In the enzymatic production of antibiotics, membranes can be used for different purposes. Since the enzyme is a cost-determining factor, membrane bioreactors in which the ultrafiltration membrane’s sole purpose is retention of the ‘free’ enzyme and recycling it to the reaction vessel, were investigated as an alternative for the current immobilised-enzyme based process. The membrane bioreactor was tested for two reactions, the production of one of the substrates for the antibiotic synthesis, and the actual enzymatic synthesis reaction. The first process could be improved considerably if the pH regulation took place right after the membrane, therewith creating a small loop that prevented direct contact of the enzyme with high pH, for which it is very sensitive. For the latter process, a big advantage was that diffusion limitation, which reduces the yield of the process with immobilised enzyme, does not take place. Besides, options for continuous operation became available. Away from application during the enzymatic reactions, (nanofiltration) membranes were useful for down-stream processing of one of the substrates for antibiotic synthesis. Some of the tested membranes showed separation behaviour, which was strongly pH dependent, therewith opening new options for isolation of the desired component.

Van Roon onderzocht de productie van cefalexine: een semi-synthetisch antibioticum, veelvuldig voorgeschreven tegen bijvoorbeeld keelontsteking. Dit antibioticum wordt door DSM gemaakt met behulp van het enzym penicilline-G acylase. Teneinde hergebruik van het enzym te vergemakkelijken, immobiliseerde (verankerde) DSM deze in een gelbolletje, Assemblase® genaamd. Het nadeel van deze procedure is, dat er tijdens de synthesereactie relatief weinig cefalexine en relatief veel bijprodukt gevormd wordt. De oorzaak hiervan is dat de substraten (beginstoffen) en producten hinder ondervinden bij het in-, respectievelijk uitdiffunderen van het gelbolletje. Van Roon relateerde dit sub-optimale gedrag van Assemblase® aan een heterogene enzymverdeling, die hij met steeds hogere resolutie kwantitatief onderzocht met verschillende licht- en elektronenmicroscopische technieken. Het enzym bleek voornamelijk gelokaliseerd te zijn aan de buitenkant van het gelbolletje, maar bleek ook op sommige interne locaties in zeer hoge concentraties aanwezig te zijn. De oorzaak hiervan werd op hoge resolutie gezocht en gevonden in de eigenschappen van het dragermateriaal, dat lokaal heterogeen bleek in zowel samenstelling als structuur. Van de opgedane kennis werd een fysisch model geconstrueerd, dat het gedrag en de limitaties van Assemblase® tijdens de cefalexinesynthese goed kon beschrijven. Dit model dient als basis voor verdere procesoptimalisatie.cum laude graduation (with distinction)

Multiple phenomena are involved in conversions by immobilized biocatalysts. A paradox is identified between analytical desires on one hand and analytical boundary conditions on the other: while the study of interdependent phenomena would call for their simultaneous analysis in an integrated context, the available experimental options may impose a series of separate and dedicated analyses. From this analysis, bottlenecks in particle performance may be identified, if possible supported by a mechanistic model and performance criteria. Subsequently, a strategy for further biocatalyst development may be chosen. Finally, possibilities for future improvement of biocatalysts are discussed for various fields of research. Some examples of recent developments in enzyme and matrix characteristics, reactor operation, and micro-technology are discussed.

A physical model was derived for the synthesis of the antibiotic cephalexin with an industrial immobilized penicillin G acylase, called Assemblase. In reactions catalyzed by Assemblase, less product and more by-product are formed in comparison with a free-enzyme catalyzed reaction. The model incorporates reaction with a heterogeneous enzyme distribution, electrostatically coupled transport, and pH-dependent dissociation behavior of reactants and is used to obtain insight in the complex interplay between these individual processes leading to the suboptimal conversion. The model was successfully validated with synthesis experiments for conditions ranging from heavily diffusion limited to hardly diffusion limited, including substrate concentrations from 50 to 600 mM, temperatures between 273 and 303 K, and pH values between 6 and 9. During the conversion of the substrates into cephalexin, severe pH gradients inside the biocatalytic particle, which were previously measured by others, were predicted. Physical insight in such intraparticle process dynamics may give important clues for future biocatalyst design. The modular construction of the model may also facilitate its use for other bioconversions with other biocatalysts.

