On-line monitoring of recombinant bacterial cultures using multi-wavelength fluorescence spectroscopy

Multi-wavelength fluorescence spectroscopy was evaluated as a tool for on-line monitoring of recombinant Escherichia coli cultivations expressing human basic fibroblast growth factor (hFGF-2). The data sets for the various combinations of the excitation and emission spectra from batch cultivations were analyzed using principal component analysis. Chemometric models (the partial least squares method) were developed for correlating the fluorescence data and the experimentally measured variables such as the biomass and glucose concentrations as well as the carbon dioxide production rate. Excellent correlations were obtained for these variables for the calibration cultivations. The predictability of these models was further tested in batch and fed-batch cultivations. The batch cultivations were well predicted by the PLS models for biomass, glucose concentrations and carbon dioxide production rate (RMSEPs were respectively 5%, 7%, 9%). However, when tested for biomass concentrations in fed-batch cultivations (with final biomass three times higher than the highest calibration data) the models had good predictability at high growth rates (RMSEPs were 3% and 4%, respectively for uninduced and induced fed-batch cultivations), which was as good as for the batch cultivations used for developing the models (RMSEPs were 3% and 5%, respectively for uninduced and induced batch cultivations). The fed-batch cultivations performed at low growth rates exhibited much higher fluorescence for fluorophores such as flavin and NAD(P)H as compared to fed-batch cultivations at high growth rate. Therefore, the PLS models tended to over-predict the biomass concentrations at low growth rates. Obviously the cells changed their concentration of biogenic fluorophores depending on the growth rate. Although multi-wavelength fluorescence spectroscopy is a valuable tool for on-line monitoring of bioprocess, care must be taken to re-calibrate the PLS models at different growth rates to improve the accuracy of predictions.     

Structured modeling and state estimation in a fermentation process: Lipase production by Candida rugosa

A simple structured mathematical model coupled with a methodology of state and parameter estimation is developed for lipase production by Candida rugosa in batch fermentation. The model describes the system according to the following qualitative observations and hypothesis: Lipase production is induced by extracellular oleic acid present in the medium. The acid is transported into the cell where it is consumed, transformed, and stored. Lipase is excreted to the medium where it is distributed between the available oil-water interphase and aqueous phase. Cell growth is modulated by the intracellular substrate concentration. Model parameters have been determined and the whole model validated against experiments not used in their determination. The estimation problem consists in the estimation of three state variables (biomass, intra- and extracellular substrate) and two kinetic parameters by using only the on-line measurement provided by exhaust gas analysis. The presented estimation strategy divides the complex problem into three subproblems that can be solved by stable algorithms. The estimation of biomass (X) and the specific growth rate (mu), is achieved by a recursive prediction error algorithm using the on-line measurement of the carbon dioxide evolution rate. mu is then used to perform an estimation of intracellular substrate and the other kinetic parameter related to substrate transport (A) by an adaptive observer. Extracellular substrate is then evaluated by means of the estimated values of intracellular substrate and biomass through the material balance of the reactor. Simulation and experimental tests showed good performance of the developed estimator, which appears suitable to be used for process control and monitoring. (c) 1995 John Wiley & Sons, Inc.     

基于动力学模型的法夫酵母发酵生产虾青素的补料策略优化

对法夫酵母的不同补料发酵方式进行了研究.基于底物抑制模型,提出了一种优化的两阶段补料策略,用于法夫酵母产虾青素的高密度发酵.在发酵的延迟期和对数生长期早期,糖浓度控制在25 g/L左右,在此条件下,生物量可以达到最大,且时间缩短.在对数生长期后期及稳定期,糖浓度控制在5 g/L,虾青素的合成时间可以有效延长.与传统的补料方式相比,采用此补料策略取得了较好的发酵效果.发酵终点细胞干重达到23.8g/L,虾青素产量达到29.05 mg/L,分别比分批发酵提高了52.8%和109%.     

