Nonlinear state estimation as tool for online monitoring and adaptive feed in high throughput cultivations.

Robotic facilities that can perform advanced cultivations (e.g., fed-batch or continuous) in high throughput have drastically increased the speed and reliability of the bioprocess development pipeline. Still, developing reliable analytical technologies, that can cope with the throughput of the cultivation system, has proven to be very challenging. On the one hand, the analytical accuracy suffers from the low sampling volumes, and on the other hand, the number of samples that must be treated rapidly is very large. These issues have been a major limitation for the implementation of feedback control methods in miniaturized bioreactor systems, where observations of the process states are typically obtained after the experiment has finished. In this work, we implement a Sigma-Point Kalman Filter in a high throughput platform with 24 parallel experiments at the mL-scale to demonstrate its viability and added value in high throughput experiments. The filter exploits the information generated by the ammonia-based pH control to enable the continuous estimation of the biomass concentration, a critical state to monitor the specific rates of production and consumption in the process. The objective in the selected case study is to ensure that the selected specific substrate consumption rate is tightly controlled throughout the complete Escherichia coli cultivations for recombinant production of an antibody fragment.

Vesicle transport and growth dynamics in Aspergillus niger: Microscale modeling of secretory vesicle flow and centerline extraction from confocal fluorescent data.

In this paper, we present a mathematical model to describe filamentous fungal growth based on intracellular secretory vesicles (SVs), which transport cell wall components to the hyphal tip. Vesicular transport inside elongating hyphae is modeled as an advection-diffusion-reaction equation with a moving boundary, transformed into fixed coordinates, and discretized using a high-order weighted essentially nonoscillatory discretization scheme. The model describes the production and the consumption of SVs with kinetic functions. Simulations are subsequently compared against distributions of SVs visualized by enhanced green fluorescent protein in young Aspergillus niger hyphae after germination. Intensity profile data are obtained using an algorithm scripted in ImageJ that extracts mean intensity distributions from 3D time-lapse confocal measurement data. Simulated length growth is in good agreement with the experimental data. Our simulations further show that a decrease of effective vesicle transport velocity towards the tip can explain the observed tip accumulation of SVs.

Flexible automation with compact NMR spectroscopy for continuous production of pharmaceuticals.

Modular plants using intensified continuous processes represent an appealing concept for the production of pharmaceuticals. It can improve quality, safety, sustainability, and profitability compared to batch processes; besides, it enables plug-and-produce reconfiguration for fast product changes. To facilitate this flexibility by real-time quality control, we developed a solution that can be adapted quickly to new processes and is based on a compact nuclear magnetic resonance (NMR) spectrometer. The NMR sensor is a benchtop device enhanced to the requirements of automated chemical production including robust evaluation of sensor data. Beyond monitoring the product quality, online NMR data was used in a new iterative optimization approach to maximize the plant profit and served as a reliable reference for the calibration of a near-infrared (NIR) spectrometer. The overall approach was demonstrated on a commercial-scale pilot plant using a metal-organic reaction with pharmaceutical relevance. Graphical abstract.

Online low-field NMR spectroscopy for process control of an industrial lithiation reaction-automated data analysis.

Monitoring specific chemical properties is the key to chemical process control. Today, mainly optical online methods are applied, which require time- and cost-intensive calibration effort. NMR spectroscopy, with its advantage being a direct comparison method without need for calibration, has a high potential for enabling closed-loop process control while exhibiting short set-up times. Compact NMR instruments make NMR spectroscopy accessible in industrial and rough environments for process monitoring and advanced process control strategies. We present a fully automated data analysis approach which is completely based on physically motivated spectral models as first principles information (indirect hard modeling-IHM) and applied it to a given pharmaceutical lithiation reaction in the framework of the European Union's Horizon 2020 project CONSENS. Online low-field NMR (LF NMR) data was analyzed by IHM with low calibration effort, compared to a multivariate PLS-R (partial least squares regression) approach, and both validated using online high-field NMR (HF NMR) spectroscopy. Graphical abstract NMR sensor module for monitoring of the aromatic coupling of 1-fluoro-2-nitrobenzene (FNB) with aniline to 2-nitrodiphenylamine (NDPA) using lithium-bis(trimethylsilyl) amide (Li-HMDS) in continuous operation. Online 43.5 MHz low-field NMR (LF) was compared to 500 MHz high-field NMR spectroscopy (HF) as reference method.

