The primary aim of the work presented in this thesis was to change the inducer specificity of the DntR protein in order to improve the response to DNT. The long-term goal is to use this protein in a biosensor for DNT, a signature compound for detection of the explosive TNT. Another aspect of this work was to understand the mechanisms of inducer binding and how the binding of an inducer molecule changes the DntR structure into a state that triggers transcriptional activation. In the papers included in this thesis the inducer specificity of wt DntR has been investigated under different conditions. The functional effects of specific mutations have also been investigated, in some cases in combination with structure determination using X-ray crystallography. In addition, structural data offering insights into the details of inducer binding and conformational changes upon inducer binding are presented and discussed in terms of mechanisms for transcriptional activation by DntR. Furthermore, a directed evolution strategy was employed in order to find variants of DntR with improved response to DNT. A variant with a large improvement in the DNT response was isolated and characterized. In optimized growth conditions, this DntR variant had a nearly 10-fold increase in fluorescence in response to DNT compared to wt DntR. Specific substitutions found in this DntR variant are suggested to be important for changing the inducer response. / Syftet med denna avhandling har varit att förbättra förmågan hos proteinet DntR att upptäcka DNT. Det långsiktiga målet har varit att använda DntR i en biosensor för att upptäcka sprängämnet TNT, som avger DNT som en ”signaturmolekyl”. En annan aspekt har varit att bättre förstå den detaljerade mekanismen för hur DntR fungerar. DntR är ett protein som binder till en viss DNA sekvens (promotor) och reglerar hur gener intill denna promotorsekvens läses av. När en inducerande molekyl som t.ex. DNT binder till DntR förändras proteinets struktur på ett sådant sätt att DntR kan aktivera transkription av de gener som finns intill promotor-sekvensen. För att mäta hur DntR reagerar på olika inducerande molekyler har DntR uttryckts i bakterien Escherichia coli, som också innehållit promotorn som DntR binder till. Intill promotorn sitter en gen som kodar för proteinet GFP. När en inducerande molekyl binder till DntR, slås avläses gfp-genen, och det fluorescerande proteinet GFP produceras. Ju mer GFP som produceras i cellerna, desto högre fluorescens kan uppmätas när cellerna analyseras.   I de artiklar som presenteras i avhandlingen har vi undersökt hur olika substitutioner i DntR proteinet påverkar specificiten och sensitiviteten och hur dessa egenskaper kan påverkas av olika experimentella faktorer. Effekten av substitutioner har relaterats till strukturdata, där bilder av hur proteinet ser ut på molekylär nivå har tagits fram. Dessutom presenteras även en bild av hur DntR förändras beroende på om inducerande molekyler är bundna eller inte. En sådan strukturbild ökar förståelsen för de mekanismer som gör att bindning av en inducerande molekyl orsakar en förändring av formen hos DntR på så sätt att avläsning av gener kan aktiveras. Vi har också använt en metod där evolutionära processer härmats för att få fram varianter av DntR med förbättrad respons till DNT. En variant med en drastisk ökning av DNT-responsen har isolerats, och dess egenskaper har karaktäriserats. / At the time of the doctoral defense, the following paper was unpublished and had a status as follows: Paper 3: Manuscript

Directed evolution is today an established strategy for generation of new affinity proteins. This thesis describes the development of a cell-display method using Escherichia coli for directed evolution of Affibody molecules. Further, the thesis describes rational design of Affibody-based tracers, intended for future patient stratification using medical imaging. Fusing recombinant proteins to various autotransporters is a promising approach for efficient surface display on the surface of E. coli, as well as for construction of high-complexity libraries. In paper I, we successfully engineered an expression vector for display of Affibody molecules using the autotransporter AIDA-I. In paper II, a large Affibody library of 2.3x109 variants was constructed and screening using FACS resulted in new specific binders in the nanomolar range. In paper III, we demonstrated Sortase-mediated secretion and conjugation of binders directly from the E. coli surface.  The three following studies describe rational design of Affibody-based tracers against two cancer-associated targets for molecular imaging. First, anti-HER3 Affibody molecules were labelled with 111In, and SPECT imaging showed that the conjugates specifically targeted HER3-expressing xenografts. Furthermore, labeling with 68Ga for PET imaging showed that tumor uptake correlated with HER3 expression, suggesting that the tracers have potential for patient stratification. The last study describes the development and investigation of anti-EGFR Affibody-based imaging agents. Labeled with 89Zr, the Affibody tracer demonstrated higher tumor uptake at 3 h post injection than the anti-EGFR antibody cetuximab at 48 h post injection.  In conclusion, this thesis describes new tools and knowledge that will hopefully contribute to the development of affinity proteins for biotechnology, therapy and medical imaging in the future. / <p>QC 20170904</p>

