Next-generation hybridization chain reaction tools with enhanced sensitivities to detect challenging targets

Abstract

Compared to traditional enzyme-based in situ amplification methods, Hybridization Chain Reaction v3.0 (HCR v3.0) offers high specificity for spatial RNA visualization but lacks the sensitivity needed for short or low-abundance targets, especially in thick tissue with high autofluorescence. We describe next-generation HCR detection methods that combine the specificity of HCR v3.0 with enzyme-based signal amplification through catalysis (HCR-Cat) or immunostaining (HCR-Immuno, HCR-Multi). These methods enhance sensitivity for robust spatial detection of both short and low-abundance targets, work well in challenging tissue environments, and enable broad utility across basic research and translational applications. These methods allow spatial detection of challenging targets that are poorly-accessible using HCR v3.0, as well as quantitative analysis of single transcripts even when targeting short RNAs with a limited number of probes.

Spatial biology has revolutionized our understanding of cellular processes by providing crucial insights into the tissue-specific organization and regulation of gene expression, particularly through RNA expression profiling, which allows for the precise mapping of molecular signatures within their native cellular environment. HCR v3.0¹ enables multiplexed, quantitative, high-resolution imaging of mRNA targets, with low background due to the use of split-initiator probes. The amplified signal scales approximately linearly with the number of probes used for each RNA target and the number of RNA target molecules, thereby limiting signal intensity for targets that are short or expressed at low levels. Detection of targets in these cases is particularly challenging when working with thick samples that have high levels of autofluorescence. Here, we describe next-generation versions of HCR v3.0 that address these limitations by increasing detection sensitivity while retaining the strengths of HCR v3.0.

To do so, we conducted HCR followed by catalytic reporter deposition (HCR-Cat) using HCR amplifiers labeled with fluorescein isothiocyanate (FITC) instead of conventional fluorophores such as Alexa Fluor dyes (Fig. 1a). These amplifiers enable binding of FITC-specific antibodies that are conjugated to enzymes such as horseradish peroxidase (HRP) or alkaline phosphatase (AP), followed by catalytic deposition of fluorescent or chromogenic reporters (Fig. 1b). This approach maintains the advantages of HCR v3.0 but has the potential to significantly increase signal. To demonstrate the enhanced detection sensitivity of HCR-Cat relative to HCR v3.0, we performed these assays on whole-mount larval zebrafish at 5 days post-fertilization (dpf) using a spatially restricted neuropeptide, hypocretin (hcrt)², as an mRNA target. We found that HCR-Cat using an HRP-conjugated antibody produced a dramatically higher signal (240-fold increase averaged over all laser powers tested) compared to HCR v3.0 while maintaining signal specificity and low background (Fig. 1c–f, Extended Data Fig. 1a–b”). To test a more challenging scenario where a target RNA is too short to accommodate multiple probes – a common limitation of HCR v3.0 – we tested whether HCR-Cat could detect hcrt mRNA using just a single probe pair (for simplicity, we will refer to a probe pair as a probe), compared to the full set of 8 probes. Remarkably, HCR-Cat produced robust RNA detection even with a single probe and as little as 10% laser power, whereas HCR v3.0 failed to produce any detectable signal even at 100% laser power (Fig. 1g–j, Extended Data Fig. 1c–d”). We obtained a similar signal enhancement for several other zebrafish mRNA targets (Extended Data Fig. 2, 3). Doubling the digital gain while imaging also did not increase the signal for HCR v3.0 to the levels seen with HCR-Cat at fixed laser powers (Extended Data Fig. 2c–c”,2e–e”). We found that amplifiers conjugated to digoxigenin (DIG), and the use of an HRP-conjugated anti-DIG antibody, significantly enhanced signal for another hypothalamic marker, neuropeptide VF (npvf)³, compared to HCR v3.0, similar to FITC-conjugated amplifiers (Extended Data Fig. 4). The use of amplifiers conjugated to FITC or DIG that targeted hcrt or npvf, respectively, enabled detection of both mRNA targets in the same sample (Fig. 1k–m, Extended Data Fig. 1e–f”).

Fig. 1: HCR-Cat enhances mRNA detection sensitivity compared to HCR v3.0 in whole-mount zebrafish larvae.

