Supplementary material: time-resolved fluorescence of Bis-ANS and DCVJ with polysorbate-containing IgG formulations

Molecular rotors DCVJ and CCVJ

The fluorescence decays of 5 µM DCVJ in 1.0 mg/ml non-stressed and heat-stressed IgG, with or without polysorbate 20, as well as in the corresponding buffers were comparably short (Figure 1). Best fitting results with respect to Chi2 and residuals of the fit (data not shown) could be obtained with three exponentials for the placebo and the non-stressed IgG and four exponentials for the 10 min 80°C stressed IgG (Table 1).

The decays of DCVJ for the placebos and the non-stressed IgG formulations were difficult to analyze because of the very short lifetime component (tau1) relative to the instrumental response function. Independent of the polysorbate concentration, a rapid component with a lifetime below 0.10 ns, which might in part be attributed to light scattering, was found. To facilitate the fitting, this rapid component, tau1, was fixed to 0.05 ns. The exact value of tau1 (if fixed to <0.10 ns) hardly affected the outcome of the other parameters to be fitted (results not shown). The short lifetime component was dominating the fits of the placebos and the non-stressed IgG, with about 90% for the polysorbate-free and about 70% for the polysorbate 20-containing formulations. Tau2 and tau3 for these formulations were shorter for the polysorbate 20-containing formulations than for the polysorbate-free preparations.

For the heat-stressed formulations, with or without polysorbate 20, four exponentials were required to obtain good fitting results. The major difference between the fits was found in the longest lifetime component tau4, which was longer for the polysorbate-free formulation (5.8 ns) than for that with polysorbate 20 (5.2 ns). Moreover, the fractional contributions of tau3 and tau4 were somewhat smaller for the polysorbate-containing formulation. Consequently, the average lifetimes were shorter for the polysorbate 20-containing formulations compared to the polysorbate-free formulations. This is in qualitative agreement with the observed lower fluorescence intensities in steady state fluorescence for DCVJ in heat-stressed IgG formulations with polysorbate.

In conclusion, time-resolved fluorescence of DCVJ confirmed the results from steady state fluorescence. Time-resolved studies with polysorbate 80, as well as with CCVJ instead of DCVJ showed similar results (data not shown).

Polarity sensitive dye Bis-ANS

In polysorbate-free formulations the fluorescence decays of Bis-ANS were relatively rapid for formulation buffer and non-stressed IgG, but clearly slower for heat-stressed IgG (Figure 2A). For polysorbate 20-containing formulations slower decays were found also for the placebo and the non-stressed formulation, which were further increased only to a minor degree for the heat-stressed preparation (Figure 2B). Table 2 shows the results from individual fitting of the lifetime decays when using the minimal number of exponentials (between 1 and 3, depending on the sample) required for obtaining good fits. For polysorbate-free formulations a short lifetime of 0.14 ns, derived from the unbound dye, dominated the decays of the placebo and of the non-stressed IgG with fractional contributions of 87.3% and 55.9%, respectively. A weak interaction of Bis-ANS with the non-stressed IgG became apparent from the appearance of a moderately long lifetime component (ca. 4.4 ns). The decay of polysorbate 20-containing placebo and non-stressed IgG could be fitted with one lifetime of 5.5-5.6 ns. This lifetime is characteristic of Bis-ANS in presence of polysorbate 20, and was comparable for polysorbate 80 (5.6-5.7 ns, data not shown). For heat-stressed IgG tau2 was 5.9 ns (without polysorbate) or 5.5 ns (with polysorbate 20) and tau3 was > 10.5 ns. Whereas tau2 has to be considered as resulting from both polysorbate 20 and heat-stressed IgG associated with Bis-ANS, the long (> 10.5 ns) lifetime component can be attributed to heat-stressed IgG interacting with Bis-ANS.

In conclusion, it was possible to distinguish between unbound dye (lifetime of 0.14 ns), dye bound to polysorbate (lifetime 5.5 ns) and dye bound to heat-stressed IgG (lifetime > 10.5 ns) by using time-resolved fluorescence of Bis-ANS. Also in polysorbate 80-containing formulations it was possible to pick up changes induced by heat, by the appearance of the long lifetime component of 10.5 ns (results not shown).

