Copper-free click chemistry for dynamic in vivo imaging

Abstract

Dynamic imaging of proteins in live cells is routinely performed by using genetically encoded reporters, an approach that cannot be extended to other classes of biomolecules such as glycans and lipids. Here, we report a Cu-free variant of click chemistry that can label these biomolecules rapidly and selectively in living systems, overcoming the intrinsic toxicity of the canonical Cu-catalyzed reaction. The critical reagent, a substituted cyclooctyne, possesses ring strain and electron-withdrawing fluorine substituents that together promote the [3 + 2] dipolar cycloaddition with azides installed metabolically into biomolecules. This Cu-free click reaction possesses comparable kinetics to the Cu-catalyzed reaction and proceeds within minutes on live cells with no apparent toxicity. With this technique, we studied the dynamics of glycan trafficking and identified a population of sialoglycoconjugates with unexpectedly rapid internalization kinetics.

Fig. 1.

Design and synthesis of Cu-free click chemistry reagents. (A) The copper-catalyzed azide–alkyne cycloaddition. (B) The Cu-free click reaction of azides and DIFOs. (C) (i) Sodium hydride, allyl bromide, 52%. (ii) Pyridinium chlorochromate, 91%. (iii) Lithium hexamethyldisilazide (LHMDS), chlorotriethylsilane (TESCl), 91%. (iv) Selectfluor, 95%. (v–vii). Cat. LHMDS, 95%. (viii) Potassium hexamethyldisilazide (KHMDS), TESCl, 97%. (ix) Selectfluor, 74%. (x) Cat. RuCl₃, NaIO₄, 96%. (xi) KHMDS, N-phenylbis(trifluoromethanesulfonamide), 46%. (xii) Lithium diisopropylamide, 11%. (D) Derivatives of DIFO and a linear alkyne (alk) containing Alexa Fluor 488, Alexa Fluor 568, or biotin.

Fig. 2.

Comparison of Cu-free click chemistry with existing bioorthogonal ligations. (A) Reactions of 10 ng of azidohomoalanine-labeled DHFR with 25 μM DIFO-488 or alk-488 were allowed to proceed for the time indicated. Reactions with alk-488 were performed as described in ref. . A negative control reaction (−) using 10 ng of azide-free DHFR was allowed to proceed for 60 min. (B) Schematic for metabolic labeling and detection of cell-surface glycans using Ac₄ManNAz and DIFO-based reagents. (C and D) Flow cytometry plots of labeling experiment described in B using Jurkat cells. (C) Cells were labeled for 1 h with 100 μM biotinylated derivatives of a phosphine (Staudinger ligation) (20), a nonfluorinated cyclooctyne (strain-promoted cycloaddition) (14), and DIFO (Cu-free click chemistry). In all cases, control cells (incubated in azido sugar-free medium but carried through an identical labeling procedure) displayed mean fluorescence intensity (MFI, arbitrary units) values < 20. (D) Cells were labeled for 1 h with 10 nM–100 μM DIFO-biotin. Error bars represent the standard deviation of three replicate experiments. Solid line, + Ac₄ManNAz; dashed line, − Ac₄ManNAz.

Fig. 3.

Time-lapse imaging of glycan trafficking using an Alexa Fluor 488 derivative of DIFO. (A–H) CHO cells were incubated with 100 μM Ac₄ManNAz (A–D) or 100 μM Ac₄ManNAc as a negative control (E–H) for 3 days and subsequently labeled with 100 μM DIFO-488 at 37°C for 1 min. (I–N) Time-lapse imaging of a single cell from the previous experiment over 1 h at 25°C (I–M, Ac₄ManNAz; N, Ac₄ManNAc).

Fig. 4.

Multicolor, dynamic imaging of glycan trafficking using Alexa Fluor derivatives of DIFO. CHO cells were grown for 2 days in 100 μM Ac₄ManNAz (A–E) or 100 μM Ac₄ManNAc (data not shown) and labeled for 1 h at 37°C with 10 μM DIFO-488 (t = 0 h, A and B). The cells were returned to medium supplemented with the appropriate sugar for 23 h and then labeled for 1 h at 37°C with 10 μM DIFO-568 (t = 24 h, C–E). Labeling in the Golgi apparatus and endosomes (filled arrowheads) and lysosomes (open arrowheads) was confirmed in colocalization experiments with known markers (SI Fig. 9).