Structured modeling of a microbial system: I. A theoretical study of lactic acid fermentation

Phenomenological models as effective tools to discover cellular design principles

Microbes have proved useful to us in many different ways. To utilize microbes, we have mostly focused on maximizing growth, to improve yield of chemicals derived from the microbes. However, to truly tap into their potential, we should also aim to understand microbial physiology. We present a historical perspective of the developments in the field of Microbial Biotechnology, focusing on how the growth-modelling approaches have changed. Starting from simple empirical growth models, we have evolved towards mechanistic and phenomenological models which use molecular and physiological details to drastically improve prediction power of these models. Lastly, we explore the as of yet unsolved questions in microbial physiology, and discuss how the ability to monitor microbial growth at single cell resolution using the lab-on-a-chip technologies is uncovering previously unobservable causal principles underlying microbial growth.

In Silico Prediction of Large-Scale Microbial Production Performance: Constraints for Getting Proper Data-Driven Models

Industrial bioreactors range from 10.000 to 700.000 L and characteristically show different zones of substrate availabilities, dissolved gas concentrations and pH values reflecting physical, technical and economic constraints of scale-up. Microbial producers are fluctuating inside the bioreactors thereby experiencing frequently changing micro-environmental conditions. The external stimuli induce responses on microbial metabolism and on transcriptional regulation programs. Both may deteriorate the expected microbial production performance in large scale compared to expectations deduced from ideal, well-mixed lab-scale conditions. Accordingly, predictive tools are needed to quantify large-scale impacts considering bioreactor heterogeneities. The review shows that the time is right to combine simulations of microbial kinetics with calculations of large-scale environmental conditions to predict the bioreactor performance. Accordingly, basic experimental procedures and computational tools are presented to derive proper microbial models and hydrodynamic conditions, and to link both for bioreactor modeling. Particular emphasis is laid on the identification of gene regulatory networks as the implementation of such models will surely gain momentum in future studies.

Operator training in recombinant protein production using a structured simulator model

Model-based operator training simulators (OTS) could be powerful tools for virtual training of operational procedures and skills of production personnel in recombinant protein processes. The applied model should describe critical events in the bioprocess so accurately that the operators' ability to observe and alertly act upon these events is trained with a high degree of efficiency. In this work is shown how this is accomplished in a structured multi-compartment model for the production of a recombinant protein in an Escherichia coli fed-batch process where in particular the induction procedure, the stress effects and overflow metabolism were highlighted. The structured model was applied on the OTS platform that virtually simulated the operational bioreactor procedures in real or accelerated time. Evaluation of training using the model-based OTS showed that trained groups of operators exhibited improved capability compared with the untrained groups when subsequently performing real laboratory scale cultivations. The results suggest that this model-based OTS may provide a valuable resource for enhancing operator skills in large scale recombinant protein manufacturing.

Application of statistical techniques for elucidating flow cytometric data of batch and fed-batch cultures

The objective of this work is to develop structured, segregated stochastic models for bioprocesses using time-series flow cytometric (FC) data. To this end, mammalian CHO cells were grown in both batch and fed-batch cultures, and their viable cell numbers (VCDs), monoclonal antibody (MAb), cell cycle phases, mitochondria membrane potential/mitochondria mass, Golgi apparatus, and endoplasmic reticulum (ER) were analyzed. For the fed-batch mode, soy hydrolysate was introduced at 24-H intervals. The cytometric data were analyzed for early indicators of growth and productivity by multiple linear regression analysis, which involved taking into account multicollinearity diagnostics, Durbin-Watson statistics, and Houston tests to determine and refine statistically significant correlations between categorical variables (FC parameters) and response variables (yield parameters). The results indicate that the percentage of G1 cells and ER was significantly correlated with VCD and MAb in the case of batch culture, whereas for fed-batch culture, the percentage of G2 cells and ER was correlated significantly. There was a significant difference between cells in the batch and fed-batch cultures in their ER content, suggesting that the increase in protein synthesis as reflected by the ER content and consequent increase in growth rate and MAb productivity both can be monitored at the cellular level by FC analysis of ER content.