Automatic identification of structured process models based on biological phenomena detected in (fed-)batch experiments.

In this paper, we present a set of methods to automatically propose structured process models from an automated analysis of (fed-)batch experiments. Therefore, the measurements are numerically compensated for the influence of feeding and sampling, and the qualitative behavior of the measurements is revealed. As measurements from fermentations are inherently noisy, we introduce a method that divides the compensated curves into several episodes in a probabilistic framework to better handle these shortcomings. The probability of biological phenomena that reveal crucial information about the underlying reaction network is calculated. Since the phenomena detection is measurement-driven, its reliability depends on the measurement situation, e.g., the number of samples taken and experiments considered, measurement noise, etc. We show a possible approach to test the uncertainty of the phenomena detection against these influences. Finally, model structures are proposed automatically based on the detected biological phenomena. An experimental validation of the approach is shown, using real fermentation data from fed-batch cultivations of Streptomyces tendae.

Secretory Vesicle and Glucoamylase Distribution in Aspergillus niger and Macromorphology in Regions of Varying Shear Stress.

In technical fermentations, filamentous microorganisms are exposed to different forms of mechanical stress, among which shear stress is prevalent in turbulent broths. Whereas small-scale bioreactors allow for realistic turbulent flow field conditions, they are not well-suited to investigate the fungal response to shear stress in more detail, as they only reveal the integral effect of a highly dynamic stress stimulus. Therefore, the widely used model system for producing constant, but rather low shear forces, the parallel plate flow chamber, is extended in this work by adding a backward-facing step (BFS). The BFS induces vortex shedding in the wake of the step and brings out distinct areas of different shear stress levels at the bottom of the chamber where mycelia grow. This allows for a stress-dependent analysis of growing cells using a confocal laser-scanning microscope. As the real stress cannot be measured in the experiment, the wall shear stress is estimated numerically using computational fluid dynamics (CFD). As a first application of the experimental setup, the relative biomass concentration, the relative amount of secretory vesicles and the relative amount of the chosen product glucoamylase produced by the filamentous fungus Aspergillus niger were measured. The obtained area scans show homogeneous mycelia growth in areas of low stress and cloud-like patterns downstream of the predicted flow reattachment length where high shear stress dominates. Quantitative analysis of the time course suggests that the amount of available secretory vesicles inside of A. niger decreases when the shear stress is increased, despite that no significant differences in biomass production could be found. In contrast, the highest level of glucoamylase was reached for intermediate volumetric flow rates, i.e., levels of shear stress.

Understanding and controlling filamentous growth of fungal cell factories: novel tools and opportunities for targeted morphology engineering.

Filamentous fungal cell factories are efficient producers of platform chemicals, proteins, enzymes and natural products. Stirred-tank bioreactors up to a scale of several hundred m³ are commonly used for their cultivation. Fungal hyphae self-assemble into various cellular macromorphologies ranging from dispersed mycelia, loose clumps, to compact pellets. Development of these macromorphologies is so far unpredictable but strongly impacts productivities of fungal bioprocesses. Depending on the strain and the desired product, the morphological forms vary, but no strain- or product-related correlations currently exist to improve process understanding of fungal production systems. However, novel genomic, genetic, metabolic, imaging and modelling tools have recently been established that will provide fundamental new insights into filamentous fungal growth and how it is balanced with product formation. In this primer, these tools will be highlighted and their revolutionary impact on rational morphology engineering and bioprocess control will be discussed.

