Perspectives in the Modeling and Optimization of PHB Production by Pure and Mixed Cultures

ABSTRACT

Poly(β-hydroxybutyrate) or PHB is an important member of the family of polyhydroxyalkanoates with properties that make it potentially competitive with synthetic polymers. In addition, PHB is biodegradable. While the biochemistry of PHB synthesis by microorganisms is well known, improvement of large-scale productivity requires good fermentation modeling and optimization. The latter aspect is reviewed here.

Current models are of two types: (i) mechanistic and (ii) cybernetic. The models may be unstructured or structured, and they have been applied to single cultures and co-cultures. However, neither class of models expresses adequately all the important features of large-scale non-ideal fermentations. Model-independent neural networks provide faithful representations of observations, but they can be difficult to design. So hybrid models, combining mechanistic, cybernetic and neural models, offer a useful compromise. All three kinds of basic models are discussed with applications and directions toward hybrid model development.     

Cognitive optimization of microbial PHB production in an optimally dispersed bioreactor by single and mixed cultures

Cognitive (or intelligent) models are often superior to mechanistic models for nonideal bioreactors. Two kinds of cognitive models—cybernetic and neural—were applied recently to fed-batch fermentation by Ralstonia eutropha in a bioreactor with optimum finite dispersion. In the present work, these models have been applied in simulation studies of co-cultures of R. eutropha and Lactobacillus delbrueckii. The results for both cognitive and mechanistic models have been compared with single cultures. Neural models were the most effective for both types of cultures and mechanistic models the least effective. Simulations with co-culture fermentations predicted more PHB than single cultures with all three types of models. Significantly, the predicted enhancements in PHB concentration by cognitive methods for mixed cultures were four to five times larger than the corresponding increases in biomass concentration. Further improvements are possible through a hybrid combination of all three types of models.     

Synthesizing cellular intelligence and artificial intelligence for bioprocesses

Microbial processes operated under realistic conditions are difficult to describe by mechanistic models, thereby limiting their optimization and control. Responses of living cells to their environment suggest that they possess some "innate intelligence". Such responses have been modeled by a cybernetic approach. Furthermore, the overall behavior of a bioreactor containing a population of cells may be described and controlled through artificial intelligence methods. Therefore, it seems logical to combine cybernetic models with artificial intelligence to evolve an integrated intelligence-based strategy that is physiologically more faithful than the current approaches. This possibility is discussed, together with practical considerations favoring a hybrid approach that includes some mathematical modeling.     

Neural and Hybrid Optimizations of the Fed-Batch Synthesis of Poly-β-Hydroxybutyrate by Ralstonia eutropha in a Nonideal Bioreactor

Poly-β-hydroxybutyrate (PHB) is synthesized by some microorganisms under stressful conditions. Despite its properties being comparable to those of synthetic polymers, and its biocompatibility and biodegradability, low productivities have dampened commercial interest in microbial PHB production. To increase production efficiency, a fed-batch fermentation with Ralstonia eutropha was optimized recently through a neural-cum-dispersion model (D-model) incorporating incomplete dispersion and noise in the feed streams. The approach described in the work has been improved in two ways: first by a model comprising neural networks only (N-model) and then by a hybrid neural model (H-model) with a mathematical component. At optimum dispersion, PHB production through the N-model optimization was 35% more than by the D-model, and this was enhanced by a further 58% using hybrid optimization. Recognizing that the D-model itself more than doubled the PHB production compared to a noise-free fully dispersed bioreactor, the present results establish hybrid neural optimization as a viable method for PHB production improvement under realistic conditions.     