The Cre/loxP system belongs to the family of site-specific recombinases (SSR) that can precisely modify DNA that is flanked by two target sites. The reaction outcome is dependent on the structure and orientation of the target sites and includes excision, inversion and exchange of a DNA fragment. The system is established for more than 30 years and is active in vitro and in vivo in several organisms. These characteristics make the Cre/loxP system the ideal tool for genome editing. However, the strict target site preference for loxP limits its use to basic research where the loxP target sites can be introduced beforehand at the anticipated genomic locus. Directed evolution strategies have overcome this limitation and allow to generate Cre-like recombinases with altered DNA specificity. During this work, I developed the first dual-recombinase system based on evolved recombinases. Using two instead of one recombinase expands the targetability of the human genome by being more flexible in the target site search. After the identification of suitable target sites, I could show an evolved dual-recombinase system that can be used for excision and inversion of a human genomic locus. The recombinase mediated inversion reaction corrected a large genomic inversion that is frequently found in patients with severe Hemophilia A. Only two days after treating human cells with the developed dual-recombinase system, RecF8, 30% inversion could be detected in a human cell line. Applying RecF8 in patient specific endothelial cells corrected around 9% of the inversion back to the wild type sequence, which would be sufficient to drastically improve the quality of life of affected individuals. This genomic correction lead to the expression of the F8 gene, which is inactive elsewise. It remains to proof that the transcriptional reactivation of the F8 gene allows for the production of the Factor VIII protein. Before using RecF8 in a clinical setting, an in vivo study in a suitable mouse model is necessary. This study introduces a dual-recombinase system and thereby broadens the use of designer recombinases for genome editing. Moreover, in a proof on concept experiment this study shows that recombinases can be applied to correct large disease-causing genomic inversions in human cells. Altogether, the use of recombinases for scarless genome editing comes a step closer to reality.

Thesis advisor: Abhishek Chatterjee / The emergence of genetic code expansion (GCE) technology, which enables sitespecific incorporation of unnatural amino acids (UAAs) into proteins, has facilitated powerful new ways to probe and engineer protein structure and function. Using engineered orthogonal tRNA/aminoacyl-tRNA synthetase (aaRS) pairs that suppress repurposed nonsense codons, a variety of structurally diverse UAAs have been incorporated into proteins in living cells. This technology offers tremendous potential for deciphering the complex biology of eukaryotes, but its scope in eukaryotic systems remains restricted due to several technical limitations. For example, development of the engineered tRNA/aaRS pairs for eukaryotic GCE traditionally relied on a eukaryotic cell-based directed evolution system, which are significantly less efficient relative to bacteria-based engineering platforms. The work described in this thesis establishes a new paradigm in GCE through the development of a novel class of universal tRNA/aaRS pairs, which can be used for ncAA incorporation in both E. coli and eukaryotes. We achieve this by developing engineered strains of E. coli, where one of its endogenous tRNA/aaRS pair is functionally replaced with an evolutionarily distant counterpart. The liberated pair can then be used for GCE in the resulting altered translational machinery (ATM) strain, as well as any eukaryote. Using this strategy, we have been able to genetically encode new bioconjugation chemistries, post-translational modifications, and facilitate the incorporation of multiple, distinct ncAAs into a single protein. The ATM technology holds enormous promise for significantly expanding the scope of the GCE technology in both bacteria and eukaryotes. / Thesis (PhD) — Boston College, 2019. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Chemistry.