a–b, Probes (blue) hybridize to their target RNA (gray), exposing initiators (cyan) that trigger the HCR. Compared to Alexa Fluor-conjugated (green stars) HCR v3.0 amplifiers, HCR-Cat amplifiers are conjugated to haptens (green circles) that are detected by an antibody conjugated to HRP (blue Y), which can catalytically deposit TSA reporter molecules (open green squares converted to closed green squares). c, Detection of hcrt with 8 probes using HCR-Cat with 50% laser power. d–e”, Detection of hcrt with 8 probes using HCR v3.0 (d-d”) and HCR-Cat (e-e”) at different laser powers. f, Quantification of signal intensity using 8 hcrt probes and either HCR v3.0 (blue, n=4 fish) or HCR-Cat (red, n=4 fish). HCR-Cat showed significantly higher signal at all laser powers (p<0.05 at 1% laser power, p<0.0001 for other laser powers; there was no significant difference in signal between HCR-Cat at 50% and higher laser powers, implying that the signal saturated at 50% laser intensity; Two-Way ANOVA with Tukey’s correction). g, Detection of hcrt with 1 probe using HCR-Cat with 100% laser power. h–i”, Detection of hcrt with 1 probe using HCR v3.0 (h–h”) and HCR-Cat (i–i”) at different laser powers. j, Quantification of signal intensity using 1 hcrt probe and either HCR v3.0 (blue, n=4 fish) or HCR-Cat (red, n=4 fish) showed significantly higher signal for HCR-Cat (p<0.001 at 50% laser power, p<0.0001 at 75% and 100% laser power; Two-Way ANOVA with Tukey’s correction). k, Co-detection of hcrt and npvf with 8 and 10 probes, respectively, using HCR-Cat with 50% laser power for both channels. l–m”, Co-detection of hcrt and npvf with 8 and 10 probes, respectively, showed higher signal for both targets using HCR-Cat (m-m”) compared to HCR v3.0 (l-l”). For each sample, the boxed region was imaged by gradually increasing the laser power from 0% to 100% while maintaining a fixed gain. Note: A higher digital gain was used for (g-j) compared to (c-f), and a higher digital gain was used to detect hcrt in (c-f) compared to (k-m). All experiments used at least 4 fish. High magnification images shown in (d, e, h, i, l, m) correspond to the boxed regions in (c, g, k) and Extended Data Fig. 1 (a”, b”, c”, d”, e”, f”). Representative images shown are maximum intensity projections of z-stacks across the entire cell populations. Alexa Fluor 488 (d, h, l), Alexa Fluor 546 (l), TSA-Flu (e, i, m), and TSA-Cy3 (m) were used. Scale bars, 50 μm (c, g, k) and 10 μm (e”, i”, m”). Mean ± SEM are shown (f, j).

We next explored whether HCR-Cat similarly improves the detection of transcripts in mouse brain tissue. To do so, we selected marker genes that show region- and cell-type–restricted expression — Dopamine Receptor D1 (Drd1), Dopamine Receptor D2 (Drd2), Tachykinin Receptor 1 (Tacr1), Tachykinin Receptor 3 (Tacr3), and Tyrosine Hydroxylase (Th) — and observed robust signal using HCR-Cat at a laser power at which we did not detect these transcripts using HCR v3.0 (Fig. 2a). The restricted expression patterns of these genes matched those shown in the Allen Brain Atlas chromogenic ISH database⁴. We observed a similar signal enhancement of the pan-neuronal markers Synaptosomal-Associated Protein 25 (Snap25) and Enolase 2 (Eno2), and the glial marker Solute Carrier Family 1 Member 3 (Slc1a3), in mouse cortex (Extended Data Fig. 5a).

Fig. 2: HCR-Cat enhances mRNA detection in mouse brain and whole-mount Drosophila larvae.