Cybernetic model predictive control of a continuous bioreactor with cell recycle

The control of poly-beta-hydroxybutyrate (PHB) productivity in a continuous bioreactor with cell recycle is studied by simulation. A cybernetic model of PHB synthesis in Alcaligenes eutrophus is developed. Model parameters are identified using experimental data, and simulation results are presented. The model is interfaced to a multirate model predictive control (MPC) algorithm. PHB productivity and concentration are controlled by manipulating dilution rate and recycle ratio. Unmeasured time varying disturbances are imposed to study regulatory control performance, including unreachable setpoints. With proper controller tuning, the nonlinear MPC algorithm can track productivity and concentration setpoints despite a change in the sign of PHB productivity gain with respect to dilution rate. It is shown that the nonlinear MPC algorithm is able to track the maximum achievable productivity for unreachable setpoints under significant process/model mismatch. The impact of model uncertainty upon controller performance is explored. The multirate MPC algorithm is tested using three controllers employing models that vary in complexity of regulation. It is shown that controller performance deteriorates as a function of decreasing biological complexity.     

Parallel hybrid modeling methods for a full-scale cokes wastewater treatment plant

Parallel hybrid modeling methods are applied to a full-scale cokes wastewater treatment plant. Within the hybrid model structure, a mechanistic model specifies the basic dynamics of the relevant process and a non-parametric model compensates for the inaccuracy of the mechanistic model. First, a simplified mechanistic model is developed based on Activated Sludge Model No. 1 and the specific process knowledge of the cokes wastewater treatment process. Then, the mechanistic model is combined with five different non-parametric models--feedforward back-propagation neural network, radial basis function network, linear partial least squares (PLS), quadratic PLS and neural network PLS (NNPLS)--in parallel configuration. These models are identified with the same data obtained from the plant operation to predict dynamic behavior of the process. The performance of each parallel hybrid model is compared based on their ease of model building, prediction accuracy and interpretability. For this application, the parallel hybrid model with NNPLS as non-parametric model gives better performance than other parallel hybrid models. In addition, the NNPLS model is used to analyze the behavior of the operation data in the reduced space and allows for fault detection and isolation.     

Hybrid neural network modeling of a full-scale industrial wastewater treatment process

In recent years, hybrid neural network approaches, which combine mechanistic and neural network models, have received considerable attention. These approaches are potentially very efficient for obtaining more accurate predictions of process dynamics by combining mechanistic and neural network models in such a way that the neural network model properly accounts for unknown and nonlinear parts of the mechanistic model. In this work, a full-scale coke-plant wastewater treatment process was chosen as a model system. Initially, a process data analysis was performed on the actual operational data by using principal component analysis. Next, a simplified mechanistic model and a neural network model were developed based on the specific process knowledge and the operational data of the coke-plant wastewater treatment process, respectively. Finally, the neural network was incorporated into the mechanistic model in both parallel and serial configurations. Simulation results showed that the parallel hybrid modeling approach achieved much more accurate predictions with good extrapolation properties as compared with the other modeling approaches even in the case of process upset caused by, for example, shock loading of toxic compounds. These results indicate that the parallel hybrid neural modeling approach is a useful tool for accurate and cost-effective modeling of biochemical processes, in the absence of other reasonably accurate process models.     

Insights into large-scale cell-culture reactors: I. Liquid mixing and oxygen supply

In the pharmaceutical industry, it is state of the art to produce recombinant proteins and antibodies with animal-cell cultures using bioreactors with volumes of up to 20 m(3) . Recent guidelines and position papers for the industry by the US FDA and the European Medicines Agency stress the necessity of mechanistic insights into large-scale bioreactors. A detailed mechanistic view of their practically relevant subsystems is required as well as their mutual interactions, i.e., mixing or homogenization of the culture broth and sufficient mass and heat transfer. In large-scale bioreactors for animal-cell cultures, different agitation systems are employed. Here, we discuss details of the flows induced in stirred tank reactors relevant for animal-cell cultures. In addition, solutions of the governing fluid dynamic equations obtained with the so-called computational fluid dynamics are presented. Experimental data obtained with improved measurement techniques are shown. The results are compared to previous studies and it is found that they support current hypotheses or models. Progress in improving insights requires continuous interactions between more accurate measurements and physical models. The paper aims at promoting the basic mechanistic understanding of transport phenomena that are crucial for large-scale animal-cell culture reactors.     