<p>Synthetic biology reprograms organisms to perform non-native functions for beneficial reasons. An important practice in synthetic biology is the ability to edit DNA to change a base pair, disrupt a gene, or insert a new DNA sequence. DNA edits are commonly made with the help of homologous recombination, which inserts new DNA flanked by sequences homologous to the target region. To increase homologous recombination efficiency, a double stranded break is needed in the middle of the target sequence. Common methods to induce double stranded breaks use nucleases, enzymes that cleave ribonucleotides (DNA and RNA). The most common nucleases are restriction enzymes, which recognize a short, fixed, palindromic DNA sequence (4-8 base pairs). Because of the short and fixed nature of the recognition sites, restriction enzymes do not make good candidates to edit large chromosomal DNA. Alternatively, scientists have turned to programmable endonucleases which recognize user-defined DNA sequences, often times much larger than the recognition sites of restriction enzymes (15-25 base pairs). Programmable endonucleases such as CRISPR-based systems and prokaryotic Argonautes are found throughout the prokaryotic kingdom and may differ significantly in activity and specificity. To compare activity levels among endonuclease enzymes, activity assays are needed. These assays must clearly delineate dynamic activity levels of different endonucleases and work with a wide variety of enzymes. Ideally, the activity assay will also function as a positive selection screen, allowing modifications to the enzymes via directed evolution. Here, we develop an <i>in vivo</i> assay for programmable endonuclease activity that can also serve as a positive selection screen using two plasmids, a lethal plasmid to cause cell death and a rescue plasmid to rescue cell growth. The lethal plasmid houses the homing endonuclease, I-SceI, which causes a deadly double-stranded break at an 18 base pair sequence inserted into an engineered <i>E. coli</i> genome. The rescue plasmid encodes for a chosen endonuclease designed to target and cleave the lethal plasmid, thereby preventing cell death. With this, cell growth is directly linked to programmable endonuclease activity. Three endonucleases were tested, SpCas9, eSpCas9, and xCas9, displaying recovered growth of 49.3%, 26.1%, and 16.4% respectively. These values translate to kinetic enzymatic activity and are congruent with current literature findings as reported values find WT-SpCas9 to have the fastest kinetics cleaving around 95% of substrate within 15 seconds, followed closely by eSpCas9 cleaving 75% of substrate within 15 seconds and finally trailed by xCas9 cleaving 20% of substrate in about 30 seconds. The differences between each endonuclease’s activity is exacerbated in our <i>in vivo</i> system when compared to similar <i>in vitro</i> methods with much lower resolution. Therefore, slight differences in activity between endonucleases within the first few minutes in an <i>in vitro</i> assay may be a few percentages different whereas in our <i>in vivo</i> assay, these differences in activity result in a more amplified signal. With the ability to display the dynamic response of enzymes, this assay can be used to compare activity levels between endonucleases, give insight into their kinetics, and serve as a positive selection screen for use in directed evolution applications. </p>

Prostate cancer is one of the most common cancer types worldwide. However, current diagnostic approaches and treatments are invasive and unspecific. Prostate-specific membrane antigen (PSMA) is an ideal biomarker for prostate cancer and can act as a target for therapeutic or diagnostic agents. Previous attempts to develop an affibody with affinity towards PSMA have been unsuccessful, therefore this thesis aimed at making the affibody selections against PSMA more efficient. In this thesis HEK293 cells expressing a modified version of PSMA containing a 3C protease cleavage site were generated, to enable extraction of the extracellular domain of PSMA during the selections. However, further analyses must be performed to determine if the extracellular domain can be successfully cleaved off. To develop an affibody that can be used both in vitro and in vivo, selections will be carried out against recombinant PSMA as well. The recombinant PSMA was previously produced incorporating an Avi tag for site-specific biotinylation and immobilization for the selections. To biotinylate the recombinant PSMA, the enzyme BirA that catalyzes the biotinylation of the Avi tag, was produced. A protein yield of 8.95 mg/liter culture was obtained and the site-specific