Limits...
Metabolic modeling of a chronic wound biofilm consortium predicts spatial partitioning of bacterial species

View Article: PubMed Central - PubMed

ABSTRACT

Background: Chronic wounds are often colonized by consortia comprised of different bacterial species growing as biofilms on a complex mixture of wound exudate. Bacteria growing in biofilms exhibit phenotypes distinct from planktonic growth, often rendering the application of antibacterial compounds ineffective. Computational modeling represents a complementary tool to experimentation for generating fundamental knowledge and developing more effective treatment strategies for chronic wound biofilm consortia.

Results: We developed spatiotemporal models to investigate the multispecies metabolism of a biofilm consortium comprised of two common chronic wound isolates: the aerobe Pseudomonas aeruginosa and the facultative anaerobe Staphylococcus aureus. By combining genome-scale metabolic reconstructions with partial differential equations for metabolite diffusion, the models were able to provide both temporal and spatial predictions with genome-scale resolution. The models were used to analyze the metabolic differences between single species and two species biofilms and to demonstrate the tendency of the two bacteria to spatially partition in the multispecies biofilm as observed experimentally. Nutrient gradients imposed by supplying glucose at the bottom and oxygen at the top of the biofilm induced spatial partitioning of the two species, with S. aureus most concentrated in the anaerobic region and P. aeruginosa present only in the aerobic region. The two species system was predicted to support a maximum biofilm thickness much greater than P. aeruginosa alone but slightly less than S. aureus alone, suggesting an antagonistic metabolic effect of P. aeruginosa on S. aureus. When each species was allowed to enhance its growth through consumption of secreted metabolic byproducts assuming identical uptake kinetics, the competitiveness of P. aeruginosa was further reduced due primarily to the more efficient lactate metabolism of S. aureus. Lysis of S. aureus by a small molecule inhibitor secreted from P. aeruginosa and/or P. aeruginosa aerotaxis were predicted to substantially increase P. aeruginosa competitiveness in the aerobic region, consistent with in vitro experimental studies.

Conclusions: Our biofilm modeling approach allows the prediction of individual species metabolism and interspecies interactions in both time and space with genome-scale resolution. This study yielded new insights into the multispecies metabolism of a chronic wound biofilm, in particular metabolic factors that may lead to spatial partitioning of the two bacterial species. We believe that P. aeruginosa lysis of S. aureus combined with nutrient competition is a particularly relevant scenario for which model predictions could be tested experimentally.

Electronic supplementary material: The online version of this article (doi:10.1186/s12918-016-0334-8) contains supplementary material, which is available to authorized users.

No MeSH data available.


Related in: MedlinePlus

Formulation and solution of the multispecies biofilm metabolic model. a Schematic representation of the chronic wound biofilm model of constant thickness W with glucose provided at the tissue-biofilm interface (z = 0), oxygen supplied at the biofilm-air interface (z = W) and the metabolic byproducts acetate, succinate and lactate removed at the tissue-biofilm interface. b Schematic representation of the biofilm metabolic model solution procedure. The multispecies biofilm with temporal and spatial variations is described by a spatiotemporal model that accounts for the diffusion of nutrients and byproducts. PDEs are written with respect to the bacterial species concentration (Xi) and the metabolite concentrations (Mj) assuming that spatial variations are limited to a single direction z. Lexicographic linear program solution of the genome-scale reconstruction of each species is performed to predict the growth rate, nutrient uptake rates and byproduct secretion rates. The PDEs are spatially discretized to yield a large-set of ODEs with embedded LPs that are integrated with the MATLAB code DFBAlab [78] to generate time and spatially resolved predictions
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC5015247&req=5

Fig1: Formulation and solution of the multispecies biofilm metabolic model. a Schematic representation of the chronic wound biofilm model of constant thickness W with glucose provided at the tissue-biofilm interface (z = 0), oxygen supplied at the biofilm-air interface (z = W) and the metabolic byproducts acetate, succinate and lactate removed at the tissue-biofilm interface. b Schematic representation of the biofilm metabolic model solution procedure. The multispecies biofilm with temporal and spatial variations is described by a spatiotemporal model that accounts for the diffusion of nutrients and byproducts. PDEs are written with respect to the bacterial species concentration (Xi) and the metabolite concentrations (Mj) assuming that spatial variations are limited to a single direction z. Lexicographic linear program solution of the genome-scale reconstruction of each species is performed to predict the growth rate, nutrient uptake rates and byproduct secretion rates. The PDEs are spatially discretized to yield a large-set of ODEs with embedded LPs that are integrated with the MATLAB code DFBAlab [78] to generate time and spatially resolved predictions

Mentions: Biofilm models were formulated by combining genome-scale reconstructions of P. aeruginosa and/or S. aureus metabolism with uptake kinetics for available nutrients and reaction-diffusion type equations for species biomass, supplied substrates and synthesized metabolic byproducts. Single species biofilm models were formulated with either the P. aeruginosa or S. aureus reconstruction, while the two species model used both reconstructions. Diffusion was assumed to occur only in the axial direction of the biofilm such that spatial variations could be captured with a single variable z (Fig. 1a). For simplicity, the biofilm was assumed to have a fixed thickness W over which the nutrients diffused and cell growth occurred. Therefore, the models were most appropriate for predicting the metabolism of biofilms of a specified thickness. Both strains were assumed to consume glucose as the primary carbon source [55]. Glucose was supplied at the tissue-biofilm interface at the assumed concentration of the wound exudate, while oxygen was supplied at the biofilm-air interface at a concentration for an aqueous solution in equilibrium with atmospheric oxygen.Fig. 1