a, Comparison of HCR-Cat with HCR v3.0 for detection of neuron subtype markers in the mouse striatum and midbrain. Consistent with known expression patterns, HCR-Cat detected Drd1 (15 probes)- and Drd2 (16 probes)-positive neurons throughout the striatum whereas Tacr1 (13 probes) and Tacr3 (13 probes) labeled sparser sets of cells. Cell-specific expression of Th (13 probes) was detected in the midbrain ventral tegmental area and substantia nigra. Each staining was performed on 100 μm thick sections from 3 separate mice. Representative images shown are maximum intensity projections of z-stacks. Boxed regions are shown at higher magnification in each inset. b, Experimental paradigm for detection of mRNA throughout thick sections of Thy1-YFP mouse brain: PACT clearing, mRNA labeling with HCR v3.0 or HCR-Cat, followed by refractive index matching and imaging. c, Representative YFP+ pyramidal cell at 500 μm depth, showing strong and specific YFP (7 probes) mRNA signal using HCR-Cat. d, Confocal stacks through 600 μm of PACT-cleared Thy1-YFP mouse cortex, with YFP transcript detected using HCR-Cat but not HCR v3.0. HCR v3.0 samples were imaged with 100% laser power, whereas HCR-Cat samples were imaged with 10% laser power. No correction for depth was applied. Representative images from 4 separate mice are shown. e, mRNA signal intensity was quantified in cell bodies, segmented using the YFP channel. Data was normalized to HCR v3.0 mRNA signal intensity at tissue surface. HCR-Cat yielded ~25-fold increase in YFP mRNA signal vs. HCR v3.0 when imaged at the same laser power. Decreasing laser power to 10% still yielded a 10-fold increase in YFP mRNA signal for HCR-Cat vs. HCR v3.0. Mean ± 95% confidence interval from 4 mice is shown. f–g, Comparison of HCR-Cat and HCR v3.0 for detecting lar in larval Drosophila VNC. Using 20 probes, HCR-Cat produced a robust signal at 10% laser power, whereas HCR v3.0 failed to yield a clear signal even at 100% laser power and double the digital gain. h–i, Comparison of HCR-Cat and HCR v3.0 for detecting eve in larval Drosophila VNC. Using 19 probes, HCR-Cat produced a robust signal at 10% laser power, whereas HCR v3.0 failed to yield any detectable signal even at 100% laser power. j–k, HCR-Cat dramatically increased the signal for dArc1 in larval Drosophila VNC compared to HCR v3.0 at the same laser power and gain. l–m, HCR-Cat produced a robust signal for dArc1 using 19 probes in motor nerves and in projections from neurons in the larval Drosophila brain compared to HCR v3.0. Each experiment used at least 7 Drosophila larvae. Representative images shown are maximum intensity projections of z-stacks across the entire cell populations. Alexa Fluor 488 (a, top panel, f-f’”, h-h’, j, l), Alexa Fluor 647 (d, left two panels), TSA-Flu (g, i, k, m-m’), TSA-Alexa Fluor 488 (a, bottom panel) and TSA-Alexa Fluor 647 (d, right two panels) were used. Scale bars, 100 μm (a), 50 μm (c), 160 μm (d, orange arrows in plane of image) and 10 μm (f–m).

Next, we tested whether HCR-Cat can be applied to cleared tissue volumes to detect transcripts at depth. For this, we used Thy1-YFP mice⁵, as the sparse expression of YFP in cortical pyramidal cells provides a convenient signal to segment cell bodies for analysis, and probed for YFP transcript using HCR v3.0 or HCR-Cat. We PACT-cleared^6,7 600 μm sections from Thy1-YFP mouse brains, then performed HCR, followed by refractive index matching and imaging (Fig. 2b). As with our results on 100 μm sections (Fig. 2a), HCR-Cat significantly increased YFP mRNA signal, which was specific to cells that expressed YFP protein (Fig. 2c), whereas HCR v3.0 produced a very dim signal, even when imaged with 10-fold higher laser power compared to HCR-Cat (Fig. 2d). Quantification of RNA signal intensity revealed that HCR-Cat provides strong signal (~25-fold increase compared to HCR v3.0) (Fig. 2e), which can be detected deep into tissue volumes, even with no correction for imaging depth. Similar results were obtained for the endogenously expressed gene Snap25 (~12-fold increase compared to HCR v3.0) (Extended Data Fig. 5b,c). As another test of thick opaque tissue, we probed whole adult zebrafish brain for a spatially restricted target, dopamine β-hydroxylase (dbh), followed by tissue clearing using iDISCO+⁸. HCR-Cat generated bright and specific labeling of dbh-expressing neurons, compared to no detectable signal using HCR v3.0 (Extended Data Fig. 2f–g).

HCR-Cat also dramatically improved signal for two low-abundance transcripts, leukocyte antigen related-like (lar) and even-skipped (eve), in the larval Drosophila ventral nerve cord (VNC) (Fig. 2f–i). HCR v3.0 was unable to detect eve signal even at 100% laser power (Fig. 2h’) and only produced weak punctate signal for lar. Moreover, using HCR-Cat, we obtained direct evidence that dArc1 is expressed in VNC motor neurons, corroborating previous genetic studies⁹ that demonstrated its presence at neuromuscular junctions and indirectly inferred its source to be motor neurons in the VNC through RNAi analyses (Fig. 2j–m). HCR-Cat revealed dArc1 mRNA expression not only in motor neuron cell bodies in the VNC (Fig. 2k), but also clearly demonstrated dArc1 mRNA “travelling” via motor nerves (Fig. 2m) for post-synaptic transfer at neuromuscular junctions. dArc1 was also expressed in a few neurons in the brain (Fig. 2m’) that showed axonal transport of dArc1 mRNA, providing the first direct evidence of dArc1 mRNA moving along axonal shafts.

While HCR-Cat yields significantly enhanced signal, the staining is qualitative, and subcellular resolution is compromised due to diffusion of the reporter molecules. Although this can be partially mitigated by shortening the amplification time or optimizing enzymatic conditions, we developed a more efficient alternative: HCR followed by immunostaining of hairpin-conjugated haptens (HCR-Immuno). In this method, haptens conjugated to HCR hairpins are detected with primary antibodies, which in turn are detected by Alexa Fluor-conjugated secondary antibodies (Fig. 3b). This approach produced an increase in signal for npvf mRNA in whole-mount larval zebrafish (~4.5-fold increase averaged over all laser powers tested) (Fig. 3c–e), while preserving spatial resolution since antibodies remain tethered to the HCR amplifiers at the target site. HCR-Immuno offers the additional advantage of allowing detection of several distinct targets simultaneously, depending upon the availability of amplifier systems, haptens, and antibodies.