biotinylation was highly efficient. To evaluate the proposed affibody selection strategy the next step is to determine if cleavage of the PSMA expressed on the HEK293 cells is possible, optimize the cleavage conditions and to start initial selections using the generated HEK293 cells and the produced BirA enzyme. / Prostatacancer är en av de mest förekommande cancertyperna över hela världen. Nuvarande diagnostiska metoder och terapeutiska behandlingar är dock invasiva och ospecifika. Prostataspecifikt membranantigen (PSMA) är en idealisk biomarkör för prostatacancer och kan agera som en målmolekyl för terapeutiska eller diagnostiska ändamål. Tidigare försök att utveckla en affibody med affinitet mot PSMA har inte lyckats, därför var målet med detta examensarbete att effektivisera selekteringen av affibodies mot PSMA. I detta projekt har HEK293 celler som uttrycker en modifierad version av PSMA, innehållande ett 3C-proteas- klyvningsställe, genererats för att möjliggöra extraktion av den extracellulära domänen av PSMA under selekteringen. Ytterligare analyser måste dock utföras för att avgöra om den extracellulära domänen kan klyvas av. För att utveckla en affibody som kan användas både in vitro och in vivo kommer selekteringen att utföras även mot rekombinant PSMA. Rekombinant PSMA har producerats tidigare med en Avi tag för specifik biotinylering och immobilisering under selekteringen. För att biotinylera det rekombinanta PSMA producerades enzymet BirA, som katalyserar biotinyleringen av en Avi tag. Ett proteinutbyte av 8,95 mg/liter kultur erhölls och den specifika biotinyleringen var effektiv. För att utvärdera den föreslagna strategin för selektering av affibodies är nästa steg att avgöra om klyvning av PSMA uttryckt av HEK293 cellerna är möjlig, optimera klyvningsförhållandena och starta initiala selektioner med de genererade HEK293-cellerna och det producerade BirA-enzymet.

This thesis concerns the development of optimisation algorithms to determine the relative co-location, (localisation), of a number of freely-flying 'Smart Dust mote' sensor platform elements using a non-deterministic data-set derived from the duplex wireless transmissions between elements. Smart dust motes are miniaturised, microprocessor based, electronic sensor platforms, frequently used for a wide range of remote environmental monitoring applications; including specific climate synoptic observation research and more general meteorology. For the application proposed in this thesis a cluster of the notional smart dust motes are configured to imitate discrete 'Radio Drop Sonde' elements of the wireless enabled monitoring system in use by meteorological research organisations worldwide. This cluster is modelled in software in order to establish the relative positions during the 'flight' ; the normal mode of deployment for the Drop Sonde is by ejection from an aeroplane into an upper-air zone of interest, such as a storm cloud. Therefore the underlying research question is, how to track a number of these independent, duplex wireless linked, free-flying monitoring devices in 3-dimensions and time (to give the monitored data complete spatio-temporal validity). This represents a significant practical challenge, the solution applied in this thesis was to generate 3-dimensional geometries using the only 'real-time' data available; the Radio Signal Strength Indicator (RSSI) data is generated through the 'normal' duplex wireless communications between motes. Individual RSSI values can be considered as a 'representation of the distance magnitude' between wireless devices; when collated into a spatio-temporal data-set it 'encodes' the relative, co-locational, 3-dimensional geometry of all devices in the cluster. The reconstruction, (or decoding), of the 3-dimensional geometries encoded in the spatio-temporal data-set is a complex problem that is addressed through the application of various algorithms. These include, Random Search, and optimisation algorithms, such as the Stochastic Hill-climber, and various forms of Evolutionary Algorithm. It was found that the performance of the geometric reconstruction could be improved through identification of salient aspects of the modelled environment, the result was heuristic operators. In general these led to a decrease in the time taken to reach a convergent solution or a reduction in the number of candidate search space solutions that must be considered. The software model written for this thesis has been implemented to generalise the fundamental characteristics of an optimisation algorithm and to incorporate them into a generic software framework; this then provides the common code to all model algorithms used.