Rapid process synthesis supported by a unified modular software framework.

Although known to be very powerful, the widespread application of model-based techniques is still significantly hampered in the area of bio-processes. Reasons for this situation can be found along the whole chain to set up and implement such approaches. In a time-consuming step, models are typically hand-crafted. Whether alternatives of better models exist to actually fulfill the final goals is undocumented, most often even unknown. In a next step, model-based process control methods are hand-coded in an error-prone procedure. For many of these methods given in the literature, only simulation studies are shown, leaving the interested reader with the unanswered question whether the implementation of a specific method in a real process is viable. As the potentially time-consuming implementation of such a method presents a risk for a rapid process development, promising candidates may be overlooked. To remediate this unsatisfactory situation, a combination of theoretical methods and information technology is proposed here. By an exemplarily realized software tool, it is shown how such an environment helps to promote model-based optimization, supervision, and control of bio-processes and allows for an inexpensive test of new ideas as well in real-life experiments. The contribution concentrates on an overview of a possible software architecture with respect to necessary methods and a meaningful information strategy, highlighting some of the more crucial building blocks. Experimental results exploiting parts of the proposed methods are given for a yeast strain synthesizing a product of industrial interest.

"Click" analytics for "click" chemistry - A simple method for calibration-free evaluation of online NMR spectra.

Driven mostly by the search for chemical syntheses under biocompatible conditions, so called "click" chemistry rapidly became a growing field of research. The resulting simple one-pot reactions are so far only scarcely accompanied by an adequate optimization via comparably straightforward and robust analysis techniques possessing short set-up times. Here, we report on a fast and reliable calibration-free online NMR monitoring approach for technical mixtures. It combines a versatile fluidic system, continuous-flow measurement of 1H spectra with a time interval of 20s per spectrum, and a robust, fully automated algorithm to interpret the obtained data. As a proof-of-concept, the thiol-ene coupling between N-boc cysteine methyl ester and allyl alcohol was conducted in a variety of non-deuterated solvents while its time-resolved behaviour was characterized with step tracer experiments. Overlapping signals in online spectra during thiol-ene coupling could be deconvoluted with a spectral model using indirect hard modeling and were subsequently converted to either molar ratios (using a calibration-free approach) or absolute concentrations (using 1-point calibration). For various solvents the kinetic constant k for pseudo-first order reaction was estimated to be 3.9h-1 at 25°C. The obtained results were compared with direct integration of non-overlapping signals and showed good agreement with the implemented mass balance.

A framework for an organelle-based mathematical modeling of hyphae.

BACKGROUND: Although highly desirable, a mechanistic explanation for the outstanding protein secretion capabilities of fungi such as Aspergilli is missing. As a result, a rational and predictive design of strains as cell factories for protein production is still out of reach. The analysis of the secretion apparatus is not only hampered by open issues concerning molecular cell biological processes, but as well by their spatial fragmentation and highly dynamic features. Whereas the former issues are addressed by many groups, an account of the space- and time-dependent processes, which is best done by means of mathematical models, is lacking. Up to now, mathematical models for hyphal organisms mainly focus on one of two extremes. Either macroscopic morphology, such as pellet or mycelium growth, is addressed, or a microscopic picture is drawn predicting, for instance, the form of a hyphal tip. How intra-hyphal transport and organelle distribution works, however, has not been tackled so far mathematically. RESULTS: The main result of this contribution is a generic modeling framework to describe the space- and time-dependent evolution of intracellular substances and organelles. It takes intrahyphal, passive and active transport of substances into account and explains exponential and then linear length growth by tugor-driven uptake of water. Experimentally observed increasing concentration levels of organelles towards the tip can be well explained within the framework without resorting to complex biological regulations. It is shown that the accumulation can be partly explained by geometrical constraints, besides a necessary deceleration of the active transport velocity. The model is formulated such that more intricate intracellular processes can be included. CONCLUSIONS: Results from steady-state experiments are easy to be interpreted. In a hyphal network, however, new branches are produced at an exponential rate. Moreover, passive and active transport processes give rise to a spatial distribution of organelles and other cytoplasmatic constituents inside hyphae. As a result, most of the data obtained in experiments will be from a non-steady and space dependent state. A quantitative and mechanistic explanation of the processes occurring will only be possible if these dependencies are taking into account while evaluating experimental findings.

