A computational model of intracellular oxygen sensing by hypoxia-inducible factor HIF1

JCS03087 Supplementary Material

Files in this Data Supplement:

• Supplemental Figure 1 -

  Fig. S1. Using the newly reported K_m value of 0.03 μM for binding of iron to PHD2 (Hirsila et al., 2005), the model predicts a maximum specific activity of 411 mol/mol/minute calculated from the slope of the tangent to the curve in (i), during the first 6 minutes of the hydroxylation reaction. This is tenfold greater than predicted by Hirsila et al. (Hirsila et al., 2005), where minimum PHD2 specific activity is estimated as 40-50 mol/mol/minute. An exhaustive search for parameter sets using K_m,_Fe2=0.03 μM found no binding on and off-rates that represented well both HIF1α half-lives and PHD2 specific activity reported in literature. The closest match, yielding a HIF1α half-life of 8 minutes (ii), was found using the rates shown in (i).

• Supplemental Figure 2 -

  Fig. S2. Calculations of half-lives (i, iii) and PHD2 specific activity (ii, iv) from the model, with different Michaelis-Menten constant K_m values. Sensitivity analysis was performed for all kinetic parameters found from experiments, and comparisons were made to experiments on HIF1α half-life (Berra et al., 2001) and PHD2 specific activity (Tuckerman et al., 2004). Examples are shown for K_m,_Fe2 (i,ii) and K_m,_DG (iii,iv).

• Supplemental Figure 3 -

  Fig. S3. Relative HIF1α hydroxylated at different O₂ levels after 5 and 20 minutes, for different values of K_ms. K_m is the experimentally determined Michaelis-Menten constant. Examples are shown for iron, 0.025K_m,_Fe2 to 2 K_m,_Fe2 (i) and for 2-oxoglutarate, 0.2K_m,_DG to 2K_m,_DG (ii).

• Supplemental Figure 4 -

  Fig. S4. Sensitivity analysis to determine parameter ranges: When other kinetic rates are kept at default values, kinetic off-rates corresponding to the minimum HIF1α half-lives of 5 and 8 minutes respectively are: k_cat,_Hα=0.098-0.164 minute^-1 (i), k_off,_Fe2=0.019-36 minute^-1 (ii), k_off,_DG=0.044-10.8 minute^-1 (iii), k_off,_O2=0.044-10.8 minute^-1 (iv), and k_off,_AS=0.0036-3.6 minute^-1 (v). Values of k_off,_DG and k_off,_O2 above 10.8 minute^-1 yield the same response curve as that for 10.8 minute^-1. (vi) The off-rate for unhydroxylated HIF1α corresponding to a minimum HIF1α half-life of 5 min is k_off,_Hα=0.7 minute^-1; the sensitivity to this parameter is low, with all response curves intersecting beyond 6 minutes.

• Supplemental Figure 5 -

  Fig. S5. Comparison of initial reaction rate to experiments was used to narrow the kinetic parameter ranges. (i) k_cat,_Hα=0.098-0.164 minute^-1 corresponds to minimum HIF1α half-lives of 5 and 8 minutes respectively, and this range compares well with the specific activity from (Tuckerman et al., 2004). This k_cat,_Hα range was used in calculations of subsequent kinetic parameters. (ii) From the minimum range of 0.019-36 minute^-1 calculated for k_off,_Fe2 from HIF1α half-lives, k_off,_Fe2=36 minute^-1 corresponds best to the experimentally-determined specific activity. (iii) k_off,_DG=10.8 minute^-1 is the analogous value for 2-oxoglutarate unbinding to PHD2. (iv) For O₂, this value is k_off,_O2=10.8 minute^-1 and for ascorbate, k_off,_AS=3.6 minute^-1 (v). k_off,_Hα=0.7 minute^-1 was the best value within the range estimated from half-life comparisons, that correlates with the specific activity data shown (vi).

• Supplemental Figure 6 -

  Fig. S6. Sensitivity of HIF1α hydroxylation to concentrations of reactants. The upper curve in both graphs corresponds to initial concentrations of unhydroxylated HIF1α₀=1 μM; PHD2₀=1 μM; and Fe2⁺₀=50 μM, as frequently used by in vitro experiments. (i) Decreasing Fe²⁺ results in a large reduction of hydroxylation. In vivo, total cellular Fe²⁺ concentrations vary from 20-200 μM (Arredondo et al., 1997; Hirsila et al., 2005). The fraction that is freely available for binding to PHD2 depends on cell type; in cultured insect cells used for HIF1α experiments, it has been estimated as 0.3-0.8 μM (without addition of Fe²⁺ to the culture) (Esposito et al., 2002). (ii) Decreasing PHD2₀ also results in a steep reduction of hydroxylation. In vivo levels of PHD2, if they can be approximated by in vitro cell extracts, are on the order of 4 nM (Tuckerman et al., 2004). Compared with Fe²⁺ and PHD2, ascorbate has minimal effects on each cycle of the hydroxylation reaction; see Figure 2.