Application of spectrofluorometry to the prediction of PHB concentrations in a fed-batch process

On-line estimation of biopolymer production during fermentation would be a useful adjunct to the development of strategies for process control and optimization. This study examined the applicability of spectrofluorometry, along with other on-line measurements, for the prediction of poly-ß-hydroxybutyric acid (PHB) concentrations in a high-cell density fed-batch fermentation of Ralstonia eutropha. Models previously used for modelling PHB evolution with time are not sufficiently accurate for situations where transient intermediate accumulations or PHB degradation occur. Thus, the mass balance in the model was modified to account for these situations. An estimation algorithm was developed that is based on a hybrid model consisting of a dynamic mass balance of PHB where the main reaction coefficient was regressed with respect to spectrofluorometric data. The regression between the kinetic parameter and the spectrofluorometric data was accomplished using partial least squares (PLS) regression to avoid high sensitivity to noise expected from highly correlated data, such as the spectrofluorometric readings. The model accounts for dynamics of intermediates and in this way allows the prediction of dynamic behaviour in PHB concentrations that cannot be predicted with other reported mathematical models.     

Neural networks for perceptual processing: from simulation tools to theories

Neural networks are modelling tools that are, in principle, able to capture the input-output behaviour of arbitrary systems that may include the dynamics of animal populations or brain circuits. While a neural network model is useful if it captures phenomenologically the behaviour of the target system in this way, its utility is amplified if key mechanisms of the model can be discovered, and identified with those of the underlying system. In this review, we first describe, at a fairly high level with minimal mathematics, some of the tools used in constructing neural network models. We then go on to discuss the implications of network models for our understanding of the system they are supposed to describe, paying special attention to those models that deal with neural circuits and brain systems. We propose that neural nets are useful for brain modelling if they are viewed in a wider computational framework originally devised by Marr. Here, neural networks are viewed as an intermediate mechanistic abstraction between 'algorithm' and 'implementation', which can provide insights into biological neural representations and their putative supporting architectures.     

Application of hybrid cybernetic model in simulating myeloma cell culture co-consuming glucose and glutamine with mixed consumption patterns

A dynamic model called hybrid cybernetic model (HCM) based on structured metabolic network is established for simulating mammalian cell metabolism featured with partially substitutable and partially complementary consumption patterns of two substrates, glucose and glutamine. Benefiting from the application of elementary mode analysis (EMA), the complicated metabolic network is decomposed into elementary modes (EMs) facilitating the employment of the hybrid cybernetic framework to investigate the external and internal flux distribution and the regulation mechanism among them. According to different substrate combination, two groups of EMs are obtained, i.e., EMs associated with glucose uptake and simultaneous uptake of glucose and glutamine. Uptake fluxes through various EMs are coupled together via cybernetic variables to maximize substrate uptake. External fluxes and internal fluxes could be calculated and estimated respectively, by the combination of the stoichiometrics of metabolic networks and fluxes through regulated EMs. The model performance is well validated via three sets of experimental data. Through parameter identification of limited number of experimental data, other external metabolites are precisely predicted. The obtained kinetic parameters of three experimental cultures have similar values, which indicates the robustness of the model. Furthermore, the prediction performance of the model is successfully validated based on identified parameters.     

New insight into the role of the PhaP phasin of Ralstonia eutropha in promoting synthesis of polyhydroxybutyrate

Phasins are proteins that are proposed to play important roles in polyhydroxyalkanoate synthesis and granule formation. Here the phasin PhaP of Ralstonia eutropha has been analyzed with regard to its role in the synthesis of polyhydroxybutyrate (PHB). Purified recombinant PhaP, antibodies against PhaP, and an R. eutropha phaP deletion strain have been generated for this analysis. Studies with the phaP deletion strain show that PhaP must accumulate to high levels in order to play its normal role in PHB synthesis and that the accumulation of PhaP to low levels is functionally equivalent to the absence of PhaP. PhaP positively affects PHB synthesis under growth conditions which promote production of PHB to low, intermediate, or high levels. The levels of PhaP generally parallel levels of PHB in cells. The results are consistent with models whereby PhaP promotes PHB synthesis by regulating the surface/volume ratio of PHB granules or by interacting with polyhydroxyalkanoate synthase and indicate that PhaP plays an important role in PHB synthesis from the early stages in PHB production and across a range of growth conditions.     

