Azide-Alkyne Click Conjugation on Quantum Dots by Selective Copper Coordination

Abstract

Functionalization of nanocrystals is essential for their practical application, but synthesis on nanocrystal surfaces is limited by incompatibilities with certain key reagents. The copper-catalyzed azide-alkyne cycloaddition is among the most useful methods for ligating molecules to surfaces, but has been largely useless for semiconductor quantum dots (QDs) because Cu⁺ ions quickly and irreversibly quench QD fluorescence. To discover nonquenching synthetic conditions for Cu-catalyzed click reactions on QD surfaces, we developed a combinatorial fluorescence assay to screen >2000 reaction conditions to maximize cycloaddition efficiency while minimizing QD quenching. We identify conditions for complete coupling without significant quenching, which are compatible with common QD polymer surfaces and various azide/alkyne pairs. Based on insight from the combinatorial screen and mechanistic studies of Cu coordination and quenching, we find that superstoichiometric concentrations of Cu can promote full coupling if accompanied by ligands that selectively compete with the Cu from the QD surface but allow it to remain catalytically active. Applied to the conjugation of a K⁺ channel-specific peptidyl toxin to CdSe/ZnS QDs, we synthesize unquenched QD conjugates and image their specific and voltage-dependent affinity for K⁺ channels in live cells.

Keywords: CuAAC; bioconjugation; combinatorial nanoscience; copper; high-throughput screen; quantum dot; quenching; synthesis.

Figure 1

Copper-mediated click reactions on QD surfaces. (a) CuAAC reaction of alkynes with amphiphilic polymer-coated core/shell QDs, listing reaction variables tested in this study. (b) Structures of surface polymers and azide linkers. (Left) Polyacrylic acid-based amphiphilic random co-polymer,^– with azide PEG linker modification. (Right) Co-maleic anhydride-octadecene polymer^, modified with azide PEG linkers and inert PEG amines. Stoichiometries are estimates based on polymer molecular weights, and positions of monomers are random. (c) Kinetic emission of 10 nM QDs with and without exposure to Cu ions. (d) QD emission spectra before (blue) and after (green, shown magnified 100-fold) addition of 20 μM Cu⁺.

Figure 2

FRET-based screen for improved CuAAC reaction with minimized QD quenching. (a) Absorbance (dashed) and emission (solid) spectra for CdSe/CdS QD donor (blue) and Cy5 acceptor (red). (b–e) Optimization of FRET emission spectra following 405-nm excitation, for reactions varying: (b) pH: 3.5 (blue), 5.0 (red). 7.5 (green), 9.0 (purple). (c) time (min): 1 (blue), 5 (red). 10 (green), 20 (purple). (d) ascorbate concentration (Cu eq.): 0 (blue), 2 (green). 20 (red), 100 (purple). (e) ligand (all 10 Cu eq.): L-cysteine (blue), THPTA (red), BTTAA (green), L-methionine (purple). See Supplemental Information Methods for full reaction details. Best CuAAC conditions are highlighted.

Figure 3

Combinatorial fluorescence analysis of ~1200 CuAAC reaction conditions for QDs. (a) Integrated Cy5 acceptor emission (670 – 700 nm) versus integrated QD emission (600 – 630 nm) for all reaction conditions. Circled data points are fully Cy5-conjugated control (black star) and closest CuAAC conditions (teal circle). (b) Emission spectra of unmodified QDs (red), control Cy5-conjugated QDs (purple), and optimized Cy5 CuAAC-conjugated QDs (teal). (c) Emission spectra of identical Cy5-QD CuAAC reactions with (teal) and without (black, shown magnified 10-fold) THPTA ligands and phosphate-citrate buffer, pH 4.5. (d–e) Pairwise correlations of reaction variables for QD emission (d) and Cy5 emission (e).

Figure 4

CuAAC chemistry and Cu quenching mechanisms on amphiphilic polymer-coated QDs. Copper ions may quench QDs by diffusion through the hydrophobic layer to the inorganic crystal or by acting as proximal charge traps when coordinated to surface carboxylates. Peptidyl toxin is shown in dark blue; THPTA ligand in light blue; CdSe core as yellow; and ZnS or CdS shell as gray. Exact structures of Cu-ligand, Cu-QD, and M²⁺-polymer are unknown.

Figure 5

Live cell imaging of QD-GxTX conjugates on CHO cells expressing Kv2.1 channels. Representative confocal images of QD-GxTX in cells with (a) high and (b) low Kv2.1 channel expression. Yellow is QD emission and blue is cellular autofluorescence. Cells with (c) high and (d) low Kv2.1 expression show similar levels of binding for QDs without conjugated GxTX. Kv2.1-expressing cells stained with QD-GxTX (e) at resting membrane potential and (f) after K⁺-induced membrane depolarization. Scale bar is 20 μm. (g) Statistical analysis of emission intensities in (a) – (f) using 560 – 610 nm (QD) and 450 – 500 nm (autofluorescence) integrated emissions. Error bars are standard error. P values: **** ≡ p < 0.0001, * ≡ p < 0.2 (not significant). N = 100 cells.