Fig. 3: HCR-Immuno and HCR-Multi enhance mRNA detection sensitivity in whole-mount zebrafish larvae compared to HCR v3.0.

a–e, Probes (blue) hybridize to their target RNA (gray), exposing initiators (cyan) that triggers the HCR. In contrast to Alexa fluorophore-labeled (green stars) HCR v3.0 amplifiers (a), HCR-Immuno amplifiers (b) are conjugated to haptens (green circles), that are detected by a primary antibody (blue Y), followed by detection using a secondary antibody (red Y) conjugated with Alexa fluorophores (green stars). HCR-Immuno produced a robust signal for npvf (d-d”) compared to HCR v3.0 (c-c”) using 10 probes. e, Quantification of signal intensity using HCR v3.0 (blue, n=4 fish) or HCR-Immuno (red, n=4 fish) showed significantly higher signal for HCR-Immuno. f–i, For HCR-Multi, amplifiers are conjugated to haptens (green circles) that are detected by a primary antibody (blue Y), followed by detection using a secondary antibody (red Y) conjugated to a second initiator (black) in a different amplifier system. This second initiator is conjugated to a different hapten (orange squares), which is detected with a second primary antibody (purple Y), followed by detection using a secondary antibody (light blue Y) conjugated to Alexa fluorophores (green stars) (f). HCR-Multi produced a robust signal for hcrt (h-h”) compared to HCR v3.0 (g-g”) using 8 probes. Quantification of signal intensity (i) using HCR v3.0 (blue, n=4 fish) or HCR-Immuno (red, n=4 fish) showed significantly higher signal for HCR-Multi. For (c, d, g, h) the same region was imaged at 10%, 50%, and 100% laser power while maintaining a fixed gain. j-k, HCR-Immuno enables visualization of individual npvf transcripts with just 10 probes (k), whereas HCR v3.0 fails to produce detectable signals under the same imaging conditions (j). All experiments used at least 4 fish. Representative images shown are maximum intensity projections of z-stacks across the cell populations (c-c”, d-d”, g-g”, h-h”) or a single optical section (j-k). Alexa Fluor 488 (c, d), Alexa Fluor 546 (g), Alexa Fluor 568 (h), and Alexa Fluor 647 (j-k). were used. ** p<0.01, **** p<0.0001 by Two-Way ANOVA with Tukey’s correction. Scale bars, 10 μm (d”, h”, k). Mean ± SEM are shown (e, i).

To further increase detection sensitivity while maintaining spatial resolution, we performed HCR in multiple rounds (HCR-Multi) for the same target. To do this, after HCR-Immuno, we used a secondary antibody labeled with an initiator instead of an Alexa Fluor-conjugated secondary antibody. This enables another round of HCR with hairpins conjugated to a different hapten, followed by a primary antibody specific for that hapten, and then an Alexa Fluor-conjugated secondary antibody (Fig. 3f). We found that HCR-Multi led to a dramatic increase in hcrt signal intensity (~70-fold increase averaged over all laser powers tested) compared to HCR v3.0 (Fig. 3g–i). While we performed HCR-Multi for only two rounds of HCR for a single target, this can be extended to detect multiple targets simultaneously, as well as perform more rounds of amplification, depending upon the availability of amplifier systems, haptens, and antibodies, but might require optimization to minimize any background caused by multiple rounds of amplification.

Critically, in addition to robust detection, both HCR-Immuno and HCR-Multi can be used to quantify transcripts even when the number of probes is limited, which is a major limitation of HCR v3.0, requiring >40 probes and reduced amplification time (~60 min) to avoid signal spread¹. This makes both methods particularly valuable for quantitative studies of challenging targets, where either the transcript length or high sequence homology with other transcripts limits the number of unique probes. To demonstrate this feature, we performed HCR-Immuno targeting npvf mRNA with 10 probes. Despite using a smaller probe set than the recommended 40 probes, HCR-Immuno produced discrete, punctate signals in larval zebrafish npvf-expressing neurons that was clearly detectable at the same laser power and imaging settings where HCR v3.0 failed to yield any visible signal (Fig. 3j–k). The uniform intensity and spatial distribution of the fluorescent puncta are consistent with single-transcript resolution. Moreover, while we did not perform single-transcript resolution analysis for HCR-Multi, its strong signal intensity and spatial confinement suggest that it may also be compatible with single-transcript detection.

