Determinants for simultaneous binding of copper and platinum to human chaperone Atox1: hitchhiking not hijacking

Abstract

Cisplatin (CisPt) is an anticancer agent that has been used for decades to treat a variety of cancers. CisPt treatment causes many side effects due to interactions with proteins that detoxify the drug before reaching the DNA. One key player in CisPt resistance is the cellular copper-transport system involving the uptake protein Ctr1, the cytoplasmic chaperone Atox1 and the secretory path ATP7A/B proteins. CisPt has been shown to bind to ATP7B, resulting in vesicle sequestering of the drug. In addition, we and others showed that the apo-form of Atox1 could interact with CisPt in vitro and in vivo. Since the function of Atox1 is to transport copper (Cu) ions, it is important to assess how CisPt binding depends on Cu-loading of Atox1. Surprisingly, we recently found that CisPt interacted with Cu-loaded Atox1 in vitro at a position near the Cu site such that unique spectroscopic features appeared. Here, we identify the binding site for CisPt in the Cu-loaded form of Atox1 using strategic variants and a combination of spectroscopic and chromatographic methods. We directly prove that both metals can bind simultaneously and that the unique spectroscopic signals originate from an Atox1 monomer species. Both Cys in the Cu-site (Cys12, Cys15) are needed to form the di-metal complex, but not Cys41. Removing Met10 in the conserved metal-binding motif makes the loop more floppy and, despite metal binding, there are no metal-metal electronic transitions. In silico geometry minimizations provide an energetically favorable model of a tentative ternary Cu-Pt-Atox1 complex. Finally, we demonstrate that Atox1 can deliver CisPt to the fourth metal binding domain 4 of ATP7B (WD4), indicative of a possible drug detoxification mechanism.

Figure 1. The copper chaperone Atox1.

Cartoon model of Atox1 made in Ccp94 using PDB 1TL4 . Stick residues are the two cysteines (Cys12 and Cys15) in the Cu-binding site, the 3:rd cysteine in the protein (Cys41) and methionine (Met10) in the conserved MxCxxC motif.

Figure 2. Near-UV CD for Atox1 variants and WD4.

Selected traces from Cu/CisPt titrations. Black: Apo-protein. Grey: +1 eq. Cu. Blue: +1 eq. Cu and +2 eq. CisPt. Red: +1 eq. CisPt +2 eq. Cu. A. WT Atox1. B. Cys41Ala Atox1. C. 3Cys3Ala Atox1. D. Cys15Ala Atox1. E. Met10Ala Atox1. F. WT WD4. See also Figure S1.

Figure 3. CisPt-triggered unfolding of Atox1 variants and WD4.

Black: Apo-protein. Grey: +1 eq. Cu. Blue: +1 eq. Cu and +5 eq. CisPt. Red: +5 eq. CisPt. A. WT Atox1. B. Cys41Ala Atox1. C. 3Cys3Ala Atox1. D. Cys15Ala Atox1. E. Met10Ala Atox1. F. WT WD4.

Figure 4. SDS-gel analysis of CisPt induced Atox1 aggregation.

Apo- and +1 eq. Cu samples treated with 5 eq. CisPt over the time of 4d, 2d, 1d, 4h and fresh made. A. WT Atox1. B. Cys41Ala Atox1. C. 3Cys3Ala Atox1. D. Cys15Ala Atox1. E. Met10Ala Atox1. F. WT WD4. See also Figure S2.

Figure 5. No loss of Cu upon CisPt binding to Atox1 WT.

Samples treated with 1∶1 of CisPt and incubated for various times. Samples were centrifuged and Cu concentration was measured in flow trough (cut off in filter 3000 Da). Dark grey: CisPt treated holo Atox1. Light gray: Holo-Atox1. The last two columns are positive controls where Atox1 was omitted.

Figure 6. Analytical gelfiltration of Atox1/WD4.

Protein +1 eq. Cu and 5 eq. CisPt. Black: 280 nm. Grey: 254 nm. A. WT Atox1. B. Cys41Ala Atox1. C. 3Cys3Ala Atox1. D. Cys15Ala Atox1. E. Met10Ala Atox1. F. WT WD4. Sample incubation time is 10 min. See also Figure S3–S8.

Figure 7. Near-UV CD of peaks from analytical gelfiltration.

Atox1+1 eq. Cu +3 eq. CisPt. Green: Monomeric peak (14.2 ml). Yellow: Dimeric peak (12.8 ml). Grey: Atox1+1 eq. Cu +2 eq. CisPt. Data is normalized to correspond to 50 µM protein in all three cases.

Figure 8. Geometry optimized model of Cu-loaded Atox1 with CisPt.

Low-energy complex (A2) between Cu-Atox1 and CisPt⁺ calculated using DFT-D3. CisPt was positioned next to Cys12 in the holo-form, since this is the more accessible of the two cysteines and thus the most likely interaction site for Pt. Pt-Cu metal-metal interaction bond is shown in light purple. See also supporting information.

Figure 9. Pt-transfer from Atox1 to WD4.

Analytical gelfiltration, Black: 280 nm, Grey: 254 nm. A. Apo-Atox1 mixed with 0.5 eq. apo-WD4. B. Control, WD4. No Atox1-CisPt added, otherwise experiment conducted as D. C. Control, Atox1-CisPt. No WD4 added, otherwise experiment conducted as D. D. Transfer experiment. Atox1-CisPt, filtrated and concentrated, mixed with 0.5 eq. WD4 and incubated for 4 h prior to SEC analysis. See also Figure S9.