Search

Sorptive bioprocesses are the basis for numerous biotechnological applications such as enzyme immobilization, biosensors, controlled drug delivery, water treatment, and molecular purification. Yet due to the complexity of these processes, their optimization is still time, labor, and cost‐intensive. This research presents a flexible self‐driving laboratory (SDL) designed for the accelerated development and optimization of solid‐phase extraction processes. As a use case, the SDL was used to optimize a DNA purification process using silica magnetic beads. Through the integration of robotics, machine learning, and data‐driven experimentation, the SDL demonstrates a highly accelerated process optimization with minimal human intervention. In the multistep purification approach, the system is able to optimize buffer compositions for DNA extraction from complex samples, demonstrating effectiveness in both conventional chaotropic salt‐based methods and innovative chaotropic salt‐free buffers. The study highlights the SDL's capability to autonomously refine process parameters, achieving significant enhancements in yield and purity of the product. This blueprint for future self‐driving optimization of bioprocess parameters showcases the potential of autonomous systems to revolutionize biochemical process development, offering insights into scalable, environmentally sustainable, and cost‐effective solutions.

Immunoglobulins are important building blocks in biology and biotechnology. With the emergence of comprehensive deep‐learning approaches, there are enormous opportunities for fast and accurate methods of classification of protein–protein interactions to arise. Herein, widely accessible image classification algorithms for species‐specific typification of a range of different immunoglobulin G (IgG) complexes are repurposed. Droplets of various immunoglobulins mixed with a B‐cell superantigen (SAg) (recombinant staphylococcal Protein A) are deposited onto hydrophobic polymer substrates and the resulting protein stains are imaged using polarized light microscopy. A comprehensive study based on 23 745 images finds that the pretrained convolutional neural network (CNN) InceptionV3 not only successfully categorizes IgGs from four different species but also predicts their binding affinity to Protein A: averaged over 36 binding pairs, the following are observed: 1) an overall accuracy of 81.4%, 2) the highest prediction accuracy for human IgG, the antibody with the highest binding affinity for Protein A, and 3) that the classification accuracy regarding the various IgG/Protein A ratios generally correlates with the binding strength of the protein–protein–complex as determined via circular dichroism spectroscopy. In addition, the CNN pretrained with IgG/Protein A stain images has been challenged with a new set of images using a different superantigen (SAg, Protein G). Despite the use of the unknown SAg, the CNN correctly classifies the images of human IgG and Protein G as indicated by a 94% accuracy over the various molar binding ratios. These findings are noteworthy because they demonstrate that appropriately pretrained CNNs can be used for the prediction of protein–protein interactions beyond the scope of the original training set. Aided by deep‐learning methods, simple stains of mixed protein solutions may serve as accurate predictors of the strength of protein–protein interactions with relevance to protein engineering, self‐aggregation, or protein stability in complex media.

Monoclonal antibodies (mAb) and other biological drugs are affected by enzymatic polysorbate (PS) degradation that reduces product stability and jeopardizes the supply of innovative medicines. PS represents a critical surfactant stabilizing the active pharmaceutical ingredients, which are produced by recombinant Chinese hamster ovary (CHO) cell lines. While the list of potential PS-degrading CHO host cell proteins (HCPs) has grown over the years, tangible data on industrially relevant HCPs are still scarce. By means of a highly sensitive liquid chromatography-tandem mass spectrometry method, we investigated seven different mAb products, resulting in the identification of 12 potentially PS-degrading hydrolases, including the strongly PS-degrading lipoprotein lipase (LPL). Using an LPL knockout CHO host cell line, we were able to stably overexpress and purify the remaining candidate hydrolases through orthogonal affinity chromatography methods, enabling their detailed functional characterization. Applying a PS degradation assay, we found nine mostly secreted, PS-active hydrolases with varying hydrolytic activity. All active hydrolases showed a serine-histidine-aspartate/glutamate catalytical triad. Further, we subjected the active hydrolases to pH-screenings and revealed a diverse range of activity optima, which can facilitate the identification of residual hydrolases during bioprocess development. Ultimately, we compiled our dataset in a risk matrix identifying PAF-AH, LIPA, PPT1, and LPLA2 as highly critical hydrolases based on their cellular expression, detection in purified antibodies, active secretion, and PS degradation activity. With this work, we pave the way toward a comprehensive functional characterization of PS-degrading hydrolases and provide a basis for a future reduction of PS degradation in biopharmaceutical drug products.

