally. Moreover, the limited repertoire of E3 ligases restricts the available chemical space for targeted degradation, reducing target specificity and ultimately hindering the development of UPS-based degraders for the majority of disease-associated proteins [42]. These limitations demonstrate the necessity of alternatives to expand the applicability of TPD. As a result, strategies utilizing the autophagy-lysosomal pathway (ALP) were explored.

Autophagy–Lysosomal Pathway (ALP)

The autophagy-lysosomal pathway (ALP) is responsible for the degradation of long-lived proteins, macromolecular complexes, and even damaged organelles. This system is essential for addressing proteinopathies that the proteasome cannot manage. The ALP encompasses three distinct modes: macroautophagy, microautophagy, and chaperone-mediated autophagy (CMA)(Figure 13) [42].

The first ALP-based approach, lysosome-targeting chimeras (LYTACs), targets extracellular and membrane proteins via the endosome-lysosome pathway. These proteins represent approximately 40% of the encoded proteome and play key roles in many diseases. A LYTAC molecule works by simultaneously binding a protein of interest (POI) outside the cell and a cell-surface lysosome-targeting receptor (LTR). This creates a ternary complex that is then internalized through clathrin-mediated endocytosis, sending the POI to the lysosome for degradation [44]. Beyond LYTACs, other lysosomal targeted proteins employing specialized scaffolds have also shown therapeutic potential. Examples include AbTACs (antibody-based PROTACs) [48], bispecific aptamer chimeras [49], and GlueTACs [50]. Furthermore, strategies that recruit the autophagy machinery have proven effective in targeting diverse proteins. These include: AUTAC (autophagy-targeting chimera) [51], ATTEC (autophagosome tethering compound) [52], AUTOTAC (p62-mediated degradation) [53], and CMA-based degraders [54]. Nevertheless, these peptide-based degraders still face significant burdens regarding stability and intracellular delivery efficiency.

The Nanobody-based P62 Degradation Factory (PDF-Bin)

To overcome the size constraints of full antibodies and the pathway limitations of small-molecule TPDs, researchers further explored p62-induced degradation. p62 is a multifunctional adaptor protein that plays a central role in selective autophagy. It acts as a scaffold, linking polyubiquitinated proteins to the autophagic machinery, and therefore, bridging autophagy to the ubiquitin-proteasome system (UPS) [55]. Using the highly modular bispecific nanobodies, the PDF-Bin system comprises a nanobody against p62 and a nanobody against the POI (Figure 14). This approach is capable of dynamically engaging multiple degradation pathways and reduces reliance on any single E3 ligase or LTR route while enabling the targeting of complex and elusive proteins [56].

Tackling the challenges - Our project

Advancements in TPD have demonstrated the transformative potential of event-driven pharmacology in addressing real-world disease challenges—particularly in areas where traditional inhibitors have failed. Nevertheless, issues regarding bioavailability, accessibility, and flexibility while maintaining affinity and sensitivity must still be resolved for TPD to achieve its full therapeutic potential. Therefore, in our iGEM project, we address these challenges by using Repebodies as modular binding domains to guide targeted protein degradation (TPD) in mammalian cells. This design aims to improve degradation specificity, increase intracellular accessibility, and expand the toolkit for programmable protein-based therapeutics.

Implementation

The idea

Now the question stands: What do we envision Concave to be? Concave is supposed to be an evolution of conventional antibodies. We intend it to be a reliable alternative to antibodies that is not only more ethical and sustainable, but also allows researchers access to a cheaper and more accessible protein binder, with their desired specifications. The Concave toolbox is supposed to also allow smaller companies and academic labs to develop their own Repebodies. For that application it would need to incorporate every part needed:

