BioModels Database logo

BioModels Database


Petelenz-Kurdzeil et al., (2013). Quantitative Analysis of Glycerol Accumulation, Glycolysis and Growth under Hyper Osmotic Stress

November 2016, model of the month by Nick Juty
Original model: BIOMD0000000603, BIOMD0000000604, BIOMD0000000605, BIOMD0000000606, BIOMD0000000607,BIOMD0000000610


A plethora of mechanisms exist by which cells handle environmental challenges and osmotic shock. Osmoregulation is a homeostatic process that ensures cellular integrity as well as maintaining optimal intracellular biochemical composition, achieved ultimately through manipulation of cellular/external water concentrations.

In yeast, a sudden increase in environmental osmolarity results in water leaving the cells, which shrink as a consequence. The cell adapts to this challenge by:

  1. accumulating glycerol through activation of the High Osmolarity Glycerol (HOG) pathway
  2. preventing glycerol outflow through closure of a glycerol facilitator Fps1
  3. a potential contribution to glycerol production through metabolic process/flux refinement
  4. up-regulation of the glycerol/proton symporter Stl1

The first enzymatic step in the glycerol production pathway converts dihydroxylacetone phosphate (DHAP) to glycerol-3-phosphate (G3P). There are two enzymes that are able to perform this conversion, with glycerol-3-phosphate dehydrogenase (GPD1) (Figure 1) expressed under osmoshock conditions, and the other (GPD2) controlled by cellular redox conditions.

This study [1] integrates and quantifies the relative contribution to osmoregulation through different and distinct mechanisms, in order to characterise the interplay between:

  1. expression of GPD1 (1st step in glycerol production)
  2. activation of glycolytic fluxes through phosphofructokinase (Pfk26)
  3. regulation of glycerol transport through Fps1
  4. effects of concentration and volume changes
  5. rerouting of overall metabolic flux towards glycerol


The model describes the osmoregulation in wildtype and in several mutant conditions. This work, an extension based on the authors' previous studies [2,3], used an extensive dataset including experimental data on a range of metabolites and proteins from both wild-type and multiple mutant strains. The model can loosely be described as an amalgamation of the following modular components:

  1. biophysical, concerning volume, surface area, osmotic pressure, ...
  2. glycolytic, comprising lumped reactions originating from glucose
  3. transport, including inputs and products such as glucose, ethanol, glycerol via Fps1, ...
  4. biomass, a measure of growth through cell number or density to evaluate the model
  5. adaption, including effects such as Hog1 phosphorylation, changes in transcription/translation, activation, ...

Each of the models used in this study are available in BioModels. Specifically, these models represent BIOMD0000000603 a wild-type osmoregulation, while the individual mutants and strains are represented by models BIOMD0000000604, BIOMD0000000605, BIOMD0000000606, BIOMD0000000607, BIOMD0000000610

Over the course of this study, intracellular and extracellular metabolite concentrations are measured (HPLC) over 4 hr assays, permitting estimation of model parameters. Over this period, external glucose is consumed, with glycerol and other byproducts (ethanol, acetate, ...) produced, with a concomitant increase in Optical Density (OD) and cell numbers increased.

Contributions to the overall osmoadpation response were characterised by determining absolute flux through glycerol for both wild-type and gpd mutant, as well as relative contributions over time (Figure 2). Ternary plots distinguished Fps1-reliant (glycerol facilitator), Gpd1-reliant (glycerol production 1st enzymatic step), and 'others' relative contributions, where 'others' included changes such as those in volume or turgor. In wild-type cells, adaption is mainly due to volume changes and closure of Fps1 (Figure 2A), with Fps reopening after around 30 mins allowing glycerol efflux. A low basal level of sustained glycerol production (Figure 2B) is evident in the reduced Hog1-dependent Gpd1 mutant. The relative contributions of each are shown (Figure 2C), where the colour coding on the path is as indicated for the time periods depicted in Figure 2A (X-axis).