Developing scalable and accurate predictive analytical methods for the classification of protein‐DNA binding is critical for advancing our understanding of molecular biology, disease mechanisms, and a wide spectrum of biotechnological and medical applications. It is discovered that histone–DNA interactions can be stratified based on stain patterns created by the deposition of various nucleoprotein solutions onto a substrate. In this study, a deep‐learning neural network is applied to categorize polarized light microscopy images of drying droplet deposits originating from different histone–DNA mixtures. These DNA stain patterns featured high reproducibility across different species and thus enabled comprehensive DNA categorization (100% accuracy) and accurate prediction of their respective binding affinities to histones. Eukaryotic DNA, which has a higher binding affinity to mammalian histones than prokaryotic DNA, is associated with a higher overall prediction accuracy. For a given species, the average prediction accuracy increased with DNA size. To demonstrate generalizability, a pre‐trained CNN is challenged with unknown images that originated from DNA samples of species not included in the training set. The CNN classified these unknown histone‐DNA samples as either strong or medium binders with 84.4% and 96.25% accuracy, respectively.

Abstract 3D printing technologies have matured to produce complex structures, still they are often limited to static materials. Introducing alkoxyamine bonds into 3D printed structures offers unprecedented possibilities for post‐synthetic modification through nitroxide exchange reaction and nitroxide‐mediated polymerization. This study provides a comprehensive molecular and macroscopic characterization of 3D‐printed alkoxyamine‐containing dynamic covalent adaptable networks. The study provides new insights into their dynamic structural and mechanical alterations, making them promising candidates for advanced applications ranging from biomedical engineering to flexible electronics.

This study presents a quasi-online method for monitoring of dissolved volatile fatty acids (VFAs) in biogas fermentation processes with a carrier gas probe by use of thermo-cyclically operated metal oxide gas sensor arrays. Each of the two sensor arrays comprises a pure SnO2 and three different SnO2/additive-composites (additives: alumina, YSZ, NASICON) but differ by SnO2 synthesis routes, namely Flame Spray Pyrolysis (FSP) and Sol-Gel (SG) technique, respectively. This allowed comparative studies of the influence of layer morphology on VFA sensing characteristics. For sensitive determination of the dissolved VFAs besides high concentrations of biogas components like CO or CH4, first a pre-treatment routine of the fermentation sample was introduced to remove those physically dissolved gases without losing VFAs. The Conductance-over-Time-Profiles (CTPs) of eight different sensing layers were measured simultaneously at exposure to the gases extracted from the fermentation sample at different pH conditions. Almost all the investigated SnO2/additive-composites show CTP-features clearly correlating with the undissociated VFA even at concentrations below 120 ppm as referenced by GC-analysis. The lower detection limit is well below inhibitory concentration for fermentation processes. As expected, most pronounced CTPs representing actual VFAs situation were measured at pH 3, well below the pKa of the VFAs. The FSP-layers highlighted clearly better sensitivity and CTP specificity of higher quality compared to SG-layers. Among the SnO2/additives, the CTP-features of the SnO2(FSP)/NASICON and SnO2(SG)/NASICON layers showed the best specificity to acetic and propionic acid. For the first time, quasi-online analysis of VFAs using metal oxide gas sensors for early warning of VFA-development in biogas fermentation processes was demonstrated.

