A Reliable Approach for Revealing Molecular Targets in Secondary Ion Mass Spectrometry

Abstract

Nano secondary ion mass spectrometry (nanoSIMS) imaging is a rapidly growing field in biological sciences, which enables investigators to describe the chemical composition of cells and tissues with high resolution. One of the major challenges of nanoSIMS is to identify specific molecules or organelles, as these are not immediately recognizable in nanoSIMS and need to be revealed by SIMS-compatible probes. Few laboratories have generated such probes, and none are commercially available. To address this, we performed a systematic study of probes initially developed for electron microscopy. Relying on nanoscale SIMS, we found that antibodies coupled to 6 nm gold particles are surprisingly efficient in terms of labeling specificity while offering a reliable detection threshold. These tools enabled accurate visualization and sample analysis and were easily employed in correlating SIMS with other imaging approaches, such as fluorescence microscopy. We conclude that antibodies conjugated to moderately sized gold particles are promising tools for SIMS imaging.

Keywords: immunogold labeling; mass spectrometry imaging; nanoSIMS; protein turnover.

Figure 1

Schema of Au-based probes measured by nanoSIMS. Secondary antibodies were conjugated to Au_Np (ball) with a size of 6 nm, 15 nm and 1.4 nm, and the 1.4 nm-conjugated antibody was additionally conjugated to a fluorophore (star). The nanobody (Nb) was conjugated to 3 nm Au_Np. All of them enable the affinity labeling of target molecules. The 5 nm Au_Np can be attached to His-tagged proteins through the binding of its conjugated nickel-NTA to His-tags.

Figure 2

NanoSIMS detection of 6 nm Au_Np-tagged HEK293 organelle markers. Schematic representation of the expected organelle distribution from thin tissue slices after embedding and sectioning. The 197Au signal from antibody labeling represents the specific targeted protein, while the 12C14N and 32S signals reveal the cells. The signals are produced by a broad range of biological materials (proteins, DNAs, RNAs, etc.) that contain a substantial amount of C, N and S elements. One representative nanoSIMS image from each organelle labeling is shown.

Figure 3

Boxplots show the 6 nm Au_Np labeling intensity of HEK293 organelle markers in comparison to the control in which no primary antibodies were applied. A significantly higher cellular labeling intensity was detected for both (a) bovine serum albumin (BSA) blocking (Wilcoxon rank sum test, N = 6 and 10 for control and calnexin, respectively; p = 2.5 × 10⁻⁴, 0.073 and 2.5 × 10⁻⁴ for 197Au, 12C14N and 197Au/12C14N, respectively) and (b) tryptone blocking (Wilcoxon rank sum test, N = 4 and 5 for control and GM130, respectively; p = 0.016, 0.29 and 0.016 for 197Au, 12C14N and 197Au/12C14N, respectively). * p ≤ 0.05, *** p ≤ 0.001.

Figure 4

Correlation of 6 nm Au_Np detected by nanoSIMS and fluorescence signals detected by fluorescent microscopy. (a) Overlay of gold images and fluorescent images for targeting PMP70. Examples of colocalized gold and fluorescent signals are indicated by white arrows. The signal variations of gold and Cy3 along the area indicated by a white horizontal line were extracted and plotted in (b). (c) shows the correlation coefficient between gold and fluorescent images (real) in comparison to that between gold and mirrored fluorescence images (mirrored). Wilcoxon rank sum test: for GM130, lamp1 and PMP70, N = 51, 45 and 50, respectively, and p = 4.9 × 10⁻⁵, 1.3 × 10⁻⁴ and 1.9 × 10⁻⁹, respectively. *** p ≤ 0.001, **** p ≤ 0.0001.

Figure 5

Detection of protein turnover heterogeneity in different organelles of ¹³C isotopically enriched mouse brain slices. For each organelle marker staining, ¹³C/¹²C ratios were calculated for 6 nm Au_Np-labeled areas (with gold) and the random selection of multiple representative unlabeled areas (no gold). A difference in the relative protein “age” in different organelles, in comparison to the average cellular protein “age”, is demonstrated by this approach. Wilcoxon rank sum test: for lamp1, calnexin, GM130 and PMP70 (no gold, with gold), N = (113, 41), (59, 80), (56, 40) and (47, 34), respectively, and p = 0.019, 0.044, 0.28 and 0.004, respectively. * p ≤ 0.05, ** p ≤ 0.01.