Our results introduce a next-generation HCR framework that enhances the sensitivity of spatial transcript detection without sacrificing specificity or resolution. HCR-Cat enables detection of challenging transcripts, including short targets, even in thick, highly autofluorescent tissues where HCR v3.0 performs poorly, and can yield signal with as few as one probe. While catalytic deposition produces strong amplification, it compromises the quantitative precision of HCR v3.0. To overcome this, we developed HCR-Immuno, which preserves subcellular localization by anchoring antibodies to HCR amplifiers and supports quantitative analysis and multiplexing. HCR-Multi further increases sensitivity through iterative signal amplification while maintaining spatial resolution. Collectively, these approaches overcome key limitations of HCR v3.0 and provide a flexible toolkit for high-resolution, high-sensitivity spatial transcriptomics across diverse tissues. These approaches are versatile, compatible with standard reagents and HCR v3.0 workflows, and readily scalable for multiplex detection.

Several strategies to improve HCR v3.0 have previously been described. USeqFISH¹⁰ combines rolling circle amplification with HCR and sequential labeling to achieve bright signals with as few as four probes and enables multiplex detection of ~50 genes, but requires enzymes, tissue clearing, and a complex workflow that limits 3D applications. Modified hairpin designs, including shortened hairpins¹¹, branched hairpins¹², and additional purification steps¹³, have also been reported to improve signal intensity or polymerization efficiency. However, these approaches often require specialized synthesis, custom reagents, or remain limited in thick tissues. Although we did not test HCR-Immuno in thick mouse brain, penetration limitations of full-length antibodies could be mitigated by using smaller binders, such as initiator-conjugated F(ab′)₂/Fab fragments¹⁴ or nanobodies¹⁵, together with extended incubations or clearing protocols, extending its applicability to large tissue volumes. In contrast to USeqFISH and hairpin-modification approaches, HCR-Immuno and HCR-Multi retain the enzyme-free simplicity of HCR, while HCR-Cat uses HRP-conjugated antibody fragments. Importantly, these small conjugates are far less demanding than the large enzymes and tissue-clearing requirements of RCA-based methods, offering a more practical route to increased sensitivity. Together, these advances highlight complementary strategies, with our methods offering a balance of sensitivity, quantification, and depth compatibility, while remaining fully compatible with existing HCR v3.0 probes to facilitate seamless adoption.

Online Methods

Zebrafish

Zebrafish husbandry

All zebrafish experiments were performed in accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines and by the Office of Laboratory Animal Resources at the California Institute of Technology (animal protocol 1836). Unless indicated otherwise, experiments used 5–7 days post fertilization (dpf) larval zebrafish from the TLAB strain, which were generated by mating AB and TL strains obtained from the Zebrafish International Resource Center (ZIRC, Oregon). Embryonic and larval zebrafish were raised at 28.5°C in E3 embryo medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgSO₄, pH 7.4) in petri dishes.

HCR v3.0 staining for larval zebrafish

HCR reagents, including probes, hairpins (based on systems B1-B5¹), and buffers, were purchased from Molecular Technologies (MT). Probes can also be designed using custom scripts^16,17 and buffers can be prepared using published recipes^1,18. For simplicity, the term ‘probe’ is used throughout this manuscript to refer to a probe set (or probe pair) designed against a given transcript. Staining was performed according to a previously published protocol¹⁹ with some modifications. Staining was performed with 10–12 larval zebrafish in a 1.5 mL microcentrifuge tube, and a volume of ~500 μL was used for each incubation and wash step. Unless otherwise stated, all incubations and washes were performed with gentle shaking. Larval zebrafish at 5–7 dpf were anesthetized in 1.5 mM tricaine (E10521, Sigma-Aldrich) and fixed with ice-cold 4% paraformaldehyde (PFA) in 1× Dulbecco’s phosphate-buffered saline (without divalent cations) with 0.1% Tween 20 (DPBST) overnight at 4°C. The following day, samples were washed three times for 5 min each with DPBST, followed by incubation in a bleaching solution (1.5% hydrogen peroxide, 1% potassium hydroxide, 0.1% Tween 20) for about 30 min (duration depends on the age of the larvae) at room temperature while monitoring pigment loss. Samples were washed 5 times for 5 min each with DPBST and incubated in DPBSTx (1× DPBS + 0.25% Triton X-100) at room temperature for 1 h to increase sample permeability. Samples were then incubated in ice-cold 100% methanol (MeOH) at −20°C for 10 min to dehydrate and further increase sample permeability, followed by rehydration by washes with 50% MeOH/50% DPBSTx and 25% MeOH/75% DPBSTx for 5 min each, and then 5 times for 5 min each with DPBSTx. Samples were then incubated in 10% hybridization buffer in 5× sodium chloride sodium citrate + 0.1% Tween 20 (SSCT) for at least 15 min in a 37°C heat-block before transferring to 100% hybridization buffer (prewarmed at 37°C) for 1–2 h for pre-hybridization. The pre-hybridization buffer was then replaced with pre-warmed probe solution (100% hybridization buffer with 16–20 nM probes for weakly-expressed targets and 2–4 nM probes for highly expressed targets) and incubated at 37°C overnight (we recommend 36–60 h for weakly expressed targets). To remove excess probe, samples were washed 4 times for 15 min each with pre-warmed probe wash buffer at 37°C. Samples were then washed twice for 5 min each with 5× SSCT at room temperature. Next, pre-amplification was performed by incubating samples in amplification buffer (pre-warmed to room temperature) for 1 h at room temperature without shaking. Hairpins h1 and h2 (30 pmol each) were individually prepared by snap-cooling 10 μL aliquots of 3 μM stock solutions. Snap-cooling was performed by heating the hairpins to 95°C for 90 s, followed by passive cooling to room temperature in the dark for at least 30 min. Subsequently, the hairpins were combined and diluted into 500 μL of amplification buffer to generate the working hairpin solution. Following removal of the pre-amplification buffer, samples were incubated with the hairpin solution for 12–16 h at room temperature in the dark without shaking. Excess hairpins were removed by four washes for 15 min each with 5× SSCT at room temperature (note: HCR-Cat and HCR-Immuno protocols diverge at this step). HCR v3.0 samples that served as controls for comparison with other methods were kept in 5× SSCT at 4°C in the dark until additional steps for the other methods were completed. Samples were then washed with 1× DPBS for 10 min without shaking, followed by a gradual increase in the concentration of VECTASHIELD PLUS (H-1900–10, Vector Laboratories) by washes in 10%, 25%, 50%, and 80% VECTASHIELD in 1× DPBS for at least 15 min each at room temperature without shaking. Samples were imaged in 80% VECTASHIELD.