Many different chemicals are used in the everyday life, like detergents and pharmaceuticals. However, their production has a big impact on health and environment as much of the raw materials are not renewable and the standard ways of production in many cases includes toxic and environmentally hazardous components. As the population and as the life standard increases all over the planet, the demand for different important chemicals, like pharmaceuticals, will increase. A way to handle this is to apply the concept of Green chemistry, where biocatalysis, in the form of enzymes, is a very good alternative. Enzymes do not normally function in industrial processes and needs modifications through protein engineering to cope in such conditions. To be able to efficiently improve an enzyme, there is a need to understand the mechanism and characteristics of that enzyme. Acyloins (α-hydroxy ketones) are important building blocks in the synthesis of pharmaceuticals. In this thesis, the enzyme alcohol dehydrogenase A (ADH-A) from Rhodococcus ruber has been in focus, as it has been shown to display a wide substrate scope, also accepting aryl-substituted alcohols. The aim has been to study the usefulness of ADH-A as a biocatalyst towards production of acyloins and its activity with aryl-substituted vicinal diols and to study substrate-, regio-, and enantioselectivity of this enzyme. This thesis is based on four different papers where the focus of the first has been to biochemically characterize ADH-A and determine its mechanism, kinetics and its substrate-, regio-, and enantioselectivity. The second and third paper aims towards deeper understanding of some aspects of selectivity of ADH-A. Non-productive binding and its importance for enantioselectivity is studied in the second paper by evolving ADH-A towards increased activity with the least favored enantiomer through protein engineering. In the third paper, regioselectivity is in focus, where an evolved variant displaying reversed regioselectivity is studied. In the fourth and last paper ADH-A is studied towards the possibility to increase its activity towards aryl-substituted vicinal diols, with R-1-phenyl ethane-1,2-diol as the model substrate, and the possibility to link ADH-A with an epoxide hydrolase to produce acyloins from racemic epoxides.

Par l'accumulation de mutations bénéfiques lors de cycles successifs de mutagénèse, l'évolution dirigée offre un cadre rationnel pour l'amélioration des protéines à vocation industrielle. Elle permet une exploration large de l'espace possible des séquences ainsi que leurs capacités fonctionnelles. Elle est cependant lourde à mettre en oeuvre et nécessite des moyens importants. Des approches in silico font usage d'un jeu minimal de données expérimentales et utilisent la modélisation statistique combinée à des algorithmes d'apprentissage machine. Elles ont été développées pour explorer de façon heuristique l'espace possible des séquences et de la fitness et d'identifier les mutations et interactions entre résidus les plus intéressantes. C'est l'objet de cette thèse qui explore la construction et l'application de modèles statistiques s'appuyant sur des jeux minimaux de données expérimentales pour relier fitness, ou activité, à la séquence biologique des variants. L'étude s'articule autour d'un choix crucial d'une méthode de numérisation, de descripteurs de la séquence et de méthodes de régression. La méthode ProSAR de R. Fox (2005) et les limites de son applicabilité sur des jeux de données expérimentales ont été étudiées. De nouvelles méthodes ont aussi été développées, prenant en compte les propriétés physico-chimiques des acides aminés et leurs périodicités. Elle a permis de découvrir de nouveaux descripteurs reliant la séquence à l'activité et propose des approches innovantes qui ont la capacité de traiter des cadres biologiques très divers, même lorsque peu de données biologiques sont disponibles. / Via the accumulation of beneficial mutations through successive rounds of mutations, directed evolution offers a rational framework for the amelioration of protein of industrial interest. It enables the large exploration of the sequence space and