Metabolic modeling of a chronic wound biofilm consortium predicts spatial partitioning of bacterial species
Formulation and solution of the multispecies biofilm metabolic model. a Schematic representation of the chronic wound biofilm model of constant thickness W with glucose provided at the tissue-biofilm interface (z = 0), oxygen supplied at the biofilm-air interface (z = W) and the metabolic byproducts acetate, succinate and lactate removed at the tissue-biofilm interface. b Schematic representation of the biofilm metabolic model solution procedure. The multispecies biofilm with temporal and spatial variations is described by a spatiotemporal model that accounts for the diffusion of nutrients and byproducts. PDEs are written with respect to the bacterial species concentration (Xi) and the metabolite concentrations (Mj) assuming that spatial variations are limited to a single direction z. Lexicographic linear program solution of the genome-scale reconstruction of each species is performed to predict the growth rate, nutrient uptake rates and byproduct secretion rates. The PDEs are spatially discretized to yield a large-set of ODEs with embedded LPs that are integrated with the MATLAB code DFBAlab [78] to generate time and spatially resolved predictions
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC5015247&req=5

Fig1: Formulation and solution of the multispecies biofilm metabolic model. a Schematic representation of the chronic wound biofilm model of constant thickness W with glucose provided at the tissue-biofilm interface (z = 0), oxygen supplied at the biofilm-air interface (z = W) and the metabolic byproducts acetate, succinate and lactate removed at the tissue-biofilm interface. b Schematic representation of the biofilm metabolic model solution procedure. The multispecies biofilm with temporal and spatial variations is described by a spatiotemporal model that accounts for the diffusion of nutrients and byproducts. PDEs are written with respect to the bacterial species concentration (Xi) and the metabolite concentrations (Mj) assuming that spatial variations are limited to a single direction z. Lexicographic linear program solution of the genome-scale reconstruction of each species is performed to predict the growth rate, nutrient uptake rates and byproduct secretion rates. The PDEs are spatially discretized to yield a large-set of ODEs with embedded LPs that are integrated with the MATLAB code DFBAlab [78] to generate time and spatially resolved predictions
Mentions: Biofilm models were formulated by combining genome-scale reconstructions of P. aeruginosa and/or S. aureus metabolism with uptake kinetics for available nutrients and reaction-diffusion type equations for species biomass, supplied substrates and synthesized metabolic byproducts. Single species biofilm models were formulated with either the P. aeruginosa or S. aureus reconstruction, while the two species model used both reconstructions. Diffusion was assumed to occur only in the axial direction of the biofilm such that spatial variations could be captured with a single variable z (Fig. 1a). For simplicity, the biofilm was assumed to have a fixed thickness W over which the nutrients diffused and cell growth occurred. Therefore, the models were most appropriate for predicting the metabolism of biofilms of a specified thickness. Both strains were assumed to consume glucose as the primary carbon source [55]. Glucose was supplied at the tissue-biofilm interface at the assumed concentration of the wound exudate, while oxygen was supplied at the biofilm-air interface at a concentration for an aqueous solution in equilibrium with atmospheric oxygen.Fig. 1

View Article: PubMed Central - PubMed

ABSTRACT

Background: Chronic wounds are often colonized by consortia comprised of different bacterial species growing as biofilms on a complex mixture of wound exudate. Bacteria growing in biofilms exhibit phenotypes distinct from planktonic growth, often rendering the application of antibacterial compounds ineffective. Computational modeling represents a complementary tool to experimentation for generating fundamental knowledge and developing more effective treatment strategies for chronic wound biofilm consortia.

Results: We developed spatiotemporal models to investigate the multispecies metabolism of a biofilm consortium comprised of two common chronic wound isolates: the aerobe Pseudomonas aeruginosa and the facultative anaerobe Staphylococcus aureus. By combining genome-scale metabolic reconstructions with partial differential equations for metabolite diffusion, the models were able to provide both temporal and spatial predictions with genome-scale resolution. The models were used to analyze the metabolic differences between single species and two species biofilms and to demonstrate the tendency of the two bacteria to spatially partition in the multispecies biofilm as observed experimentally. Nutrient gradients imposed by supplying glucose at the bottom and oxygen at the top of the biofilm induced spatial partitioning of the two species, with S. aureus most concentrated in the anaerobic region and P. aeruginosa present only in the aerobic region. The two species system was predicted to support a maximum biofilm thickness much greater than P. aeruginosa alone but slightly less than S. aureus alone, suggesting an antagonistic metabolic effect of P. aeruginosa on S. aureus. When each species was allowed to enhance its growth through consumption of secreted metabolic byproducts assuming identical uptake kinetics, the competitiveness of P. aeruginosa was further reduced due primarily to the more efficient lactate metabolism of S. aureus. Lysis of S. aureus by a small molecule inhibitor secreted from P. aeruginosa and/or P. aeruginosa aerotaxis were predicted to substantially increase P. aeruginosa competitiveness in the aerobic region, consistent with in vitro experimental studies.

Conclusions: Our biofilm modeling approach allows the prediction of individual species metabolism and interspecies interactions in both time and space with genome-scale resolution. This study yielded new insights into the multispecies metabolism of a chronic wound biofilm, in particular metabolic factors that may lead to spatial partitioning of the two bacterial species. We believe that P. aeruginosa lysis of S. aureus combined with nutrient competition is a particularly relevant scenario for which model predictions could be tested experimentally.

Electronic supplementary material: The online version of this article (doi:10.1186/s12918-016-0334-8) contains supplementary material, which is available to authorized users.

No MeSH data available.


Related in: MedlinePlus