1. The toolbox needs a custom E. coli strain with all required mutations. This is not limited to mutations in the genome, to enable the functionality of the mutation and selection system, but also an integrated Repebody, for ease of selection, although this is not strictly necessary.
2. The ssDNA-optimized-Repebody (BBa_25IQCJ3B) gene itself, ideally included in the strains genome to ensure that every bacteria has only one Repebody gene. Thus the selection of the bacteria with the best Repebody and the purification of it is simplified. Alternatively, a plasmid with the Repebody gene can be introduced into the bacteria.
3. Plasmids with multiple retron cassettes, for the production of the mutating ssDNAs. Multiple plasmids with retrons cassettes, causing different types of mutations (e.g. a cassette that mostly encodes for hydrophilic amino acids), can be included.
4. All parts of the selection system, that would be required in addition to the custom strain. This could include a second custom strain or additional plasmids.

The Concave toolbox could either be used by research labs to create their own custom Repebodies, or by companies specialized on developing a vast array of different Repebodies. We also wanted to test a pre-existing application using Repebodies, targeted protein degradation (TPD). This approach involves using a molecule possessing an affinity for a target and for proteins involved in protein degradation. The basic principle is that this bispecific-molecule would connect a target to the degradation machinery of a cell, causing it to selectively be degraded [7]. We intended to employ this principle by using the Repebody as the component of the bispecific molecule binding the target, mCherry.

Where we are at

We managed to assemble the mCherry-binding-Repebody (BBa_25X0LA8K), as well as a Repebody, which is optimized to be mutated by single-stranded DNA pieces (BBa_25IQCJ3B). We have proven both Repebodies to be capable of specifically binding mCherry, while not interacting with eGFP (see results page). Despite not directly proving mutation of our Repebody, we developed a robust approach on how to troubleshoot and improve the proof of concept (see results page). We also explored different selection systems. In the case of the BACTH system, we showed that it is capable of detecting the interaction between a Repebody and its specific target. On the TPD side, we have successfully shown expression of our target proteins and the degrader construct containing the Repebody in mammalian cells (HeLa and U2OS). For a more detailed breakdown of our results have a look at our results page. Visit our engineering success page to see how we got our results.

What’s missing?

We managed to lay the foundation for a complete toolbox that allows for the creation, selection and application of Repebodies. The Concave toolbox however is not finished yet. Critical components are still missing. The implementation of the Repebody, mutation and selection system were impossible, due to the time constraints we experienced. We would also need to characterize the mutation of the Repebodies further, to give more solid predictions on how large the potential library would be and how long it would take to develop a custom Repebody. Furthermore, the selection systems would need optimization and refinement. They were not tested in conjunction with the mutation system, so we couldn’t explore their limitations. All parts would need to be assembled into a system of plasmids, including the Repebody and at least parts of the selection system on an integrative plasmid, to guarantee only one Repebody species per bacterial cell. Finally, an E. coli strain with all necessary mutations, required by the mutation and selection system (e.g. cya^-, recJ^-, exoI^-) needs to be engineered. Unfortunately, we fell short of confirming degradation using the Repebody, as the complex nature of mammalian cell expression and behavior has resulted in several experimental challenges that must be addressed through iterative refinements regarding the design and execution of the experiments. Only if all these parts are successfully built and functional, the Concave toolbox can be considered finished.

Outlook

Mutation system

To fix the problems with the mutation system we have encountered, a more established system for retron mutation, keeping the plasmid architecture intact, while only changing the ssDNA-donor sequence, should be employed.

Artifical intelligence

When thinking about implementing AI in our project, the general idea was to use it to generate or optimize Repebody sequences. To develop a more targeted approach to the topic, we contacted Dr. Gabriel Kalweit, whose research background in bioinformatics includes the generation of antibody sequences. There are advantages and disadvantages to Repebodies compared to antibodies when working with AI. On one hand, there is no large database for Repebodies to train the AI. On the other hand, the number of variables and subsequently the number of possible sequences is significantly lower. There are different approaches, one of which seems well suited for our project, because a pre-existing database is not required. Instead, the sequences would be generated by constructing a physical model through fixed sequence blocks. By implementing this AI pipeline, we could generate Repebody-diversity more efficiently. Instead of creating random variants and subsequently selecting promising candidates, the AI could directly propose viable sequences specific to a given target. These could then be synthesized and experimentally validated, therefore skipping the step of generating diversity via in-vivo mutation.