Figure 1: Fluorescence decays of 5 mM DCVJ (Excitation: 450 nm, Emission: 495 nm cut-on filter) without polysorbate 20 (PS20) (A) and with 0.02% (w/v) polysorbate 20 (B) for buffer, non-stressed IgG (NS) and 10 min 80°C heat-stressed IgG.

Table 1: Fluorescence decay parameters for DCVJ in the different formulations*

sample / tau1 [ns] / tau2 [ns] / tau3 [ns] / tau4 [ns] / f1 [%] / f2 [%] / f3 [%] / f4 [%] / average lifetime [ns] / Chi2
buffer, 0% PS20 / 0.05** / 0.42 ± 0.11 / 3.89 ± 0.02 / - / 88.8 ± 0.9 / 8.3 ± 3.2 / 2.9 ± 2.4 / - / 0.19 / 1.11
NS, 0% PS20 / 0.05** / 0.54 ± 0.14 / 3.91 ± 0.02 / - / 89.9 ± 0.7 / 5.8 ± 2.2 / 4.3 ± 0.4 / - / 0.24 / 1.15
10 min 80°C, 0% PS20 / 0.05** / 0.45 ± 0.13 / 1.57 ± 0.02 / 5.84 ± 0.01 / 29.3 ± 2.2 / 27.2 ± 9.8 / 33.9 ± 2.5 / 9.5 ± 1.1 / 1.23 / 1.01
buffer, 0.02% PS20 / 0.05** / 0.08 ± 0.67 / 2.49 ± 0.02 / - / 65.2 ± 11.9 / 30.8 ± 26.1 / 3.97 ± 0.16 / - / 0.15 / 1.15
NS, 0.02% PS20 / 0.05** / 0.10 ± 2.9 / 2.31 ± 0.01 / - / 78.4 ± 17.8 / 17.3 ± 5.2 / 4.3 ± 0.2 / - / 0.16 / 1.09
10 min 80°C, 0.02% PS20 / 0.05** / 0.42 ± 0.16 / 1.54 ± 0.025 / 5.16 ± 0.01 / 38.1 ± 2.0 / 23.5 ± 11.5 / 30.6 ± 2.31 / 7.9 ± 1.1 / 0.99 / 1.09

* Resulting from fits using the minimal number of exponents yielding a good fit (with respect to Chi2 and residuals). Shown are individual lifetime components (tau1-4), the corresponding fractional contributions (f1-4) including the fitting error and average lifetime and Chi2.

**tau1 fixed to 0.05 ns (see text).

Figure 2: Fluorescence decays of 5 mM Bis-ANS (Excitation: 375 nm, Emission: 495 nm cut-on filter) without polysorbate 20 (PS20) (A) and with 0.02% (w/v) polysorbate 20 (B) for buffer, non-stressed IgG (NS) and 10 min 80°C heat-stressed IgG.

Table 2: Fluorescence decay parameters for Bis-ANS in the different formulations*

sample / tau1 [ns] / tau2
[ns] / tau3 [ns] / f1
[%] / f2
[%] / f3
[%] / average lifetime [ns] / Chi2
buffer, 0% PS20 / 0.14** / 2.74 ± 0.008 / - / 87.3 ± 2.2 / 12.7 ± 0.36 / - / 0.46 / 1.10
NS, 0% PS20 / 0.14** / 1.18 ± 0.003 / 4.42 ± 0.004 / 55.9 ± 0.44 / 16.5 ± 1.05 / 27.6 ± 1.13 / 1.49 / 1.07
10 min 80°C, 0% PS20 / - / 5.94 ± 0.002 / 10.71 ± 0.001 / - / 48.8 ± 1.46 / 51.2 ± 2.60 / 8.38 / 1.06
buffer, 0.02% PS20 / 5.59 ± 0.001 / - / 5.59 / 1.05
NS, 0.02% PS20 / 5.56 ± 0.001 / - / 5.56 / 1.09
10 min 80°C, 0.02% PS20 / 5.67 ± 0.001 / 11.34 ± 0.002 / 85.5 ± 0.79 / 14.5 ± 1.51 / 6.49 / 1.06

* Resulting from fits using the minimal number of exponents yielding a good fit (with respect to Chi2 and residuals). Shown are individual lifetime components (tau1-3), the corresponding fractional contributions (f1-3) including the fitting error and average lifetime and Chi2.

**tau1 fixed to 0.14 ns.

5