Abstract Magnetic properties of superparamagnetic iron oxide nanoparticles are controlled mainly by their particle size and by their particle size distribution. Magnetic properties of multi-core iron oxide nanoparticles, often called iron oxide nanoflowers (IONFs), are additionally affected by the interaction of magnetic moments between neighboring cores. The knowledge about the hierarchical structure of IONFs is therefore essential for understanding the magnetic properties of IONFs. In this contribution, the architecture of multi-core IONFs was investigated using correlative multiscale transmission electron microscopy (TEM), X-ray diffraction and dynamic light scattering. The multiscale TEM measurements comprised low-resolution and high-resolution imaging as well as geometric phase analysis. The IONFs contained maghemite with the average chemical composition $$\gamma$$ γ -Fe $$_{2.72\pm 0.02}$$ 2.72 ± 0.02 O $$_4$$ 4 . The metallic vacancies located on the octahedral lattice sites of the spinel ferrite structure were partially ordered. Individual IONFs consisted of several cores showing frequently a specific crystallographic orientation relationship between direct neighbors. This oriented attachment may facilitate the magnetic alignment within the cores. Individual cores were composed of partially coherent nanocrystals having almost the same crystallographic orientation. The sizes of individual constituents revealed by the microstructure analysis were correlated with the magnetic particle sizes that were obtained from fitting the measured magnetization curve by the Langevin function.

For a comparative study of the gas sensing performance to dissolved volatile organic compounds (VOCs), the SnO2-powders have been prepared using two different fabrication routes; the flame-spray-pyrolysis (FSP) and the sol-gel (SG) route; and were admixed with the same additive-powders (alumina, YSZ, and NASICON). The morphology of the two different SnO2/additive families were investigated by ESEM analysis and Energy-dispersive X-Ray Spectroscopy (EDS). Both SnO2/additive material families were separately deposited as thick-film-layers on two four-fold sensor-chips which were simultaneously thermo-cyclically operated in a measurement cell combined with a carrier gas probe, which enables sensing tests with evaporated VOCs (acetic acid, propionic acid, ethanol, acetone) dissolved in water. The resulting Conductance-over-Time-Profiles (CTPs) highlight better sensitivities of most of the FSP-layers to all the analytes compared to the SG-prepared layers. Furthermore, the CTP shapes of the FSP layers show clearly enhanced specificity representing the individual analyte components. This was interpreted to be the consequence of the extremely fine, scarcely agglomerated grain morphology of FSP-prepared powders and their very narrow grain size distribution which provide better conditions for enhanced gas specific surface reactions. Results promise a better chemical analysis capability of dissolved VOCs by numerical analysis of the CTP of FSP-prepared gas sensitive layers.

Dynamic crossflow filtration (DCF) is the state-of-the-art technology for solid–liquid separation from viscous and sensitive feed streams in the food and biopharma industry. Up to now, the potential of industrial processes is often not fully exploited, because fixed recipes are usually applied to run the processes. In order to take the varying properties of biological feed materials into account, we aim to develop a digital twin of an industrial brownfield DCF plant, allowing to optimize setpoint decisions in almost real time. The core of the digital twin is a mechanistic–empirical process model combining fundamental filtration laws with process expert knowledge. The effect of variation in the selected process and model parameters on plant productivity has been assessed using a model-based design-of-experiments approach, and a regression metamodel has been trained with the data. A cyclic program that bidirectionally communicates with the DCF asset serves as frame of the digital twin. It monitors the process dynamics membrane torque and transmembrane pressure and feeds back the optimum permeate flow rate setpoint to the physical asset in almost real-time during process runs. We considered a total of 24 industrial production batches from the filtration of grape juice from the years 2022 and 2023 in the study. After implementation of the digital twin on site, the campaign mean productivity increased by 15% over the course of the year 2023. The presented digital twin framework is a simple example how an industrial established process can be controlled by a hybrid model-based algorithm. With a digital process dynamics model at hand, the presented metamodel optimization approach can be easily transferred to other (bio)chemical processes.

Abstract One of the main steps in the biotechnological production of chemical building blocks, such as, e.g. bio‐based succinic acid which is used for lubricants, cosmetics, food, and pharmaceuticals, is the isolation and purification of the target molecule. A new approach to isolate charged, bio‐based chemicals is by electrosorption onto carbon surfaces. In contrast to ion exchange, electrosorption does not require additional chemicals for elution and regeneration. However, while the electrosorption of inorganic salts is well understood and in commercial use, the knowledge about electrosorption of weak organic acids including the strong implications of the pH‐dependent dissociation and their affinity towards physical adsorption must be expanded. Here, we show a detailed discussion of the main pH‐dependent effects determining the achievable charge efficiencies and capacities. An explicit set of equations allows the fast prediction of the named key figures for constant voltage and constant current operation. The calculated and experimental results obtained for the electrosorption of maleic acid show that the potential‐free adsorption of differently protonated forms of the organic acid play a dominating role in the process. At pH 8 and a voltage threshold of 1.3 V, charge efficiencies of 25% and capacities around 40 mmol/kg could be reached for a constant current experiment. While this capacity is clearly below that of ion exchange resins, the required carbon materials are inexpensive and energy costs are only about 0.013 €/mol. Therefore, we anticipate that electrosorption has the potential to become an interesting alternative to conventional unit operations for the isolation of charged target molecules.