Applications

The Repebodies chemical and biological properties make them viable to be used in many different applications in analytics, diagnostics and therapeutics.

Analytics

As a part of our project we already developed a possible application of Repebodies for analytical purposes. Targeted protein degradation with a Repebody could be used to study the functions of proteins in live cells, by degrading them. From the resulting reaction of the cell the protein’s function could be deduced. But other applications are possible. During the BACTH assays we managed to prove that domains and even proteins can be fused N- and C-terminally to the Repebody without impeding its functionality. This opens the possibility to replace antibodies in protein analysis assays such as western blots or ELISA, since common reporters, such as HRP or fluorophors, could be fused to a Repebody.

Diagnostics

Instead of detecting certain proteins of interest for analytical purposes, it is also possible to use them for diagnosing diseases.The targeted protein could be spike proteins of viruses, certain antibodies or other pathogen associated molecules. The Concave toolbox could make development of such diagnostics assays easier and more accessible.

Therapeutics

Targeted Protein Degradation with Repebodies could not only be used for analytical applications, but also therapeutics. With the right delivery system, TPD could degrade for example viral proteins in human cells and thereby act as an antiviral drug. Prof. Schäfer also gave us the idea that you could potentially use Repebodies as a clotting agent, by allowing them to accumulate in place, where bleeding occurs. The unique binding properties of Repebodies also make them exceptionally interesting for combating cancer. Prof. Römer told us that detecting specific glyco-motives could help identify cancerous cells. The ability of Repebodies to detect those, opens the possibility to use them as markers to enable the body’s immune system to detect transformed cells more easily. There already exist multiple Repebodies that can detect certain types of cancer. The Concave toolbox could make research in this direction significantly easier.

References

[1] Palma, J., Tokarz-Deptuła, B., Deptuła, J., & Deptuła, W. (2018). Natural antibodies – facts known and unknown. Central European Journal of Immunology, 43(4), 466–475. https://doi.org/10.5114/ceji.2018.81354

[2] Fraunhofer-Gesellschaft. (2021, February). Top speed antibody development. Fraunhofer. https://www.fraunhofer.de/en/press/research-news/2021/february-2021/top-speed-antibody-development.html, retrieved on: 07.09.2025.

[3] Roche. (n.d.). Antibody development: From discovery to therapeutic application. Roche. https://www.roche.com/stories/antibody-development, retrieved on: 07.09.2025.

[4] MBL. (n.d.). Antibody development technology: Strengths of MBL. MBL. https://www.mblbio.com/e/company/strengths/antibody_development_technology02.html, retrieved on: 07.09.2025.

[5] Zhou, Y., Zhang, Y., & Xu, W. (2023). Antibody engineering for infectious diseases: Advances and prospects. mAbs, 15(1), 2273018. https://doi.org/10.1080/19420862.2023.2273018

[6] Lee, S.-C., Park, K., Han, J., Lee, J.-J., Kim, H. J., Hong, S., Heu, W., Kim, Y. J., Ha, J.-S., Lee, S.-G., Cheong, H.-K., Jeon, Y. H., Kim, D., & Kim, H.-S. (2012). Design of a binding scaffold based on variable lymphocyte receptors of jawless vertebrates by module engineering. Proceedings of the National Academy of Sciences of the United States of America, 109(9), 3299–3304. https://doi.org/10.1073/pnas.1113193109

[7] Zhao, L., Zhao, J., Zhong, K., Tong, A., & Jia, D. (2022). Targeted protein degradation: Mechanisms, strategies and application. Signal Transduction and Targeted Therapy, 7, 113. https://doi.org/10.1038/s41392-022-00966-4