The Cell Factory Aspergillus Enters the Big Data Era: Opportunities and Challenges for Optimising Product Formation.

Living with limits. Getting more from less. Producing commodities and high-value products from renewable resources including waste. What is the driving force and quintessence of bioeconomy outlines the lifestyle and product portfolio of Aspergillus, a saprophytic genus, to which some of the top-performing microbial cell factories belong: Aspergillus niger, Aspergillus oryzae and Aspergillus terreus. What makes them so interesting for exploitation in biotechnology and how can they help us to address key challenges of the twenty-first century? How can these strains become trimmed for better growth on second-generation feedstocks and how can we enlarge their product portfolio by genetic and metabolic engineering to get more from less? On the other hand, what makes it so challenging to deduce biological meaning from the wealth of Aspergillus -omics data? And which hurdles hinder us to model and engineer industrial strains for higher productivity and better rheological performance under industrial cultivation conditions? In this review, we will address these issues by highlighting most recent findings from the Aspergillus research with a focus on fungal growth, physiology, morphology and product formation. Indeed, the last years brought us many surprising insights into model and industrial strains. They clearly told us that similar is not the same: there are different ways to make a hypha, there are more protein secretion routes than anticipated and there are different molecular and physical mechanisms which control polar growth and the development of hyphal networks. We will discuss new conceptual frameworks derived from these insights and the future scientific advances necessary to create value from Aspergillus Big Data.

Toward Microbioreactor Arrays: A Slow-Responding Oxygen Sensor for Monitoring of Microbial Cultures in Standard 96-Well Plates.

In this study, a slow-responding chemo-optical sensor for dissolved oxygen (DO) integrated into a 96-well plate was developed. The slow response time ensures that the measured oxygen value does not change much during plate transport to the microplate reader. The sensor therefore permits at-line DO measurement of microbial cultures. Moreover, it eliminates the necessity of individual optical measurement systems for each culture plate, as many plates can be measured successively. Combined with the 96-well format, this increases the experimental throughput enormously. The novel sensor plate (Slow OxoPlate) consists of fluorophores suspended in a polymer matrix that were placed into u-bottom 96-well plates. Response time was measured using sodium sulfite, and a t90 value of 9.7 min was recorded. For application, DO values were then measured in Escherichia coli and Saccharomyces cerevisiae cultures grown under fed-batch-like conditions. Depending on the DO sensor's response time, different information on the oxygenation state of the culture plate was obtained: a fast sensor variant detects disturbance through sampling, whereas the slow sensor indicates oxygen limitation during incubation. A combination of the commercially available OxoPlate and the Slow OxoPlate enables operators of screening facilities to validate their cultivation procedures with regard to oxygen availability.

A multiparallel bioreactor for the cultivation of mammalian cells in a 3D-ceramic matrix.

For adherently growing cells, cultivation is limited by the provided growth surface. Excellent surface-to-volume ratios are found in highly porous matrices, which have to face the challenge of nutrient supply inside the matrices' caverns. Therefore, perfusion strategies are recommended which often have to deal with the need of developing an encompassing bioreactor periphery. We present a modular bioreactor system based on a porous ceramic matrix that enables the supply of cells with oxygen and nutrients by perfusion. The present version of the reactor system focuses on simple testing of various inoculation and operation modes. Moreover, it can be used to efficiently test different foam structures. Protocols are given to set-up the system together with handling procedures for long-time cultivation of a CHO cell line. Experimental results confirm vital growth of cells inside the matrices' caverns.