Capacitive deionization (CDI) has become an important research topic in terms of water purification as it is a cost- and energy efficient approach for the desalination of brackish water. Carbon- based materials are often used as electrodes in this process due to a variety of morphologies and their suitable performance in this field. In recent studies it has been already shown that a conscientiously chosen set of electrode properties has an enormous effect on the performance behavior of CDI cells. However, most of the studies focus on the optimization of the electrosorption capacity, while the question on how material properties influence the kinetic behavior of a CDI setup has been less intensively studied so far. Here we show that the kinetic effects of electrode materials can be studied in great detail by using electrochemical impedance spectroscopy (EIS). EIS studies of CDI electrodes have been reported before, however, in contrast to these we introduce a method of presenting and analyzing EIS results. This is especially suitable for extracting the frequency ranges in which different rate limiting mechanisms dominate. The new Nyquist Incline Frequency plot (NIF) shows the local slope of the Nyquist plot in dependency of the applied frequency. By this, electron transfer and mass transfer mechanisms which show a characteristic slope in the classical Nyquist plot can be directly visualized together with the information in which time domain they occur. The data derived from the new EIS plot show a clear correlation to pore size distributions from BET measurements, as well as effective capacities and electrosorption kinetics, extracted from CDI experiments. Therefore, we think that by using the NIF plot electrode materials for CDI processes can be conveniently characterized and compared within a single diagram, helping to standardize CDI material development.

In biotechnology, immobilization of functional reactants is often done as a surface immobilization on small particles. Examples are chromatography columns and fixed-bed reactors. However, the available surface for immobilization is directly linked to particle diameter and bed porosity for these systems, leading to high backpressure for small particle sizes. When larger molecules, such as enzymes are immobilized, physical entrapment within porous materials like hydrogels is an alternative. An emerging technique for the production of geometrically structured, three-dimensional and scalable hollow bodies is 3D-printing. Different bioprinting methods are available to produce structures of the desired size, resolution and solids content. However, in case of entrapped enzymes mass transfer limitations often determine the achievable reactivities. With increasing complexity of the system, for example a fixed-bed reactor, 3D-simulation is indispensable to understand the local reaction conditions to be able to highlight the optimization potential. Based on experimental data, this manuscript shows the application of the dimensionless numbers effectiveness factor and Thiele modulus for the design of a 3D-printed flow-through reactor. Within the reactor, enzymes are physically entrapped in 3D-printed hydrogel lattices. The local reaction rate of the enzymes is directly dependent on the provided substrate amount at the site of reaction which is limited by the diffusion properties of the hydrogel matrix and the diffusion distance. All three parameters can be summed up by one key figure, the Thiele modulus, which, in short, quantifies mass transfer limitations of a catalytic system. Depending on the rate of the enzymatic reaction in correlation to the diffusional transport, mass transfer limitations will shift the optimum of the system, favoring slow enzyme kinetics and small diffusion distances. Comparison with the enzymatic reaction rate in solution yields the effectiveness factor of the system. As a result, the optimization potential of varying the 3D-printed geometries or the reaction rate within the experimentally available design space can be estimated.