Optimization of process parameters by response surface methodology and kinetic modeling for batch production of canthaxanthin by Dietzia maris NIT-D (accession number: HM151403)

Dietzia maris NIT-D, a canthaxanthin producer, was isolated during routine screening of pigment-producing bacteria. Response surface methodology was applied for statistical designing of process parameters for biomass and canthaxanthin production. The effects of four process parameters (considered as independent variables), namely temperature (10-30?°C), pH (4.75-5.75), shaker speed (75-135?rpm) and percentage inoculum (0.5-2.5?%) on the biomass and canthaxanthin yield (considered as dependent variables) were studied. As much as 122?mg?L(-1) of canthaxanthin was obtained when Dietzia maris NIT-D was incubated for 120?h at 25?°C and 120?rpm, initial pH and percentage inoculum being 5.5 and 2?% respectively. The pigment yield is the highest reported till date, with Dietzia maris as the test organism. The maximum biomass yield was 7.39?g?L(-1) under optimized process parameters. The predicted values were also verified by validation experiments in 5-day fermentation. Different mathematical models were used to describe growth and production, considering the effect of glucose in batch mode. The kinetic constants were calculated by fitting the experimental data to the models. Cell growth was inhibited beyond a glucose concentration of 15?g?L(-1). Andrews' model gave the best fit with a R (2) value of 0.9993. During the exponential growth phase, the specific growth rate was found to remain fairly constant with respect to time. There was no inhibitory effect due to intracellular product accumulation for all concentrations of glucose. This observation is the first of its kind, as previous studies have reported that increasing accumulation of intracellular carotenoid exerts greater degree of inhibition on growth.     

High-throughput analysis of the plasmid bioproduction process in Escherichia coli by FTIR spectroscopy

Monitoring plasmid production systems is a lab intensive task. This article proposes a methodology based on FTIR spectroscopy and the use of chemometrics for the high-throughput analysis of the plasmid bioproduction process in E. coli. For this study, five batch cultures with different initial medium compositions are designed to represent different biomass and plasmid production behavior, with the maximum plasmid and biomass concentrations varying from 11 to 95 mg L(-1) and 6.8 to 12.8 g L(-1), respectively, and the plasmid production per biomass varying from 0.4 to 5.1 mg g(-1). After a short sample processing consisting of centrifugation and dehydration, the FTIR spectra are recorded from the collected cellular biomass using microtiter plates with 96 wells. After spectral pre-processing, the predictive FTIR spectra models are derived by using partial least squares (PLS) regression with the wavenumber selection performed by a Monte-Carlo strategy. Results show that it is possible to improve the PLS models by selecting specific spectral ranges. For the plasmid model, the spectral regions between 590-1,130, 1,670-2,025, and 2,565-3,280 cm(-1) are found to be highly relevant. Whereas for the biomass, the best wavenumber selections are between 900-1,200, 1,500-1,800, and 2,850-3,200 cm(-1). The optimized PLS models show a high coefficient of determination of 0.91 and 0.89 for the plasmid and biomass concentration, respectively. Additional PLS models for the prediction of the carbon sources glucose and glycerol and the by-product acetic acid, based on metabolism-induced correlations between the nutrients and the cellular biomass are also established.     

Influence of nutrients on enhancing laccase production by Botryosphaeria rhodina MAMB-05.

The physiological requirements needed to enhance the production of laccases by the ascomycete Botryosphaeria rhodina MAMB-05 in submerged cultivation were examined under non-induced and induced (veratryl alcohol, VA) conditions. Under non-induced conditions (-VA), the initial pH, C:N ratio, and inorganic N source did not influence laccase production, in contrast to Tween 80, soybean oil, and copper, which significantly increased laccase production, and proline and urea, which suppressed laccase formation. In addition, Tween 60 could serve as the sole carbon source for the production of these enzymes. Under VA-induced conditions of fungal growth, factors such as inoculum type, time-point of addition of inducer, initial pH, C:N ratio, and type of N source, influenced the production of laccases; however, unlike the non-induced conditions, proline and urea did not act as suppressors. Each of these physiological conditions exerted different effects on biomass production. The nutritional conditions examined for B. rhodina MAMB-05 are discussed in relation to their influence on fungal growth and laccase production.     