[8] Alder, M. N., Rogozin, I. B., Iyer, L. M., Glazko, G. V., Cooper, M. D., & Pancer, Z. (2005). Diversity and Function of Adaptive Immune Receptors in a Jawless Vertebrate. Science, 310(5756), 1970–1973. https://doi.org/10.1126/science.1119420

[9] Herrin, B. R., Alder, M. N., Roux, K. H., Sina, C., Ehrhardt, G. R. A., Boydston, J. A., Turnbough, C. L., & Cooper, M. D. (2008). Structure and specificity of lamprey monoclonal antibodies. Proceedings of the National Academy of Sciences, 105(6), 2040–2045. https://doi.org/10.1073/pnas.0711619105

[10] Bella, J., Hindle, K. L., McEwan, P. A., & Lovell, S. C. (2008). The leucine-rich repeat structure. Cellular and Molecular Life Sciences, 65(15), 2307–2333. https://doi.org/10.1007/s00018-008-8019-0

[11] Waters, E. A., & Shusta, E. V. (2018). The variable lymphocyte receptor as an antibody alternative. Current Opinion in Biotechnology, 52, 74–79. https://doi.org/10.1016/j.copbio.2018.02.016

[12] Han, B. W., Herrin, B. R., Cooper, M. D., & Wilson, I. A. (2008). Antigen Recognition by Variable Lymphocyte Receptors. Science, 321(5897), 1834–1837. https://doi.org/10.1126/science.1162484

[13] Velikovsky, C. A., Deng, L., Tasumi, S., Iyer, L. M., Kerzic, M. C., Aravind, L., Pancer, Z., & Mariuzza, R. A. (2009). Structure of a lamprey variable lymphocyte receptor in complex with a protein antigen. Nature Structural & Molecular Biology, 16(7), 725–730. https://doi.org/10.1038/nsmb.1619

[14] Bassing, C. H., Swat, W., & Alt, F. W. (2002). The Mechanism and Regulation of Chromosomal V(D)J Recombination. Cell, 109(2), S45–S55. https://doi.org/10.1016/S0092-8674(02)00675-X

[15] Nagawa, F., Kishishita, N., Shimizu, K., Hirose, S., Miyoshi, M., Nezu, J., Nishimura, T., Nishizumi, H., Takahashi, Y., Hashimoto, S., Takeuchi, M., Miyajima, A., Takemori, T., Otsuka, A. J., & Sakano, H. (2007). Antigen-receptor genes of the agnathan lamprey are assembled by a process involving copy choice. Nature Immunology, 8(2), 206–213. https://doi.org/10.1038/ni1419

[16] Boehm, T., McCurley, N., Sutoh, Y., Schorpp, M., Kasahara, M., & Cooper, M. D. (2012). VLR-based adaptive immunity. Annual Review of Immunology, 30, 203–220. https://doi.org/10.1146/annurev-immunol-020711-075038

[17] Das, S., Boehm, T., Holland, S. J., Rast, J. P., Fontenla-Iglesias, F., Morimoto, R., Valadez, J. G., Heimroth, R. D., Hirano, M., & Cooper, M. D. (2023). Evolution of two distinct variable lymphocyte receptors in lampreys: VLRD and VLRE. Cell Reports, 42(8), 112933. https://doi.org/10.1016/j.celrep.2023.112933

[18] Bela-ong, D. B., Kim, J., Thompson, K. D., & Jung, T. S. (2024). Leveraging the biotechnological promise of the hagfish variable lymphocyte receptors: Tools for aquatic microbial diseases. Fish & Shellfish Immunology, 150, 109565. https://doi.org/10.1016/j.fsi.2024.109565

[19] McKitrick, T. R., Goth, C. K., Rosenberg, C. S., Nakahara, H., Heimburg-Molinaro, J., McQuillan, A. M., Falco, R., Rivers, N. J., Herrin, B. R., Cooper, M. D., & Cummings, R. D. (2020). Development of smart anti-glycan reagents using immunized lampreys. Communications Biology, 3(1), 91. https://doi.org/10.1038/s42003-020-0819-2