HCR-Cat staining for larval zebrafish

HCR hairpins labeled with FITC were purchased from Molecular Technologies and HCR hairpins labeled with DIG were purchased from Molecular Instruments (Los Angeles, CA, USA). All hairpins were purchased from Molecular Technologies or Molecular Instruments, but both Alexa Fluor-labeled and hapten-labeled hairpins can be prepared using published protocols for HCR amplifier design and synthesis¹. For comparison, we used directly fluorophore-labeled HCR hairpins, whereas in HCR-Cat the signal detection relies on detecting the haptens. Although haptens can exhibit intrinsic fluorescence, our aim was to benchmark HCR-Cat against widely used, brighter fluorophores to illustrate relative performance. For HCR-Cat, the HCR v3.0 protocol was followed until the step for washing off excess hairpins but instead of washing the samples with 1× DPBS for 10 min, they were washed three times for 5 min each with DPBSTx at room temperature. Samples were then incubated in blocking solution (1158576200z, Roche) for at least 1 h at room temperature. Blocking solution was replaced with antibody solution containing anti-Fluorescein-POD (anti-Flu-POD) Fab fragments (11426346910, Roche) at a dilution of 1:400 in blocking solution and samples were incubated at 4°C for at least overnight. Excess antibody was then removed by washing the samples 5 times for 15 min each with DPBSTx at room temperature. The signal was developed by replacing DPBSTx from the last wash with ~100 μL (or enough to cover the samples) of freshly made tyramide solution (TSA PLUS Fluorescein (NEL741001KT, Akoya Biosciences) suspended in TSA 1× Plus Amplification Diluent (FP1498, Akoya Biosciences). We used a dilution of 1:200 for weakly-expressed targets and 1:300 for highly-expressed targets. The signal was developed for 30 min in the dark at room temperature without shaking. TSA Fluorescein (TSA-Flu) was used with hairpins that were labeled with FITC to match its endogenous fluorescence. Tyramide solution was removed by a quick wash followed by 5 washes for 5 min each with DPBSTx and a single wash with 1× DPBS for 10 min, followed by a gradual increase in the concentration of VECTASHIELD PLUS as described for the HCR v3.0 protocol. Optional: Samples can be left in DPBSTx overnight at 4°C to reduce background before moving to 1× DPBS and VECTASHIELD PLUS. For single HCR-Cat with DIG-labeled hairpins, the same workflow was used except that Alexa Fluor 647-labeled hairpins were applied for HCR v3.0, while anti-digoxigenin–POD (Fab fragments; 11207733910, Roche) and tyramide solution (TSA PLUS Cyanine 5; NEL745001KT, Akoya Biosciences) were used for HCR-Cat. For double HCR-Cat, after the first round of signal development, samples were washed 5 times for 5 min each with DPBSTx (as described above), followed by incubation in 0.02 N HCl²⁰ at room temperature without shaking for 30 min to inactivate the peroxidase from the first target. The peroxidase inhibitor was removed by washing the samples 5 times for 5 min each with DPBSTx. Samples were then incubated in blocking solution (11585762001, Roche) for 1 h at room temperature. Blocking solution was replaced with antibody solution containing anti-Digoxigenin-POD (anti-DIG-POD) Fab fragments (11207733910, Roche) at a dilution of 1:400 in blocking solution and samples were incubated at 4°C for at least overnight. Excess antibody was then removed by washing the samples 5 times for 15 min each with DPBSTx at room temperature. The signal was developed by replacing DPBSTx from the last wash with ~100 μL (or enough to cover the samples) of freshly made tyramide solution (TSA PLUS Cyanine 3 (TSA-Cy3; NEL744001KT, Akoya Biosciences) suspended in TSA 1× Plus Amplification Diluents. We used a dilution of 1:200 for weakly-expressed targets and 1:300 for highly-expressed targets. The signal was developed for 30 min in the dark at room temperature without shaking. Tyramide solution was removed by a quick wash followed by 5 washes for 5 min each with DPBSTx and a single wash with 1× DPBS for 10 min, followed by gradual increase in the concentration of VECTASHIELD PLUS as described for the HCR v3.0 protocol. Optional: Samples can be left in DPBSTx overnight at 4°C to reduce background before moving to 1× DPBS and VECTASHIELD PLUS. Although we used HRP-conjugated anti-FITC and anti-DIG antibodies in this study, the approach is broadly compatible with primary antibodies against other haptens (e.g., FITC, DIG, biotin, and Alexa fluorophores) in combination with HRP- or AP-conjugated secondaries. Depending on substrate choice, the same strategy can be applied for chromogenic or fluorescent readouts.