Prediction of dynamic metabolic behavior of Pediococcus pentosaceus producing lactic acid from lignocellulosic sugars

A dynamic metabolic model is presented for Pediococcus pentosaceus producing lactic acid from lignocellulose-derived mixed sugars including glucose, mannose, galactose, arabinose, and xylose. Depending on the pairs of mixed sugars, P. pentosaceus exhibits diverse (i.e., sequential, simultaneous or mixed) consumption patterns. This regulatory behavior of P. pentosaceus is portrayed using the hybrid cybernetic model (HCM) framework which views elementary modes of the network as metabolic options dynamically modulated. Comprehensive data are collected for model identification and validation through fermentation experiments involving single substrates and various combinations of mixed sugars. Most sugars are metabolized rather sequentially while co-consumption of galactose and arabinose is observed. It is demonstrated that the developed HCM successfully predicts mixed sugar data based on the parameters identified mostly from single substrate data only. Further, we discuss the potential of HCMs as a tool for predicting intracellular flux distribution with comparison with flux balance analysis.

Structured modeling of a microbial system: II. Experimental verification of a structured lactic acid fermentation model

A two-compartment model for the lactic acid fermentation with Streptococcus cremoris is experimentally verified. The seven parameters of the model are determined using steady-state chemostat data at varying values of dilution rate, D, but with a constant feed concentration, s(f), of a single carbohydrate source (glucose, lactose, or galactose), and a constant feed concentration of s(Nf) of the N source. Steady-state measurements of the RNA content at different exit concentrations, s, of the carbohydrate are included to calculate kinetic parameters that determine the cell composition for varying operating conditions. The model is tested using data from a large set of steady-state and non-steady-state experiments: batch fermentations and step and pulse experiments in a chemostat. Both qualitatively and quantitatively the major features of the model are confirmed: the external substrates enter into intracellular high-energy building blocks, and lactic acid is formed as a by-product of these reactions. Cell growth depends on the fraction of active components (X(A)) of the cell and is not accompanied by lactic acid production. Possible model modifications are discussed, primarily to obtain a better description of lactic acid fermentation at nongrowth conditions.

Modelling the growth of filamentous fungi

Despite the considerable industrial importance of filamentous fungi there have been very few attempts to model the complex growth process of these microorganisms. With a new generation of high performance, computerized bioreactors and new analytical techniques it is possible to obtain the necessary experimental data for setting up reliable structured models describing the growth process of filamentous fungi. It is therefore interesting to review the mathematical models described previously in the literature and the experimental data on which these models are built. Only structured models are considered due to the complex metabolism of filamentous fungi and to the natural cellular structuring of the biomass, i.e. the biomass can be divided into different cell types. In order to set up good structured models it is strictly necessary to have a detailed knowledge of the mechanisms underlying the growth process. This involves both biochemical insight and understanding of the interactions between different macromolecules and cytological organelles.

Batch cultures of supplemented whey permeate using Lactobacillus helveticus: unstructured model for biomass formation, substrate consumption and lactic acid production

The Luedeking and Piret expression can not account for the cessation of production observed at the end of batch; so an empiric term has been previously added to this equation which accounted in a global way for possible substrate limitations. In the model developed in this work, a carbon substrate limitation appeared explicitly in the production expression. Assuming a sigmoidal variation with time of specific growth rate previously validated, the new production model matched well the entire experimental production kinetics. It has been successfully tested for a wide range of nitrogen supplementations, i.e. from an almost total coupling between growth and production for largely supplemented media, to a high decoupling in case of few available nitrogen. Since all the parameters of this model have an obvious biologic meaning, it may be an unvaluable tool for the comprehension of the phenomenon. The model accounted also well for the variation of the specific production rate versus specific growth rate, avoiding the noise due to the direct differentiation of experimental data.

Instrumentation of biotechnological processes

Modern bioprocesses are monitored by on-line sensing devices mounted either in situ or externally. In addition to sensor probes, more and more analytical subsystems are being exploited to monitor the state of a bioprocess on-line and in real time. Some of these subsystems deliver signals that are useful for documentation only, other, less delayed systems generate signals useful for closed loop process control. Various conventional and non-conventional monitoring instruments are evaluated; their usefulness, benefits and associated pitfalls are discussed.