[20] Appelt, E. A., Thoden, J. B., Dancy, C., Bachmeier, H. D., Gehrke, S. A., Staffenson, M. A., Rayment, I., Katt, M. E., Holden, H. M., & Shusta, E. V. (2025). A method for facile production of variable lymphocyte receptors using SHuffle Escherichia coli. Biotechnology Progress, 41(3), e3530. https://doi.org/10.1002/btpr.3530

[21] Lee, S.-C., Park, K., Han, J., Lee, J., Kim, H. J., Hong, S., Heu, W., Kim, Y. J., Ha, J.-S., Lee, S.-G., Cheong, H.-K., Jeon, Y. H., Kim, D., & Kim, H.-S. (2012). Design of a binding scaffold based on variable lymphocyte receptors of jawless vertebrates by module engineering. Proceedings of the National Academy of Sciences, 109(9), 3299–3304. https://doi.org/10.1073/pnas.1113193109

[22] Marino, M., Braun, L., Cossart, P., & Ghosh, P. (1999). Structure of the InlB Leucine-Rich Repeats, a Domain that Triggers Host Cell Invasion by the Bacterial Pathogen L. monocytogenes. Molecular Cell, 4(6), 1063–1072. https://doi.org/10.1016/S1097-2765(00)80234-8

[23] Seo, H.-D., Lee, J., Kim, Y. J., Hantschel, O., Lee, S.-G., & Kim, H.-S. (2017). Alkaline phosphatase-fused Repebody as a new format of immuno-reagent for an immunoassay. Analytica Chimica Acta, 950, 184–191. https://doi.org/10.1016/j.aca.2016.11.013

[24] Higgins, S. A., Ouonkap, S. V. Y., & Savage, D. F. (2017a). Rapid and Programmable Protein Mutagenesis Using Plasmid Recombineering. ACS Synthetic Biology, 6(10), 1825–1833. https://doi.org/10.1021/acssynbio.7b00112

[25] Caldwell, B. J., & Bell, C. E. (2019). Structure and mechanism of the Red recombination system of bacteriophage λ. Progress in Biophysics and Molecular Biology, 147, 33–46. https://doi.org/10.1016/j.pbiomolbio.2019.03.005

[26] Al-ramahi, Y., Nyerges, A., Margolles, Y., Cerdán, L., Ferenc, G., Pál, C., Fernández, L. Á., & de Lorenzo, V. (2021). ssDNA recombineering boosts in vivo evolution of nanobodies displayed on bacterial surfaces. Communications Biology, 4(1), 1169. https://doi.org/10.1038/s42003-021-02702-0

[27] Zheng, X., Xing, X.-H., & Zhang, C. (2017). Targeted mutagenesis: A sniper-like diversity generator in microbial engineering. Synthetic and Systems Biotechnology, 2(2), 75–86. https://doi.org/10.1016/j.synbio.2017.07.001

[28] Simon, A. J., Ellington, A. D., & Finkelstein, I. J. (2019). Retrons and their applications in genome engineering. Nucleic Acids Research, 47(21), 11007–11019. https://doi.org/10.1093/nar/gkz865

[29] González-Delgado, A., Lopez, S. C., Rojas-Montero, M., Fishman, C. B., & Shipman, S. L. (2024). Simultaneous multi-site editing of individual genomes using retron arrays. Nature Chemical Biology, 20(11), 1482–1492. https://doi.org/10.1038/s41589-024-01665-7

[30] Schubert, M. G., Goodman, D. B., Wannier, T. M., Kaur, D., Farzadfard, F., Lu, T. K., Shipman, S. L., & Church, G. M. (2021). High-throughput functional variant screens via in vivo production of single-stranded DNA. Proceedings of the National Academy of Sciences, 118(18), e2018181118. https://doi.org/10.1073/pnas.2018181118