HCR-Immuno staining for larval zebrafish

For HCR-Immuno, the HCR v3.0 protocol was followed until the step for washing off excess hairpins but instead of washing the samples with 1× DPBS for 10 min, they were washed three times for 5 min each with DPBSTx at room temperature. Samples were then incubated in blocking solution (2% normal goat serum and 2% DMSO in DPBSTx) for 1 h at room temperature. Blocking solution was replaced with 1:100 rabbit anti-digoxigenin primary antibody (700772, Invitrogen) in blocking solution at 4°C for at least overnight. Excess primary antibody was removed by washing the samples 5 times for 15 min each with DPBSTx at room temperature. Following the last wash, samples were incubated in 1:500 Alexa Fluor 488 conjugated goat anti-rabbit secondary antibody (A-11008, Thermo Fisher) at 4°C overnight. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature, followed by a single wash with 1× DPBS for 10 min, and a gradual increase in the concentration of VECTASHIELD PLUS as described for the HCR v3.0 protocol. While we only tested this method for a single target, it can be extended to more targets, based on the availability of initiator systems, haptens in the hairpins, primary antibodies, and secondary antibodies, without any increase in the duration of the protocol. For single-molecule detection, we used the same conditions as HCR-Immuno, except the amplification was reduced to 60 min at room temperature. HCR v3.0 hairpins labeled with Alexa Fluor 647 were used as controls, and HCR-Immuno was performed with detection using Alexa Fluor 647-conjugated goat anti-rabbit secondary antibody (A-21245, Thermo Fisher). HCR-Immuno experiments were performed using an ethylene carbonate-based buffer, which provided superior signal-to-noise performance compared to a traditional formamide-based buffer (data not shown), and is less hazardous than formamide-based buffers²¹.

HCR-Multi staining for larval zebrafish

Antibody buffer and secondary antibodies conjugated to initiators (B1-B5)^1,22 were purchased from Molecular Technologies and Molecular Instruments but can also be prepared using published protocols for antibody-initiator conjugation²². For HCR-Multi, HCR v3.0 protocol was followed until the step for washing off excess hairpins but instead of washing the samples with 1× DPBS for 10 min, they were washed three times for 5 min each with DPBSTx at room temperature. For the data shown (Figure 3), we used B1 amplifiers¹ labeled with DIG for the first round of amplification. Samples were blocked in antibody buffer for at least 1 h at room temperature and then incubated with 1:100 rabbit anti-digoxigenin primary antibody (700772, Invitrogen) in antibody buffer at 4°C for at least overnight. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature and then incubated in 1:500 donkey anti-rabbit secondary antibody (with B2 initiator conjugated) in antibody buffer at 4°C overnight. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature. Next, after pre-amplification for 1 h, amplification was performed overnight as in the HCR v3.0 protocol using hairpins labeled with a second hapten, FITC (B2¹-FITC). Samples were then washed 4 times for 15 min each with 5× SSCT at room temperature. Samples were then blocked again in the antibody buffer for at least 1 h at room temperature. Samples were then incubated in 1:500 mouse anti-FITC primary antibody (31242, Invitrogen) at 4°C for at least overnight. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature. Following the last wash, samples were incubated in 1:500 Alexa Fluor 568 conjugated goat anti-mouse (A-11011, Thermo Fisher) at 4°C overnight. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature, followed by a single wash with 1× DPBS for 10 min, and a gradual increase in the concentration of VECTASHIELD PLUS as described in the HCR v3.0 protocol. While we only tested this protocol for one target, it can be extended to additional targets based on the availability of initiator systems, haptens in the hairpins, primary antibodies, and secondary antibodies. Also, while the protocol tested here used two rounds of amplification, it can potentially be extended to additional rounds of amplification depending upon the availability of initiator systems, haptens in the hairpins, primary antibodies, and secondary antibodies.