An empirical model on extractive lactic acid bioconversion

The commercial production of lactic acid through fermentation process has always been in competition with its chemical synthesis process (Kirk Othmer, 1995). Lactic acid produced through the fermentation process has to cope with the problems of purification to meet the required quality standards. An attempt to improve the fermentative production is possible by proper design of an industrial process involving low capital cost for the plant. Also, the low energy costs both in its fermentation and purification, are required. In the commercial interest, the investment cost should be minimised, which is possible only when the cell density in fermenter is high. It means that the inhibitory effect of the product on process kinetics must be minimised. Based on these requirements, the extractive bioconversion technique is one of the approaches to achieve the commercially viable lactic acid production. Extractive lactic acid bioconversion using ion-exchange resin process has already been described in our earlier publications (Srivastava e al., 1992: Roychoudhury et al., 1995) It is always an advantage to develop a process model, thus opening an area of biotechnological improvements to the process. In the present paper, an empirical mathematical model has been described to explain this extractive bioconversion using ion-exchange resin process. It was based on generalised Monod's growth model and Leudeking and Piret equation. The system was defined with the assumption that the microbial growth can be represented as a single reaction; only a very little part of the substrate is utilised for the maintenance of the cells. The effect of end product inhibition on growth and product formation kinetics has also been considered in this model. A non-linear regression technique was used for evaluation of bioconversion kinetic parameters. The fourth order Runge Kutta method was used for solving the differential equations. The results of this process simulation are also discussed in the present paper. It indicates that the use of present technique has minimised the effect of lactic acid inhibition on process kinetics and hence higher productivity and least substrate utilisation for maintenance of cells. A statistical F-test has been performed for determining the validity of the model for a given set of experimental data with a level of significance alpha = 0.05 selected for this extractive batch recycle bioconversion process using ion-exchange resin.

Intracellular compounds quantification by means of flow cytometry in bacteria: application to xanthan production by Xanthomonas campestris

The use of flow cytometry (FCM) to quantitatively analyze intracellular compounds is studied. FCM is a very useful technique for individual cell studies in microbial systems, and gives access to information which cannot be obtained in any other way. Nevertheless, it provides data in arbitrary units, that is, relative data. This analytical technique could be employed for kinetic modeling of microbial systems and even for internal phenomena analysis, but for this purpose, absolute data-that is concentration of intracellular compounds-must be used. In this work, relative flow cytometry data are transformed into absolute data by means of calibrations employing the same fluorochromes with another technique: spectrofluorymetry. Calibrations of DNA, RNA, and protein intracellular concentrations are presented for the bacteria, Xanthomonas campestris. Other analytical methods, based on biochemical determinations, were also employed to quantify intracellular compounds, but the results obtained are very poor compared with those achieved by means of spectrofluorymetry (SFM). Calibration equations and data obtained by both techniques are given. Evolutions of protein and nucleic acids during Xanthomonas campestris growth and xanthan gum production are shown.

Lactic acid production in a cell retention continuous culture using lignocellulosic hydrolysate as a substrate

The effect of lignocellulosic hydrolysate of crushed corn cobs on the kinetics of growth and lactic acid production of Lactobacillus casei and L. lactis in the cell retention continuous culture was studied. The continuous cultivations were carried out in a continuous flow stirred bioreactor combined in a recycle loop with an ultrafiltration module retaining all biomass and allowing the continuous removal of metabolites, including lactic acid, as a cell-free permeate. Based on computer-aided analysis of experimental data, a simple physiological model of lactic acid cultivation was developed. The parameters of the model were estimated by non-linear regression.

Methods and strategies available for the process control and optimization of monoclonal antibody production

The objective of this paper is to explore the range of methods and strategies available for the process control and optimization of monoclonal antibody production by hybridoma cell culture. Emphasis will be placed on the choice of the level of complexity incorporated into the process control and optimisation procedure. It will be shown that the behaviour of hybridomas in culture is influenced by sophisticated cellular metabolic activities and various interactive environmental factors and that the understanding and modelling of the way hybridomas grow in the bioreactor should enable optimisation of bioreactor operating conditions to achieve maximum monoclonal antibody formation. However, due to the lack of on-line instrumentation of important biological variables and the incomplete knowledge of hybridoma cultivation process, there exist many limitations and challenges to the advent of applications of process control and optimisation in this field. To solve the problem, introduction of industrially practical biological measurements and development of new control concepts are inevitable. At the end of this paper, we shall discuss possible schemes for the control of the physiological state of cells in order that balanced cell growth and maximum monoclonal antibody synthesis may be achieved.