Vitreoscilla hemoglobin expression in engineered Escherichia coli: improved performance in high cell-density batch cultivations

High cell-density cultivations are the preferred system for biomolecules production by Escherichia coli. It has been previously demonstrated that a strain of E. coli with a modified substrate transport system is able to attain high cell densities in batch mode, due to the very low overflow metabolism displayed. The use of elevated amounts of glucose from the beginning of the cultivation, eliminates the existence of substrate gradients due to deficient mixing at large-scale. However, the large amounts of oxygen demanded resulted in microaerobic conditions after some hours of cultivation, even at small-scale. In this work, the effect of expressing the Vitreoscilla hemoglobin (VHb) in the engineered strain during batch cultures using high-glucose concentrations was tested. Together, the expression of VHb and the modified substrate transport system resulted in a 33% increase of biomass production compared to the parental strain (W3110) lacking the VHb in batch cultivations using 25 g/L of glucose. When 50 g/L of glucose were used, expression of VHb in the modified strain led to 11% higher biomass production compared to W3110. The VHb also increased the growth rates of the strains by about 30% in the aerobic phase and more than 200% in the microaerobic phase of batch cultivation.     

Sorbitol co-feeding reduces metabolic burden caused by the overexpression of a Rhizopus oryzae lipase in Pichia pastoris

To improve the specific production rate of Rhizopus oryzae lipase (ROL) in Pichia pastoris, a protein that triggers the unfolded protein response in P. pastoris, the effect of sorbitol/methanol mixed substrates was tested in batch and fed-batch cultures. Remarkably, a different substrate consumption behaviour was observed depending on the host's phenotype (Mut(+) or Mut(s)) in batch cultures: when the methanol assimilation capacity is genetically reduced (Mut(s) phenotype), both substrates were consumed simultaneously, allowing not only a higher specific growth rate but also higher lipase levels (8.7-fold) compared to those obtained by cells growing on methanol as a sole carbon source in batch culture. This effect was not observed in Mut(+) phenotype, where the two substrates were consumed sequentially and the levels of heterologous product were only slightly higher (1.7-fold). A mixed substrate strategy was also applied to a Mut(s) fed-batch culture at a low methanol concentration set-point (0.5 gl(-1)). This resulted in a 2.2-fold increase in the heterologous protein level achieved, compared with the methanol-only feeding strategy. In addition, sorbitol co-feeding permitted the achievement of higher specific growth rates, and avoided the drastic decrease of the specific production rate observed after the start of the induction phase when methanol was used as sole carbon source This resulted in a significant increase in the overall bioprocess volumetric productivity (2.2-fold) and specific productivity (1.7-fold). Moreover, whereas increased ROL gene dosage in Mut(s) strains have been previously reported to be deleterious for P. pastoris cells growing on methanol, sorbitol co-feeding allowed for sustained cell growth and lipase production.     

Modeling and simulation of Streptomyces peucetius var. caesius N47 cultivation and epsilon-rhodomycinone production with kinetic equations and neural networks

This study focuses on comparing different kinetic growth models and the use of neural networks in the batch cultivation of Streptomyces peucetius var. caesius producing epsilon-rhodomycinone. Contois, Monod and Teissier microbial growth models were used as well as the logistic growth modeling approach, which was found best in the simulations of growth and glucose consumption in the batch growth phase. The lag phase was included in the kinetic model with a CO2 trigger and a delay factor. Substrate consumption and product formation were included as Luedeking-Piret and logistic type equations, respectively. Biomass formation was modeled successfully with a 6-8-2 network, and the network was capable of biomass prediction with an R2-value of 0.983. Epsilon-rhodomycinone production was successfully modeled with a recursive 8-3-1 network capable of epsilon-rhodomycinone prediction with an R2-value of 0.903. The predictive power of the neural networks was superior to the kinetic models, which could not be used in predictive modeling of arbitrary batch cultivations.     

