Toluene diisocyanate reactivity with glutathione across a vapor/liquid interface and subsequent transcarbamoylation of human albumin

Abstract

Glutathione has previously been identified as a reaction target for toluene diisocyanate (TDI) in vitro and in vivo, and has been suggested to contribute to toxic and allergic reactions to exposure. In this study, the reactivity of reduced glutathione (GSH) with TDI in vitro was further investigated using a mixed phase (vapor/liquid) exposure system to model the in vivo biophysics of exposure in the lower respiratory tract. HPLC/MS/MS was used to characterize the observed reaction products. Under the conditions tested, the major reaction products between TDI vapor and GSH were S-linked bis(GSH)-TDI and to a lesser extent mono(GSH)-TDI conjugates (with one N═C═O hydrolyzed). The vapor-phase-generated GSH-TDI conjugates were capable of transcarbamoylating human albumin in a pH-dependent manner, resulting in changes in the self-protein's conformation/charge, on the basis of electrophoretic mobility under native conditions. Specific sites of human albumin-TDI conjugation, mediated by GSH-TDI, were identified (Lys(73), Lys(159), Lys(190), Lys(199), Lys(212), Lys(351), Lys(136/137), Lys(413/414), and Lys(524/525)) along with overlap with those susceptible to direct conjugation by TDI. Together, the data extend the proof-of-principle for GSH to act as a "shuttle" for a reactive form of TDI, which could contribute to clinical responses to exposure.

Figure 1

Depiction of methodology used in the study. Panel A (top) depicts vapors of reactive TDI interacting with GSH in solution. Panel B (bottom) depicts the co-incubation of GSH-TDI reaction products with human albumin in solution.

Figure 2

HPLC separation of GSH-TDI reaction products. The absorbance at 210 nm (top) and 245 nm (bottom) was measured (Y-axis) for each 1 min eluent sample collected over time (X-axis). Samples of TDI vapor (<200 ppb) exposed GSH (10mM), shown as blue diamonds, were compared with “mock” exposed GSH, shown as red squares, to identify TDI reaction products, highlighted as peaks A-D, and unreacted GSH (peak ∅).

Figure 3

Mass spectrometry analysis of HPLC fractions containing GSH-TDI reaction products. Each panel A through B corresponds to samples from peaks A through D in Figure 2. Tandem MS/MS data (X-axis = m/z) and deduced structures for the singly charged m/z 456.2, the doubly and singly charged m/z 395.2 and 789.4 are shown in Figures 4 through 6.

Figure 4

Tandem mass spectrometry analysis of peak A from HPLC purified GSH-TDI reaction products. A sample of peak A from the HPLC fractionation of GSH exposed to TDI vapor was analyzed by MS/MS (X-axis = m/z), focusing on the major m/z 456.2 product. The MS/MS spectrum shows a major fragment with an m/z consistent with the Gly-Cys-TDI* structure shown in Figure 6, suggesting linkage of TDI via the thiol of Cys, rather than the glutamate amino acid of GSH. Identical results were obtained with peak B from the HPLC fractionation shown in Fig. 2.

Figure 5

Tandem mass spectrometry analysis of peak C from HPLC purified GSH-TDI reaction products. A sample of peak C from the HPLC fractionation of GSH exposed to TDI vapor was analyzed by MS/MS (X-axis = m/z), focusing on the major m/z 789.4 product. The MS/MS spectrum shows a major fragment with an m/z consistent with the Gly-Cys-TDI-GSH, Gly-Cys-TDI-Cys-Gly, and Gly-Cys-TDI structures shown in Figure 6, suggesting linkage of TDI via the thiol of Cys, rather than the glutamate amino acid of GDH. Identical results were obtained with peak D from the HPLC fractionation shown in Fig. 2.

Figure 6

Chemical structures of the major reaction products between TDI vapors and GSH in solution. The structure of the major GSH-TDI reaction products and their predicted ionization fragments, including the m/z of the singly charged species, consistent with the observed MS/MS data shown in Figures 4 and 5.

Figure 7

Native gel analysis of human albumin transcarbamoylated by purified GSH-TDI reaction products. Human albumin co-incubated with HPLC purified GSH-TDI reaction products, or controls (labeled above the lane) was subjected to electrophoresis under native conditions and stained for protein. A downward shift in electrophoresis is consistent with TDI conjugation based on previous published data. All co-incubations used albumin in 0.1 M carbonate pH 8.8, plus the following: Lane 1- GSH from HPLC Peak ∅, Lanes 2 and 3- mono(GSH)-TDI* from HPLC peaks A and B respectively, Lanes 4 and 5- bis(GSH)-TDI from HPLC peaks C and D respectively, Lane 6- mock exposed GSH (unfractionated), Lane 7- TDI vapor exposed GSH (unfractionated), Lane 8-mock exposed albumin, Lane 9- albumin exposed directly to TDI vapors.

Figure 8

Dependence of human albumin transcarbamoylation by GSH-TDI on pH and free thiol of cysteine. Results of transcarbamoylation reactions of human albumin with total unfractionated GSH-TDI or GSSG-TDI reaction products. Co-incubations were performed at different pH levels (as labeled) and analyzed by gel electrophoresis (Panels A and C), or anti-TDI Western blot (Panels B and D). *Note- A/B, and C/D are paired gel / Western blots. Protein-stained gels were run under native conditions to highlight shift in migration due to conformational/charge differences, while Western blots were run under reducing conditions to maximize anti-TDI mAb binding.