In the last decade, the fabrication of microfluidic chips was revolutionized by 3D printing. It is not only used for rapid prototyping of molds, but also for manufacturing of complex chips and even integrated active parts like pumps and valves, which are essential for many microfluidic applications. The manufacturing of multiport injection valves is of special interest for analytical microfluidic systems, as they can reduce the injection to detection dead volume and thus enhance the resolution and decrease the detection limit. Designs reported so far use radial compression of rotor and stator. However, commercially available nonprinted valves usually feature axial compression, as this allows for adjustable compression and the possibility to integrate additional sealing elements. In this paper, we transfer the axial approach to 3D-printed valves and compare two different printing techniques, as well as six different sealing configurations. The tightness of the system is evaluated with optical examination, weighing, and flow measurements. The developed system shows similar performance to commercial or other 3D-printed valves with no measurable leakage for the static case and leakages below 0.5% in the dynamic case, can be turned automatically with a stepper motor, is easy to scale up, and is transferable to other printing methods and materials without design changes.

Owing to the hierarchical organization of biology, from genomes over transcriptomes and proteomes down to metabolomes, there is continuous debate about the extent to which data and interpretations derived from one level, e.g. the transcriptome, are in agreement with other levels, e.g. the metabolome. Here, we tested the effect of ocean acidification (OA; 400 vs. 1000 μatm CO2) and its modulation by light intensity (50 vs. 300 μmol photons m-2 s-1) on the biomass composition (represented by 75 key metabolites) of diploid and haploid life-cycle stages of the coccolithophore Emiliania huxleyi (RCC1216 and RCC1217) and compared these data with interpretations from previous physiological and gene expression screenings. The metabolite patterns showed minor responses to OA in both life-cycle stages. Whereas previous gene expression analyses suggested that the observed increased biomass buildup derived from lipid and carbohydrate storage, this dataset suggests that OA slightly increases overall biomass of cells, but does not significantly alter their metabolite composition. Generally, light was shown to be a more dominant driver of metabolite composition than OA, increasing the relative abundances of amino acids, mannitol and storage lipids, and shifting pigment contents to accommodate increased irradiance levels. The diploid stage was shown to contain vastly more osmolytes and mannitol than the haploid stage, which in turn had a higher relative content of amino acids, especially aromatic ones. Besides the differences between the investigated cell types and the general effects on biomass buildup, our analyses indicate that OA imposes only negligible effects on E. huxleyi´s biomass composition.

The development of process steps catalyzed by immobilized enzymes usually encompasses the screening of enzyme variants, as well as the optimization of immobilization protocols and process parameters. Direct immobilization of biocatalysts by physical entrapment into hydrogels can be applied to reduce the effort required for immobilization, as the enzyme-specific optimization of the immobilization procedure is omitted. Physical entrapment is applicable for purified enzymes as well as crude cell extracts. Therefore, it can be used to quickly assess and compare activities of immobilized enzymes. For the application in flow reactors, we developed 3D-printed hydrogel lattices for enzyme entrapment as well as matching housings, also manufactured by 3D-printing. Testing the resulting enzyme reactors for three different enzymes, namely alcohol dehydrogenase from Lactobacillus brevis, benzoylformate decarboxylase from Pseudomonas putida and β-galactosidase from Aspergillus oryzae, and four different enzymatic reactions showed the broad applicability of the approach but also its limitations. The activity of the immobilized biocatalysts was measured in batch experiments and compared to the kinetics of the respective free enzymes in solution. This comparison yields an effectiveness factor, which is a key figure to describe the extent the immobilized catalyst is effectively utilized. For the examined systems the effectiveness factor ranged between 6 and 14% and decreased with increasing absolute activity of the entrapped enzymes due to mass transfer limitations. To test the suitability of the hydrogel lattices for continuous operation, they were inserted into 3D-printed reactor housings and operated at constant flow. Stable product formation could be monitored over a period of 72 h for all four enzymatic systems, including two reactions with redox cofactor regeneration. Comparing calculated and experimental conversion in the continuous setup, higher values of the effectiveness factor in batch experiments also hint at good performance in continuous flow. This can be used to optimize complex biocatalytic reactions on a small scale.