Software sensors for biomass concentration in a SSC process using artificial neural networks and support vector machine

The lack of sensors for some relevant state variables in fermentation processes can be coped by developing appropriate software sensors. In this work, NARX-ANN, NARMAX-ANN, NARX-SVM and NARMAX-SVM models are compared when acting as software sensors of biomass concentration for a solid substrate cultivation (SSC) process. Results show that NARMAX-SVM outperforms the other models with an SMAPE index under 9 for a 20 % amplitude noise. In addition, NARMAX models perform better than NARX models under the same noise conditions because of their better predictive capabilities as they include prediction errors as inputs. In the case of perturbation of initial conditions of the autoregressive variable, NARX models exhibited better convergence capabilities. This work also confirms that a difficult to measure variable, like biomass concentration, can be estimated on-line from easy to measure variables like CO₂ and O₂ using an adequate software sensor based on computational intelligence techniques.     

氨水流加用于粪产碱杆菌热凝胶发酵

热凝胶是粪产碱杆菌(Alcaligenes faecalis)在氮源限制条件下生成的水不溶性胞外多糖,分泌到胞外后就附着在菌体外壁,因此在细胞生长期提高生物量对促进热凝胶合成有重要意义。热凝胶分批发酵时, 起始NH4Cl浓度提高到3.6 g/L时能促进菌体生长和热凝胶合成,但是过量NH4Cl会抑制热凝胶合成,且生物量提高不是很明显。为了进一步提高菌体浓度, 在菌体生长期, 氨水取代NaOH溶液进行流加控制pH为7.0, 随后又用2 mol/L NaOH控制pH 5.6。实验表明, 氨水流加使菌体浓度大大提高,流加24 h使菌体浓度达到18.8 g/L。但是菌体浓度过高也会抑制热凝胶的合成,在氨水流加14 h时,菌体浓度在11.9 g/L左右, 热凝胶产量最高（72 g/L）。     

氨水流加用于粪产碱杆菌热凝胶发酵

热凝胶是粪产碱杆菌(Alcaligenes faecalis)在氮源限制条件下生成的水不溶性胞外多糖,分泌到胞外后就附着在菌体外壁,因此在细胞生长期提高生物量对促进热凝胶合成有重要意义。热凝胶分批发酵时, 起始NH4Cl浓度提高到3.6 g/L时能促进菌体生长和热凝胶合成,但是过量NH4Cl会抑制热凝胶合成,且生物量提高不是很明显。为了进一步提高菌体浓度, 在菌体生长期, 氨水取代NaOH溶液进行流加控制pH为7.0, 随后又用2 mol/L NaOH控制pH 5.6。实验表明, 氨水流加使菌体浓度大大提高,流加24 h使菌体浓度达到18.8 g/L。但是菌体浓度过高也会抑制热凝胶的合成,在氨水流加14 h时,菌体浓度在11.9 g/L左右, 热凝胶产量最高（72 g/L）。     

Making sense of parameter estimation and model simulation in bioprocesses

Most articles that report fitted parameters for kinetic models do not include meaningful statistical information. This study demonstrates the importance of reporting a complete statistical analysis and shows a methodology to perform it, using functionalities implemented in computational tools. As an example, alginate production is studied in a batch stirred-tank fermenter and modeled using the kinetic model proposed by Klimek and Ollis (1980). The model parameters and their 95% confidence intervals are estimated by nonlinear regression. The significance of the parameters value is checked using a hypothesis test. The uncertainty of the parameters is propagated to the output model variables through prediction intervals, showing that the kinetic model of Klimek and Ollis (1980) can simulate with high certainty the dynamic of the alginate production process. Finally, the results obtained in other studies are compared to show how the lack of statistical analysis can hold back a deeper understanding about bioprocesses.     