fitness. However, they are wet-lab intensive and may reveal to be time consuming and costly. In silico approaches using minimal sets of experimental data and statistical models combined with machine learning algorithms have been developed to explore heuristically the sequence space and to identify the effect of the potential epistatic interactions between residues on protein fitness. This work focused on the construction and application of statistical models relying on minimal experimental datasets to study protein sequence to activity relationships (ProSAR). In particular, the choices of appropriate numerical encoding methods, of descriptors extracted from protein sequences and of regression methods were investigated. The original ProSAR method from R. Fox (2005) and the limits of its applicability on experimental datasets have been studied. New methods that consider physico-chemical features of amino acids and their periodicities have been explored. This study unveils novel descriptors of the sequence-activity relationship and provides innovative approaches that can deal with very diverse biological datasets, even when few biological data are available.

Élargir le répertoire de composés accessibles dans le domaine des Glycosciences est d’un intérêt majeur pour la communauté des biologistes du fait que ces composés, oligosaccharides et glyco-conjugués, sont impliqués dans diverses fonctions biologiques, aussi bien au niveau structurel, qu’énergétique voire même signalétique jouant un rôle primordial dans les interactions inter- ou intracellulaires. L’assemblage, la modification ou la déconstruction de ces glyco-structures complexes est possible grâce à l’action d’enzymes, parmi lesquelles l’on retrouve les CAZymes (Carbohydrate Active enZymes). Ces enzymes font partie du répertoire de la base de données CAZy, incluant les Glycoside Hydrolases (GHs) qui représentent le groupe le plus important et ayant pour fonction biologique principale l’hydrolyse des liens glycosidiques. Cependant, un certain nombre de GHs possède aussi la capacité de catalyser des réactions de synthèse (transglycosylation) en tant qu’activité secondaire mineure, voire en tant qu’activité principale pour un nombre restreint d’entre elles, qui sont alors appelées transglycosylases. Sachant que ces deux types de comportements peuvent se retrouver au sein d’une même famille de GH (donc étroitement liés sur le plan évolutif), la découverte et la compréhension des déterminants moléculaires qui ont été développés par les GHs au cours de leur évolution pour permettre cette partition d’activité, entre hydrolyse et transglycosylation, est d’une importance capitale pour le domaine de la synthèse chimio-enzymatique et des Glycosciences de manière plus générale.Ce travail de thèse décrit une proposition de synthèse pour apporter une réponse à cette question fondamentale via une revue critique de la littérature sur le sujet. Sur le plan expérimental, a été réalisée l’évolution moléculaire d’une enzyme spécifique des pentoses, l’α-L-arabinofuranosidase de Thermobacillus xylanilyticus (TxAbf) de la famille GH51, vers les premières transarabinofuranosylases de type ‘non-Leloir’. Cette évolution itérative a été développée en utilisant un panel d’outils d’ingénierie enzymatique combinant des approches aléatoire, semi-rationnelle, de prédiction in silico suivie de recombinaison dans un processus d’évolution dirigée global. Une analyse fine des mutants générés sur le plan mécanistique en lien avec la partition hydrolyse/transglycosylation mène à des conclusions en accord avec la proposition de synthèse issue de la revue de la littérature sur le sujet. Sur un plan plus appliqué, ces nouveaux biocatalyseurs ont ensuite été mis en oeuvre dans des voies de synthèse chimio-enzymatiques pour la préparation de composés furanosylés de structure contrôlée. Le transfert d’L-arabinofuranosyles permet la génération d’arabinoxylo-oligosaccharides (AXOS) ainsi que la conception d’oligosaccharides non naturels, tel que des galactofuranoxylo-oligosaccharides ou des arabinofuranogluco-oligosaccharides. Dans son ensemble, ce travail de recherche constitue les premières étapes clés du