Dynamic metabolic flux analysis using a convex analysis approach: Application to hybridoma cell cultures in perfusion.

In recent years, dynamic metabolic flux analysis (DMFA) has been developed in order to evaluate the dynamic evolution of the metabolic fluxes. Most of the proposed approaches are dedicated to exactly determined or overdetermined systems. When an underdetermined system is considered, the literature suggests the use of dynamic flux balance analysis (DFBA). However the main challenge of this approach is to determine an appropriate objective function, which remains valid over the whole culture. In this work, we propose an alternative dynamic metabolic flux analysis based on convex analysis, DMFCA, which allows the determination of bounded intervals for the fluxes using the available knowledge of the metabolic network and information provided by the time evolution of extracellular component concentrations. Smoothing splines and mass balance differential equations are used to estimate the time evolution of the uptake and excretion rates from this experimental data. The main advantage of the proposed procedure is that it does not require additional constraints or objective functions, and provides relatively narrow intervals for the intracellular metabolic fluxes. DMFCA is applied to experimental data from hybridoma HB58 cell perfusion cultures, in order to investigate the influence of the operating mode (batch and perfusion) on the metabolic flux distribution.

Modeling of the ComRS Signaling Pathway Reveals the Limiting Factors Controlling Competence in Streptococcus thermophilus.

In streptococci, entry in competence is dictated by ComX abundance. In Streptococcus thermophilus, production of ComX is transient and tightly regulated during growth: it is positively regulated by the cell-cell communication system ComRS during the activation phase and negatively regulated during the shut-off phase by unidentified late competence gene(s). Interestingly, most S. thermophilus strains are not or weakly transformable in permissive growth conditions (i.e., chemically defined medium, CDM), suggesting that some players of the ComRS regulatory pathway are limiting. Here, we combined mathematical modeling and experimental approaches to identify the components of the ComRS system which are critical for both dynamics and amplitude of ComX production in S. thermophilus. We built a deterministic, population-scaled model of the time-course regulation of specific ComX production in CDM growth conditions. Strains LMD-9 and LMG18311 were respectively selected as representative of highly and weakly transformable strains. Results from in silico simulations and in vivo luciferase activities show that ComR concentration is the main limiting factor for the level of comX expression and controls the kinetics of spontaneous competence induction in strain LMD-9. In addition, the model predicts that the poor transformability of strain LMG18311 results from a 10-fold lower comR expression level compared to strain LMD-9. In agreement, comR overexpression in both strains was shown to induce higher competence levels with deregulated kinetics patterns during growth. In conclusion, we propose that the level of ComR production is one important factor that could explain competence heterogeneity among S. thermophilus strains.

On extremum seeking in bioprocesses with multivalued cost functions.

Finding optimal operating modes for bioprocesses has been, for a long time, a relevant issue in bioengineering. The problem is of special interest when it implies the simultaneous optimization of competing objectives. In this paper, we address the problem of finding optimal steady states that achieve the best tradeoff between yield and productivity by using nonmodel-based extremum-seeking control with semiglobal practical stability and convergence properties. A special attention is paid to processes with multiple steady states and multivalued cost functions.

Identification of reaction networks for bioprocesses: determination of a partially unknown pseudo-stoichiometric matrix.

In this paper we propose a methodology to determine the structure of the pseudo-stoichiometric coefficient matrix K in a mass balance-based model and to identify its coefficients from a set of available data. The first stage consists in estimating the number of reactions that must be taken into account to represent the main mass transfer within the bioreactor. This provides the dimension of K. Then we propose a method to directly determine the structure of the matrix (i.e. mainly its zeros and the signs of its coefficients). These methods are illustrated with simulations of a process of lipase production from olive oil by Candida rugosa.

On the estimation of the pseudo-stoichiometric matrix for macroscopic mass balance modelling of biotechnological processes.

In this paper we propose a methodology to determine the structure of the pseudo-stoichiometric coefficient matrix kappa in a macroscopic mass balance based model. The first step consists in estimating the minimal number of reactions that must be taken into account to represent the main mass transfer within the bioreactor. This provides the dimension of kappa. Then we discuss the identifiability of the components of kappa and we propose a method to estimate their values. Finally we present a method to select among a set of possible macroscopic reaction networks those which are in agreement with the available measurements. These methods are illustrated with three examples: real data of the growth and biotransformation of the filamentous fungi Pycnoporus cinnabarinus, real data of an anaerobic digester involving a bacterial consortium degrading a mixture of organic substrates and a process of lipase production from olive oil by Candida rugosa.