Forecasting and Uncertainty Quantification Using a Hybrid of Mechanistic and Non-mechanistic Models for an Age-Structured Population Model

In this paper, we present a new method for the prediction and uncertainty quantification of data-driven multivariate systems. Traditionally, either mechanistic or non-mechanistic modeling methodologies have been used for prediction; however, it is uncommon for the two to be incorporated together. We compare the forecast accuracy of mechanistic modeling, using Bayesian inference, a non-mechanistic modeling approach based on state space reconstruction, and a novel hybrid methodology composed of the two for an age-structured population data set. The data come from cannibalistic flour beetles, in which it is observed that the adults preying on the eggs and pupae result in non-equilibrium population dynamics. Uncertainty quantification methods for the hybrid models are outlined and illustrated for these data. We perform an analysis of the results from Bayesian inference for the mechanistic model and hybrid models to suggest reasons why hybrid modeling methodology may enable more accurate forecasts of multivariate systems than traditional approaches.     

Extending, changing, and explaining the brain

This paper addresses concerns raised recently by Datteri (Biol Philos 24:301–324, 2009) and Craver (Philos Sci 77(5):840–851, 2010) about the use of brain-extending prosthetics in experimental neuroscience. Since the operation of the implant induces plastic changes in neural circuits, it is reasonable to worry that operational knowledge of the hybrid system will not be an accurate basis for generalisation when modelling the unextended brain. I argue, however, that Datteri’s no-plasticity constraint unwittingly rules out numerous experimental paradigms in behavioural and systems neuroscience which also elicit neural plasticity. Furthermore, I propose that Datteri and Craver’s arguments concerning the limitations of prosthetic modelling in basic neuroscience, as opposed to neuroengineering, rests on too narrow a view of the ways models in neuroscience should be evaluated, and that a more pluralist approach is needed. I distinguish organisational validity of models from mechanistic validity. I argue that while prosthetic models may be deficient in the latter of these explanatory virtues because of neuroplasticity, they excel in the former since organisational validity tracks the extent to which a model captures coding principles that are invariant with plasticity. Changing the brain, I conclude, is one viable route towards explaining the brain.     

Interaction between O-GlcNAc Modification and Tyrosine Phosphorylation of Prohibitin: Implication for a Novel Binary Switch

Prohibitin (PHB or PHB1) is an evolutionarily conserved, multifunctional protein which is present in various cellular compartments including the plasma membrane. However, mechanisms involved in various functions of PHB are not fully explored yet. Here we report for the first time that PHB interacts with O-linked β-N-acetylglucosamine transferase (O-GlcNAc transferase, OGT) and is O-GlcNAc modified; and also undergoes tyrosine phosphorylation in response to insulin. Tyrosine 114 (Tyr114) and tyrosine 259 (Tyr259) in PHB are in the close proximity of potential O-GlcNAc sites serine 121 (Ser121) and threonine 258 (Thr258) respectively. Substitution of Tyr114 and Tyr259 residues in PHB with phenylalanine by site-directed mutagenesis results in reduced tyrosine phosphorylation as well as reduced O-GlcNAc modification of PHB. Surprisingly, this also resulted in enhanced tyrosine phosphorylation and activity of OGT. This is attributed to the presence of similar tyrosine motifs in PHB and OGT. Substitution of Ser121 and Thr258 with alanine and isoleucine respectively resulted in attenuation of O-GlcNAc modification and increased tyrosine phosphorylation of PHB suggesting an association between these two dynamic modifications. Sequence analysis of O-GlcNAc modified proteins having known O-GlcNAc modification site(s) or known tyrosine phosphorylation site(s) revealed a strong potential association between these two posttranslational modifications in various proteins. We speculate that O-GlcNAc modification and tyrosine phosphorylation of PHB play an important role in tyrosine kinase signaling pathways including insulin, growth factors and immune receptors signaling. In addition, we propose that O-GlcNAc modification and tyrosine phosphorylation is a novel previously unidentified binary switch which may provide new mechanistic insights into cell signaling pathways and is open for direct experimental examination.     