[31] Ni, Y., Wang, Y., Shi, X., Yu, F., Ruan, Q., Tian, N., He, J., & Wang, X. (2024). Reducing competition between msd and genomic DNA improves retron editing efficiency. EMBO Reports, 25(12), 5316–5330. https://doi.org/10.1038/s44319-024-00311-6

[32] Karimova, G., Pidoux, J., Ullmann, A., & Ladant, D. (1998). A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proceedings of the National Academy of Sciences, 95(10), 5752–5756. https://doi.org/10.1073/pnas.95.10.5752

[33] Battesti, A., & Bouveret, E. (2012). The bacterial two-hybrid system based on adenylate cyclase reconstitution in Escherichia coli. Methods, 58(4), 325–334. https://doi.org/10.1016/j.ymeth.2012.07.018

[34] Gancedo, J. (2013). Biological roles of cAMP: variations on a theme in the different kingdoms of life. Biological Reviews, 88(3), 645-668. https://doi.org/10.1111/brv.12020

[35] Carbonetti, N. H. (2010). Pertussis toxin and adenylate cyclase toxin: Key virulence factors of Bordetella pertussis and cell biology tools. Future Microbiology, 5(3), 455–469. https://doi.org/10.2217/fmb.09.133

[36] Jung, B., & Hoilat, G. J. (2022). MacConkey Medium. In StatPearls. StatPearls Publishing. https://www.ncbi.nlm.nih.gov/books/NBK557394/

[37] X-Gal 1 g | Buy Online | Thermo ScientificTM | thermofisher.com. (n.d.). https://www.thermofisher.com/order/catalog/product/de/en/R0404

[38] Mary Aubry, Wei-An Wang, Yohan Guyodo, Eugénia Delacou, Jean-Michel Guigner, Olivier Espeli, Alice Lebreton, François Guyot, and Zoher Gueroui. Engineering E. coli for Magnetic Control and the Spatial Localization of Functions. ACS Synthetic Biology 2020 9 (11), 3030-3041. https://doi.org/10.1021/acssynbio.0c00286

[39] Honarmand Ebrahimi, K., Hagedoorn, PL., Jongejan, J.A. et al. Catalysis of iron core formation in Pyrococcus furiosus ferritin. J Biol Inorg Chem 14, 1265–1274 (2009). https://doi.org/10.1007/s00775-009-0571-z

[40] Bai, L., Xie, T., Hu, Q. et al. Genome-wide comparison of ferritin family from Archaea, Bacteria, Eukarya, and Viruses: its distribution, characteristic motif, and phylogenetic relationship. Sci Nat 102, 64 (2015). https://doi.org/10.1007/s00114-015-1314-3

[41] Yasmin J. Dias, Alexander Kolbasov, Suman Sinha-Ray, Behnam Pourdeyhimi, Alexander L. Yarin. Theoretical and experimental study of dissolution mechanism of cellulose. Journal of Molecular Liquids, Volume 312, 2020, 113450, ISSN 0167-7322, https://doi.org/10.1016/j.molliq.2020.113450.

[42] Zhao, L., Zhao, J., Zhong, K. et al. Targeted protein degradation: mechanisms, strategies and application. Sig Transduct Target Ther 7, 113 (2022). https://doi.org/10.1038/s41392-022-00966-4

[43] Shukla, M., Narayan, M. Proteostasis and Its Role in Disease Development. Cell Biochem Biophys 83, 1725–1741 (2025). https://doi.org/10.1007/s12013-024-01581-6

[44] Banik, S.M., Pedram, K., Wisnovsky, S. et al. Lysosome-targeting chimaeras for degradation of extracellular proteins. Nature 584, 291–297 (2020). https://doi.org/10.1038/s41586-020-2545-9

[45] Wang, C., Zhang, Y., Chen, W. et al. New-generation advanced PROTACs as potential therapeutic agents in cancer therapy. Mol Cancer 23, 110 (2024). https://doi.org/10.1186/s12943-024-02024-9

[46] Aishwarya L. Saraswat, Richa Vartak, Rehab Hegazy, Akanksha Patel, Ketan Patel, Drug delivery challenges and formulation aspects of proteolysis targeting chimera (PROTACs),Drug Discovery Today,Volume 28, Issue 1, 2023, 103387, ISSN 1359-6446, https://doi.org/10.1016/j.drudis.2022.103387.