Effect of Partial Medium Replacement on Cell Growth and Protein Production for the High-Fivetrade mark insect cell line.

The potential of spent medium to support the growth and recombinant protein production of High-Fivetrade mark cells was investigated. Growth in medium consisting of three parts fresh and one part spent medium was comparable to that in fresh medium (maximal specific growth rates of 0.028 and 0.029 h(-1), and maximal cell densities of 4 and 4.5 x 10(6) cells ml(-1), respectively). Glucose exhaustion coincided with an abrupt decrease of viability. Of 15 amino acids analyzed, not a single one was completely exhausted at the end of the growth phase. Growth in medium consisting of equal parts spent and fresh medium led to lower maximal cell concentration (2.9 x 10(6) cells ml(-1)) with a smoother death phase. Glucose supplementation at the beginning of the culture or at the end of the growth phase did not lead to an increase of either the maximal cell density or the specific growth rate. Infection of High-Fivetrade mark cells at three different densities (1.4, 2.5 and 4.2 x 10(6) cells ml(-1)) without medium change led to monotonically decreased specific productions for beta-galactosidase. Partial (75%) or total medium replacement at the higher infection density restored the specific production at the levels of the intermediate density infection (321, 292 and 389 U.(10(6) cells)(-1), respectively).

Microcarrier culture of lepidopteran cell lines: implications for growth and recombinant protein production.

Several microcarrier systems were screened with Sf-9 and High-Five cell lines as to their ability to support cell growth and recombinant (beta-galactosidase) protein production. Growth of both cell lines on compact microcarriers, such as Cytodex-1 and glass beads, was minimal, as cells detached easily from the microcarrier surface and grew as single cells in the medium. Cell growth was also problematic on Cytopore-1 and -2 porous microcarriers. Cells remained attached for several days inside the microcarrier pores, but no cell division and proliferation were observed. On the contrary, insect cells grew well in the interior of Fibra-Cel disks mainly as aggregates at points of fiber intersection, reaching final (plateau) densities of about 4 x 10(6) (Sf-9) and 2.7 x 10(6) (High-Five) cells mL(-1) (8 x 10(6) and 5.5 x 10(6) cells per cm(2) of projected disk area, respectively). Their growth was described well by the logistic equation, which takes into account possible inhibition effects. Beta-Galactosidase (beta-gal) production of Sf-9 cells on Fibra-Cel disks (infected at 3.3 x 10(6) cells mL(-1)) was prolonged (192 h), and specific protein production was similar to that of high-density free cell infection. Cultispher-S microcarriers were found to be a very efficient system for the growth of High-Five cells, whereas no growth of Sf-9 cells took place for the same system. Concentrations of about 9 x 10(6) cells mL(-1) were reached within 120 h, with cell growth in both microcarriers and aggregates, appearance of cellular bridges between microcarriers and aggregates, and eventual formation of macroaggregates incorporating several microcarriers. Specific protein productions after beta-gal baculovirus infection at increasing cell concentrations were almost constant, thus leading to elevated volumetric protein production: final beta-gal titers of 946, 1728, and 1484 U mL(-1) were obtained for infection densities of 3.4, 7.2, and 8.9 x 10(6) cells mL(-1), respectively.

Feedback stabilization of fed-batch bioreactors: non-monotonic growth kinetics.

This paper deals with the design of a feedback controller for fed-batch microbial conversion processes that forces the substrate concentration C(S) to a desired setpoint, starting from an arbitrary (initial) substrate concentration when non-monotonic growth kinetics apply. This problem is representative for a lot of industrial fermentation processes, with the baker's yeast fermentation as a well-known example. It is assumed that the specific growth rate mu is function of the substrate concentration only. A first approach exploits the availability of on-line measurements of both the substrate and biomass concentration. A second approach is merely based on on-line measurements of the biomass concentration, which provide an estimate for the specific growth rate. After a reformulation of the substrate concentration setpoint into a specific growth rate setpoint, it is demonstrated that the fed-batch process can still be stabilized around any desired operating point along the non-monotonic kinetics.