Growth of microbial populations in nonminimal media: some considerations for modeling

Ramkrishna and his co-workers have developed so-called cybernetic models which purport to describe, among other things, how microorganisms make choices when presented with two or more functionally equivalent, or substitutable, nutrients that are sources of carbon and available energy. In general, however, organisms are presented with choices not just between nutrients that are substitutable for one another, but also between sets of nutrients some of which are by no means substitutable for one another. It is postulated herein that the main ideas of cybernetic modeling apply to these more general choices as they seem to apply to the choices considered by Ramkrishna and his co-workers. Some consequences of the postulate are worked out for steady-state growth situations where two, or in one case three, nutrients limit or potentially limit growth rate. If predicted phenomena are observed experimentally so as to verify the postulate, a significantly improved basis for understanding growth of microorganisms in practical fermentation media as well as in natural situations will be provided by this application of cybernetic modeling ideas.     

Phosphate removal under denitrifying conditions by Brachymonas sp. strain P12 and Paracoccus denitrificans PP15

In this study, we used the denitrifying phosphorus-removing bacterium Brachymonas sp. strain P12 to investigate the enhanced biologic phosphorus-removal (EBPR) mechanism involved with polyhydroxybutyrate (PHB), glycogen, and phosphorus uptake in the presence of acetate under anoxic or aerobic conditions. The results showed that excess acetate concentration and aerobic cultivation can enhance PHB formation efficiency and that PHB formation might be stimulated by glycogenolysis of the cellular glycogen. The efficiency of the uptake of anoxic phosphorus was greater when PHB production was lower. The EBPR mechanism of Brachymonas sp. strain P12 for PHB, phosphorus, and glycogen was similar to the conventional anaerobic-aerobic (or anaerobic-anoxic) EBPR models, but these models were developed under anoxic or aerobic conditions only, without an anaerobic stage. The anoxic or aerobic log phase of growth is divided into two main phases: the early log phase, in which acetate and glycogen are consumed to supply enough energy and reducing power for PHB formation and cell growth (phosphorus assimilation), and the late log phase, which ends the simultaneous degradation of PHB and remaining acetate for polyphosphate accumulation. Glycogenolysis plays a significant role in the alternate responses between PHB formation and phosphorus uptake under anoxic or aerobic conditions. After the application of the denitrifying phosphorus-removing bacterium Brachymonas sp. strain P12, aerobic cultivation increases the level of PHB production, and anoxic cultivation further increases phosphorus uptake.     

PHB/PLLA组织工程前交叉韧带支架材料改性的实验研究

目的:探索体外构建组织工程前交叉韧带(anterior cruciate ligament,ACL)的三维支架材料。方法:以聚羟基丁酸已酯/聚左旋乳酸(PHB/PLLA1:1)制备"三明治"样结构共聚物并测量其孔隙率等指标。以I型胶原对制备的PHB/PLLA支架进行杂化,获得PHB/PLLA胶原杂化支架。扫描电镜观察其表面结构。将兔皮肤成纤维细胞(SF)接种于PHB/PLLA支架与PHB/PLLA胶原杂化支架,观察其在材料上生长情况。结果:PHB/PLLA支架杂化后胶原填充于纤维空隙,分布比较均匀。体外培养的胶原杂化支架材料上要比PHB/PLLA支架有更多的皮肤成纤维细胞生长。结论:胶原杂化有利于细胞种植和生长,PHB/PLLA胶原杂化支架具有良好的三维构型和生物相容性,有望为前交叉韧带损伤的修复提供了一种新型的支架材料。