Glycine significantly enhances bacterial membrane vesicle production: a powerful approach for isolation of LPS-reduced membrane vesicles of probiotic Escherichia coli

Bacterial membrane vesicles (MVs) have attracted strong interest in recent years as novel nanoparticle delivery platforms. Glycine is known to induce morphological changes in the outer layer of bacteria. We report here that glycine dramatically facilitates MV production in a flagella-deficient mutant of the non-pathogenic probiotic Escherichia coli strain Nissle 1917. Supplementation of culture medium with 1.0% glycine induced cell deformation at the early exponential phase, eventually followed by quasi-lysis during the late exponential to stationary phase. Glycine supplementation also significantly increased the number of MVs with enlarged particle size and altered the protein profile with an increase in the inner membrane and cytoplasmic protein contents as compared to non-induced MVs. Of note, the endotoxin activity of glycine-induced MVs was approximately eightfold or sixfold lower than that of non-induced MVs when compared at equal protein or lipid concentrations respectively. Nevertheless, glycine-induced MVs efficiently induced both immune responses in a mouse macrophage-like cell line and adjuvanticity in an intranasal vaccine mouse model, comparable to those of non-induced MVs. We propose that the present method of inducing MV production with glycine can be used for emerging biotechnological applications of MVs that have immunomodulatory activities, while dramatically reducing the presence of endotoxins.     

Chemometric analysis of in-line multi-wavelength fluorescence measurements obtained during cultivations with a lipase producing Aspergillus oryzae strain

The filamentous fungus, Aspergillus oryzae, was cultivated in batch and fed-batch cultivations in order to investigate the use of multi-wavelength fluorescence for monitoring course of events during filamentous fungi cultivations. The A. oryzae strain applied expressed a fungal lipase from Thermomyces lanuginosus. Spectra of multi-wavelength fluorescence were collected every 5 min with the BioView system (DELTA, Denmark) and both explorative and predictive models, correlating the fluorescence data with cell mass and lipase activity, were built. During the cultivations, A. oryzae displayed dispersed hyphal growth and under these conditions no fouling of the multi-wavelength fluorescence sensor was observed. The scores of a parallel factor analysis (PARAFAC) model, based on the fluorescence spectra, gave clear evidence of, for example, the on-set of the feeding phase. The predictive models, estimating the cell mass, showed correlations between 0.73 and 0.97 with root mean square error of cross validation (RMSECV) values between 1.48 and 0.77 g . kg(-1). A model estimating the lipase activity was also constructed for the fed-batch cultivations with a correlation of 0.93. The results presented here clearly show that multi-wavelength fluorescence is a useful tool for monitoring fed-batch cultivations of filamentous fungi.     

Enabling PAT in insect cell bioprocesses: In situ monitoring of recombinant adeno-associated virus production by fluorescence spectroscopy

The process analytical technology (PAT) initiative shifted the bioprocess development mindset towards real-time monitoring and control tools to measure relevant process variables online, and acting accordingly when undesirable deviations occur. Online monitoring is especially important in lytic production systems in which released proteases and changes in cell physiology are likely to affect product quality attributes, as is the case of the insect cell-baculovirus expression vector system (IC-BEVS), a well-established system for production of viral vectors and vaccines. Here, we applied fluorescence spectroscopy as a real-time monitoring tool for recombinant adeno-associated virus (rAAV) production in the IC-BEVS. Fluorescence spectroscopy is simple, yet sensitive and informative. To overcome the strong fluorescence background of the culture medium and improve predictive ability, we combined artificial neural network models with a genetic algorithm-based approach to optimize spectra preprocessing. We obtained predictive models for rAAV titer, cell viability and cell concentration with normalized root mean squared errors of 7%, 4%, and 7%, respectively, for leave-one-batch-out cross-validation. Our approach shows fluorescence spectroscopy allows real-time determination of the best time of harvest to maintain rAAV infectivity, an important quality attribute, and detection of deviations from the golden batch profile. This methodology can be applied to other biopharmaceuticals produced in the IC-BEVS, supporting the use of fluorescence spectroscopy as a versatile PAT tool.     