[47] Stoeckli E. T. (2014). Protocadherins: not just neuron glue, more too!. Developmental cell, 30(6), 643–644. https://doi.org/10.1016/j.devcel.2014.09.008

[48] Cotton, A. D. et al. Development of antibody-based PROTACs for the degradation of the cell-surface immune checkpoint protein PD-L1. J. Am. Chem. Soc. 143,593–598 (2021). DOI: 10.1021/jacs.0c10008

[49] Miao, Y., Gao, Q., Mao, M., Zhang, C., Yang, L., Yang, Y., & Han, D. (2021). Bispecific Aptamer Chimeras Enable Targeted Protein Degradation on Cell Membranes. Angewandte Chemie (International ed. in English), 60(20), 11267–11271. https://doi.org/10.1002/anie.202102170

[50] Zhang, H., Han, Y., Yang, Y., Lin, F., Li, K., Kong, L., Liu, H., Dang, Y., Lin, J., & Chen, P. R. (2021). Covalently Engineered Nanobody Chimeras for Targeted Membrane Protein Degradation. Journal of the American Chemical Society, 143(40), 16377–16382. https://doi.org/10.1021/jacs.1c08521

[51] Takahashi, D., Moriyama, J., Nakamura, T., Miki, E., Takahashi, E., Sato, A., Akaike, T., Itto-Nakama, K., & Arimoto, H. (2019). AUTACs: Cargo-Specific Degraders Using Selective Autophagy. Molecular cell, 76(5), 797–810.e10. https://doi.org/10.1016/j.molcel.2019.09.009

[52] Li, Z., Zhu, C., Ding, Y., Fei, Y., & Lu, B. (2020). ATTEC: a potential new approach to target proteinopathies. Autophagy, 16(1), 185–187. https://doi.org/10.1080/15548627.2019.1688556

[53] Ji, C. H., Kim, H. Y., Lee, M. J., Heo, A. J., Park, D. Y., Lim, S., Shin, S., Ganipisetti, S., Yang, W. S., Jung, C. A., Kim, K. Y., Jeong, E. H., Park, S. H., Bin Kim, S., Lee, S. J., Na, J. E., Kang, J. I., Chi, H. M., Kim, H. T., Kim, Y. K., … Kwon, Y. T. (2022). The AUTOTAC chemical biology platform for targeted protein degradation via the autophagy-lysosome system. Nature communications, 13(1), 904. https://doi.org/10.1038/s41467-022-28520-4

[54] Fan, X., Jin, W. Y., Lu, J., Wang, J., & Wang, Y. T. (2014). Rapid and reversible knockdown of endogenous proteins by peptide-directed lysosomal degradation. Nature neuroscience, 17(3), 471–480. https://doi.org/10.1038/nn.3637

[55] Cha-Molstad, H., Yu, J. E., Feng, Z., Lee, S. H., Kim, J. G., Yang, P., Han, B., Sung, K. W., Yoo, Y. D., Hwang, J., McGuire, T., Shim, S. M., Song, H. D., Ganipisetti, S., Wang, N., Jang, J. M., Lee, M. J., Kim, S. J., Lee, K. H., Hong, J. T., … Kim, B. Y. (2017). p62/SQSTM1/Sequestosome-1 is an N-recognin of the N-end rule pathway which modulates autophagosome biogenesis. Nature communications, 8(1), 102. https://doi.org/10.1038/s41467-017-00085-7

[56] Jia, W., Li, W., Li, Z. et al. An all-in-one targeted protein degradation platform guided by degradation condensates-bridging bi-specific nanobodies. Cell Res 34, 389–392 (2024). https://doi.org/10.1038/s41422-024-00942-3.