développement de méthodes de synthèse chimio-enzymatique plus élaborées pour la conception d’arabinoxylanes artificiels. Dans le contexte actuel de transition vers une bio-économie, reposant sur des concepts tels que ceux de la bioraffinerie ou de la chimie verte, nous espérons que les outils de glycosynthèse développés au cours de ces travaux trouveront leur application dans la valorisation des pentoses issus de la biomasse. La synthèse à-façon d’arabinoxylooligo- et polysaccharides présente nombre de valorisations possibles allant de la préparation de prébiotiques à la conception de matériaux bio-inspirés en passant par la synthèse de modèles de parois végétales. / Widening the spectrum of available compounds in the field of Glycosciences is of utmost importance for the entire biology community, because carbohydrates are determinants of a myriad of life-sustaining or threatening processes. The assembly, modification or deconstruction of complex carbohydrate-based structures mainly involves the action of enzymes, among which one can identify Carbohydrate Active enZymes (CAZymes). These enzymes form part of the CAZy database repertoire and include Glycoside Hydrolases (GHs), which are the biggest group of CAZymes, whose main role is to hydrolyze glycosidic linkages. However, some GHs also display the ability to perform synthesis (transglycosylation), an activity that mostly manifests itself as a minor one alongside hydrolysis, but which is the only activity displayed by a rather select group of GHs that are often called transglycosylases. Understanding how transglycosylases have resulted from the process of evolution is both intringuing and crucial, because it holds the key to the creation of tailored glycosynthetic enzymes that will revolutionize the field of glycosciences.In this thesis, an extensive review of relevant scientific literature that treats the different aspects of GH-catalyzed transglycosylation and glycosynthesis is presented, along with experimental results of work that has been performed on a family GH-51 α-L-arabinofuranosidase, a pentose-acting enzyme from Thermobacillus xylanilyticus (TxAbf). The conclusions of the literature are presented in the form of a hypothesis, which describes the molecular basis of the hydrolysis/transglycosylation partition and thus provides a proposal on how to engineer dominant transglycosylation activity in a GH. Afterwards, using a directed evolution approach, including random mutagenesis, semi-rational approaches, in silico predictions and recombination it has been experimentally possible to create the very first ‘non-Leloir’ transarabinofuranosylases. The mechanistic analysis of the resultant TxAbf mutants notably focusing on the hydrolysis/transglycosylation partition reveals that the results obtained are consistent with the initial hypothesis that was formulated on the basis of the literature review.To demonstrate the applicative value of the experimental work performed in this study, the TxAbf mutants were used to develop a chemo-enzymatic methodology that has procured a panel of well-defined furanosylated compounds. Enzyme-catalyzed transfer of arabinofuranosyl moities can be used to generate arabinoxylo-oligosaccharides (AXOS), but the design of non-natural oligosaccharides, such as galactofuranoxylo-oligosaccharides or arabinofuranogluco-oligosaccharides is also possible. Overall, the work presented constitutes the first steps towards the development of more sophiscated methodologies that will procure the means to synthesize artificial arabinoxylans, with a first proof of concept being presented at the very end of this manuscript.In the present context of the bioeconomy transition, which relies on technologies such as biorefining and green chemistry, it is expected that the glycosynthetic tools that have been developed in this work will be useful for the conversion of pentose sugars obtained from biomass. The synthesis of tailor-made arabinoxylo-oligo- and polysaccharides may lead to a variety of potential applications including the production of prebiotics, surfactants or bio-inspired materials and, more fundamentally, the synthesis of artificial models of plant cell wall.