In a study of Assemblase®, an industrial immobilized penicillin-G acylase, various electron microscopic techniques were used to relate intra-particle enzyme heterogeneity with the morphological heterogeneity of the support material at various levels of detail. Transmission electron microscopy was used for the study of intra-particle penicillin-G acylase distribution in Assemblase® particles of various sizes; it revealed an abrupt increase in enzyme loading at the particle surface (1.4-fold) and in the areas (designated halo's) surrounding internal macro-voids (7.7-fold). Cryogenic field-emission scanning electron microscopy related these abrupt local enzyme heterogeneities to local heterogeneity of the support material by revealing the presence of dense top layers surrounding both the particle exterior and the internal macro-voids. Furthermore, it showed a very distinct morphological appearance of the halo. Most probably, all these regions contained relatively more chitosan than gelatin (the polymers Assemblase® was constructed of), which suggested local polymer demixing during particle production. A basic thermodynamic line of reasoning suggested that a difference in hydrophilicity between the two polymers induced local demixing. In the future, thermodynamic knowledge on such polymer interactions resulting in matrix heterogeneity may be used as a tool for biocatalyst design.

The quantitative intraparticle enzyme distribution of Assemblase, an industrially employed polydisperse immobilized penicillin-G acylase, was measured. Because of strong autofluorescence of the carrier, the generally applied technique of confocal scanning microscopy could not be used; light microscopy was our method of choice. To do so, Assemblase particles of various sizes were sectioned, labeled with antibodies specifically against the enzyme, and analyzed light microscopically. Image analysis software was developed and used to determine the intraparticle enzyme distribution, which was found to be heterogeneous, with most enzyme located in the outer regions of the particles. Larger particles showed steeper gradients than smaller ones. A mathematical representation of the intraparticle profiles, based on in-stationary enzyme diffusion into the particles, was validated successfully for a broad range of particle sizes using data for volume-averaged particle size and enzyme loading. The enzyme gradients determined in this work will be used as input for a physical model that quantitatively describes the complex behavior of Assemblase. Such a physical model will lead to identification of the current bottlenecks in Assemblase and can serve as a starting point for the design of improved biocatalysts that also may be based on intelligent use of enzyme gradients.

Field-emission scanning electron microscopy (FESEM) was used in a technical feasibility study to obtain insight into the internal morphology and the intraparticle enzyme distribution of Assemblase, an industrial biocatalytic particle containing immobilized penicillin-G acylase. The results were compared with previous studies based on light and transmission electron microscopic techniques. The integrated FESEM approach yielded the same quantitative results as the microscopic techniques used previously. Given this technical equivalence, the integrated approach offers several advantages. First, the single preparation method and detection system avoids interpretation discrepancies between corresponding areas that were examined for different properties with different detection techniques in different samples. Second, the specimen size suitable for whole particle study is virtually unlimited, which simplifies sectioning and puts less stringent demands on the embedding technique. Furthermore, the sensitivity toward enzyme presence and distribution increases because the epitopes inside thick sections become available for labeling. Quick and unambiguous analysis of the relation between particle morphology and enzyme distribution is important because this information may be used in the future for the design of enzyme distributions in which the particle morphology can be used as a control parameter.

The macroscopic kinetic behavior of an industrially employed immobilized penicillin‐G acylase, called Assemblase, formed the basis for a discussion on some simple intraparticle biocatalytic model distributions. Assemblase catalyzes the synthesis of the widely used semisynthetic antibiotic cephalexin. Despite the obvious advantages of immobilization, less cephalexin and more of the unwanted byproduct d‐(–)‐phenylglycine are obtained due to diffusional limitations when the immobilized enzyme is employed. To rationally optimize Assemblase, the parameters particle size, enzyme loading, and enzyme distribution, which severely determine the macroscopic particle performance, were studied on the basis of macroscopic observations. Laser diffraction measurements showed that the particle sizes in Assemblase vary as much as 100‐fold. The relative and total enzyme loadings in Assemblase and fractions thereof of different sizes were determined by initial‐rate d‐(–)‐phenylglycine amide hydrolysis, cephalexin synthesis experiments, and active‐site titration. These experiments revealed that the loading of penicillin‐G acylase in Assemblase was inversely correlated with the particle diameter. Apart from enzyme loadings, estimates on the intraparticle enzyme distribution came from cephalexin synthesis experiments, where mass‐transport limitations were present. Although this method cannot provide the level of detail of specific labeling experiments, it is simple, fast, and cheap. Within the set of simple model predictions, a heterogeneous enzyme distribution with most biocatalyst present in the outer region of the particle (within the outer 100 μm) gave the best description of the observed behavior, although no exact correlation was established. Highly detailed determination of intraparticle enzyme distributions must come from immunolabeling.

Research has shown that the intraparticle biocatalytic distribution has extensive effects on the properties of various (industrial) biocatalytic particles and their performance in (bio-) chemical reactions. In recent years, advances in molecular chemistry have led to the development of many different specific (immuno-) labeling and light-microscopic detection techniques. Furthermore, high-quality image-digitizing devices and enhanced computing power have made image analysis readily accessible. These technologies may lead to the assessment and improvement of the internal biocatalyst profile as an integral part of biocatalytic particle optimization.