In the course of their development, industrial biocatalysis processes have to be optimized in small-scale, e. g., within microfluidic bioreactors. Recently, we introduced a novel microfluidic reactor device, which can handle defined reaction compartments of up to 250 μL in combination with magnetic micro carriers. By transferring the magnetic carriers between subsequent compartments of differing compositions, small scale synthesis, and bioanalytical assays can be conducted. In the current work, this device is modified and extended to broaden its application range to the screening and optimization of bioprocesses applying immobilized enzymes. Besides scaling the maximum compartment volume up to 3 mL, a temperature control module, as well as a focused infrared spot were integrated. By adjusting the pump rate, compartment volumes can be accurately dosed with an error

Protein modification by covalent coupling of small ligands or markers is an important prerequisite for the use of proteins in many applications. Well-known examples are the use of proteins with fluorescent markers in many in vivo experiments or the binding of biotinylated antibodies via biotin–streptavidin coupling in the frame of numerous bioassays. Multiple protocols were established for the coupling of the respective molecules, e.g., via the C and N-terminus, or via cysteines and lysines exposed at the protein surface. Still, in most cases the conditions of these standard protocols are only an initial guess. Optimization of the coupling parameters like reagent concentrations, pH, or temperature may strongly increase coupling yield and the biological activity of the modified protein. In order to facilitate the process of optimizing coupling conditions, a method was developed which uses a compartmented microfluidic reactor for the rapid screening of different coupling conditions. In addition, the system allows for the integration of an enzymatic digest of the modified protein directly after modification. In combination with a subsequent MALDI-TOF analysis of the resulting fragments, this gives a fast and detailed picture not only of the number and extent of the generated modifications but also of their position within the protein sequence. The described process was demonstrated for biotinylation of green fluorescent protein. Different biotin-excesses and different pH-values were tested in order to elucidate the influence on the modification extent and pattern. In addition, the results of solid-phase based modifications within the microfluidic reactor were compared to modification patterns resulting from coupling trials with unbound protein. As expected, modification patterns of immobilized proteins showed clear differences to the ones of dissolved proteins.

Several health benefits, associated with human milk oligosaccharides (HMOS), have been revealed in the last decades. Further progress, however, requires not only the establishment of a simple “routine” method for absolute quantification of complex HMOS mixtures but also the development of novel synthesis strategies to improve access to tailored HMOS. Here, we introduce a combination of salvage-like nucleotide sugar-producing enzyme cascades with Leloir-glycosyltransferases in a sequential pattern for the convenient tailoring of stable isotope-labeled HMOS. We demonstrate the assembly of [13C6]galactose into lacto-N- and lacto-N-neo-type HMOS structures up to octaoses. Further, we present the enzymatic production of UDP-[15N]GlcNAc and its application for the enzymatic synthesis of [13C6/15N]lacto-N-neo-tetraose for the first time. An exemplary application was selected—analysis of tetraose in complex biological mixtures—to show the potential of tailored stable isotope reference standards for the mass spectrometry-based quantification, using matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF-MS) as a fast and straightforward method for absolute quantification of HMOS. Together with the newly available well-defined tailored isotopic HMOS, this can make a crucial contribution to prospective research aiming for a more profound understanding of HMOS structure-function relations.

Blends of biodiesels produced from soapnut oil and high-oleic free fatty acids (FFAs), which are potential non-edible oil feedstocks, were investigated with respect to their fuel properties. The soapnut oil methyl esters (SNME) had satisfactory fuel properties with the exception of its high cold filter plugging point. In contrast, the biodiesel from the FFAs had favorable fuel properties such as a low cold filter plugging point of −6 °C; however, it exhibits poor oxidation stability with an induction period (IP) of 0.2 h. The complementary blend of the SNME and the FFA-based biodiesel at various weight ratios was studied to improve the fuel properties. As a result, the biodiesel blend at a weight ratio of 70:30 can successfully meet all the biodiesel specifications, except the marginal oxidation stability. Furthermore, the effectiveness of N,N’-di-sec-butyl-p-phenylenediamine at the concentration between 100 and 500 ppm on the improvement in the oxidation stability of the biodiesel blend was examined. The relationship between the IP values associated with the consumption of antioxidants in the biodiesel blends was described by first-order reaction rate kinetics. In addition, the natural logarithm of IP (ln IP) at various concentrations of antioxidant presented a linear relation with the test temperature. The IP at ambient temperature can be predicted based on the extrapolation of the temperature dependence relation.