Figure 2

Figure 2. Relative contribution to glycerol accumulation mechanisms
Absolute fluxes to glycerol are shown for wild-type (A) and gpd1Δ (B). Contributions to the total flux are indicated for the different mechanisms (Gpd1, glycerol-3-phosphate dehydrogenase; basal, basal glycerol production; Stl1, glycerol/proton symporter; Fps1, glycerol facilitator; volume, volume changes of cell). Ternary plots (C) use the colours use those indicated on the x-axis of (A).Figure taken from [1].

Since the final relative flux compositions approached, but did not match, pre-stress conditions, an additional analysis was performed which included consideration of the state of the cells, including indirectly affected reactions. This used scaled time-dependent response coefficients (RCs), which reflect the effect of a parameter perturbation on a time course; a positive value indicates a time course increase when a parameter has been increased. Through such analysis, it was found that Pfk26/27 (phosphofructokinase, Pfk) itself does not have a major effect on intracellular glycerol. Furthermore, downstream reactions were positively affected, while upstream reactions were negatively affected, indicating that Pfk may be a key component in rerouting metabolic flux downstream of itself. This would ensure ATP-production downstream of pyruvate is continued during the osmoadaption period.

Biomass regulation (Figure 3) can be seen to contribute to glycerol accumulation. As shown (Figure 3A), the doubling time for all cells tested increases following osmoshock, suggesting that there is a cost incurred by the cell in maintaining a certain cell volume, resulting in a balance between glycerol accumulation and biomass production. This trade off is depicted (Figure 3B), showing model simulations of carbon flux between these two processes for wild-type and mutant strains.


This paper illustrates an increasingly useful and common approach, integrating both modeling and experimental work, in a data-driven manner, and utilising a wide spectrum of available genetic resources, such as mutants and deficient strains.

This work demonstrates that osmoadaption in yeast involves short-term metabolic responses, membrane transport changes, as well as longer term transcription/translational responses. The temporal contributions of different types of response were studied, and it seems clear that osmoadaptation prioritizes energy balance in glycolysis, with the energy cost of osmoadaption being taken through the rerouting of flux from biomass production to glycerol accumulation. Robustness of the network was also studied, along with a potential role for trehalose in yeast osmoadaptation.

Mammalian stress-activated protein kinase p38 is a homologue of Hog1 and plays an analogous role in adaptive response. Since response to osmomotic shock is in principle conserved across many species, the approaches and techniques employed in this work may be useful for similar studies in other mammalian systems.

Figure 1

Figure 1. Model topology visualised using SBGN.
An SBGN (Systems Biology Graphical Notation) representation of the model is shown. The model is composed of an integrated set of different colour-coded modules: red, 'adaptation'; yellow, 'biophysical'; brown, 'transport'; blue, 'glycolytic'; green, 'biomass'. Metabolic components whose concentrations were measured during the study are indicated using a green background. Mutant strains used are also indicated: I, hog1Δ; II, pfk26/27Δ; III, HOG1-att; IV, gpd1Δ; V, FPS1-Δ1. Figure taken from the supplementary material S10 of [1].

Figure 3

Figure 3. Stress effect on growth rate
A: Bar graph comparing in vivo doubling times of wild-type and multiple mutant strains (see text), with '-' indicating pre-stress growth rate, and '+' indicating growth following stress through addition of 0.4 M NaCl.
B: Simulation of flux towards biomass production and towards glycerol production in wild-type and mutant strains. The 3 bars for each strain indicate time periods of 0, 20, and 90 minutes following stress onset (A). These indicate a correlation between insufficient glycerol accumulation (osmoprotection) and a prolonged decrease in growth rate. Figure taken from [1]


  1. Petelenz-Kurdzeil et al. Quantitative Analysis of Glycerol Accumulation, Glycolysis and Growth under Hyper Osmotic Stress. PLoS Comput Biol. 2013 June; 9(6): e1003084.
  2. Kuhn et al. Exploring the impact of osmoadaptation on glycolysis using time-varying response-coefficients. Genome Inform. 2008 20: 77-90.
  3. Klipp et al. Integrative model of the response of yeast to osmotic shock. Nat Biotechnol. 2005 23: 975-982.