Testing the generality of above‐ground biomass allometry across plant functional types at the continent scale

Accurate ground‐based estimation of the carbon stored in terrestrial ecosystems is critical to quantifying the global carbon budget. Allometric models provide cost‐effective methods for biomass prediction. But do such models vary with ecoregion or plant functional type? We compiled 15 054 measurements of individual tree or shrub biomass from across Australia to examine the generality of allometric models for above‐ground biomass prediction. This provided a robust case study because Australia includes ecoregions ranging from arid shrublands to tropical rainforests, and has a rich history of biomass research, particularly in planted forests. Regardless of ecoregion, for five broad categories of plant functional type (shrubs; multistemmed trees; trees of the genus Eucalyptus and closely related genera; other trees of high wood density; and other trees of low wood density), relationships between biomass and stem diameter were generic. Simple power‐law models explained 84–95% of the variation in biomass, with little improvement in model performance when other plant variables (height, bole wood density), or site characteristics (climate, age, management) were included. Predictions of stand‐based biomass from allometric models of varying levels of generalization (species‐specific, plant functional type) were validated using whole‐plot harvest data from 17 contrasting stands (range: 9–356 Mg ha^?1). Losses in efficiency of prediction were <1% if generalized models were used in place of species‐specific models. Furthermore, application of generalized multispecies models did not introduce significant bias in biomass prediction in 92% of the 53 species tested. Further, overall efficiency of stand‐level biomass prediction was 99%, with a mean absolute prediction error of only 13%. Hence, for cost‐effective prediction of biomass across a wide range of stands, we recommend use of generic allometric models based on plant functional types. Development of new species‐specific models is only warranted when gains in accuracy of stand‐based predictions are relatively high (e.g. high‐value monocultures).     

Production of poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) by Ralstonia eutropha in high cell density palm oil fermentations

Improved production costs will accelerate commercialization of polyhydroxyalkanoate (PHA) polymer and PHA-based products. Plant oils are considered favorable feedstocks, due to their high carbon content and relatively low price compared to sugars and other refined carbon feedstocks. Different PHA production strategies were compared using a recombinant strain of Ralstonia eutropha that produces high amounts of P(HB-co-HHx) when grown on plant oils. This R. eutropha strain was grown to high cell densities using batch, extended batch, and fed batch fermentation strategies, in which PHA accumulation was triggered by nitrogen limitation. While extended batch culture produced more biomass and PHA than batch culture, fed batch cultivation was shown to produce the highest levels of biomass and PHA. The highest titer achieved was over 139 g/L cell dry weight (CDW) of biomass with 74% of CDW as PHA containing 19 mol% HHx. Our data suggest that the fermentation process is scalable with a space time yield (STY) better than 1 g PHA/L/h. The achieved biomass concentration and PHA yield are among the highest reported for the fermentation of recombinant R. eutropha strains producing P(HB-co-HHx).     

Biodegradable polymers from organic acids by using activated sludge enriched by aerobic periodic feeding

This article describes a new process for the production of biopolymers (polyhydroxyalkanoates, PHAs) based on the aerobic enrichment of activated sludge to obtain mixed cultures able to store PHAs at high rates and yields. Enrichment was obtained through the selective pressure established by feeding the carbon source in a periodic mode (feast and famine regime) in a sequencing batch reactor. A concentrated mixture of acetic, lactic, and propionic acids (overall concentration of 8.5 gCOD L(-1)) was fed every 2 h at 1 day(-1) overall dilution rate. Even at such high organic load (8.5 gCOD L(-1) day(-1)), the selective pressure due to periodic feeding was effective in obtaining a biomass with a storage ability much higher than activated sludges. The immediate biomass response to substrate excess (as determined thorough short-term batch tests) was characterized by a storage rate and yield of 649 mgPHA (as COD) g biomass (as COD)(-1) h(-1) and 0.45 mgPHA (as COD) mg removed substrates (as COD(-1)), respectively. When the substrate excess was present for more than 2 h (long-term batch tests), the storage rate and yield decreased, whereas growth rate and yield significantly increased due to biomass adaptation. A maximum polymer fraction in the biomass was therefore obtained at about 50% (on COD basis). As for the PHA composition, the copolymer poly(beta-hydroxybutyrate/beta-hydroxyvalerate) with 31% of hydroxyvalerate monomer was produced from the substrate mixture. Comparison of the tests with individual and mixed substrates seemed to indicate that, on removing the substrate mixture for copolymer production, propionic acid was fully utilized to produce propionylCoA, whereas the acetylCoA was fully provided by acetic and lactic acid.