Imaging, quantification, and data analysis for larval zebrafish samples

Samples were imaged in 80% VECTASHIELD using a Zeiss 880 confocal microscope. To mount samples, a stack of 5–6 reinforcement labels was placed on a glass slide, and the sample (each larva) was transferred to the well inside the labels in 80% VECTASHIELD, followed by covering the sample with a coverslip. The viscosity of the mounting media helped to maintain the dorsal-up orientation of the samples. For low-magnification images we used a 20× 1.0 NA water dipping objective and imaged the samples using the Zeiss tiling function and the pairwise stitching function of Fiji. Higher-magnification images were acquired using either a 25× or 40× objective. In both cases, a z-stack was acquired that captured the entire cell population of interest. Imaging parameters for each experiment were set using HCR-Cat/HCR-Immuno/HCR-Multi samples to avoid any over-saturation of pixels. We used different gain settings for low-magnification and high-magnification images. Only three laser powers (10%, 50%, and 100%) were used when HCR v3.0 was performed using Alexa Fluor 546 and for HCR-Immuno experiments to avoid bleaching. In other cases, a constant gain was used, and images were acquired at increasing laser powers (0%, 0.5%, 1.0%, 2%, 5%, 10%, 25%, 50%, 75%, and 100%). To quantify the signal, z-stacks corresponding to different laser powers were concatenated in increasing order of laser power using the “concatenate” plugin in Fiji²³, followed by resetting the combined z-stack to 65535. The z-stack was converted to 8-bit and the freehand tool was used to select the region of interest and measure both the integrated density and the area of the region. In controls where no cells were visible at the set digital gain during imaging, we increased the gain post-acquisition in Fiji to identify cell boundaries. Regions of interest were defined using these adjusted images, and all quantitative analyses were performed using the original acquisition settings. To measure any change in the background, a region was selected near the cells of interest. For each laser power, the integrated density was summed for the entire stack for both the cells of interest and the control region, and the final signal was calculated by subtracting the total integrated density of the control region from the total integrated density of the cells of interest. Statistical significance was assessed using two-way ANOVA followed by Tukey’s correction for multiple comparisons using Prism (Graphpad).

HCR v3.0 staining for adult zebrafish brain

HCR v3.0 staining for adult zebrafish brain was performed according to a modified version of the published protocol⁸. Adult zebrafish (2–3 years old) were anesthetized in 4% tricaine and then euthanized by immersion in ice-cold water for 5 min. Animals were then decapitated using a pair of scissors and heads were placed in ice-cold 1× DPBS for 5 min to let the blood drain. Dorsal head skin and bones were removed to expose the top of the brain, and samples were fixed in 4% PFA in DPBST overnight, after which the brains were dissected out into cold 1× DPBS and fixed in ice cold 4% PFA in DPBST overnight. Unless stated otherwise, all incubations and washes were performed using gentle shaking. Dehydration and rehydration steps were performed without shaking. Probe incubations and washes at 37°C were done using a heat block. During dissections, a small section of the spinal cord was kept intact to help in manipulating and mounting the brains for imaging. Samples were washed three time for 5 min each with DPBSTx at room temperature, followed by serial dehydration in methanol/water mixtures (30 min to 1 h each in 20%, 40%, 60%, 80%, and 100% methanol). Samples were then incubated for 1 h in ice-cold 100% methanol, chilled on ice, transferred to chilled 5% hydrogen peroxide in methanol and incubated overnight at 4°C. The next day samples were rehydrated in a methanol/water series at room temperature (30 min to 1 h each in 80%, 60%, 40%, and 20% methanol) followed by a 1× DPBS wash and two 1 h washes in DPBSTx. Samples were permeabilized by overnight incubation in permeabilization solution (0.3 M glycine and 20% DMSO in DPBSTx) at 37°C. Permeabilization buffer was removed from the samples by washing them three times for 5 min each with DPBSTx at 37°C. Two to five brains were transferred into each 2.0 mL microcentrifuge tube and a volume of at least 500–750 μL was used for each incubation and wash step. We used the same HCR v3.0 protocol for adult zebrafish brains as was used for larvae, except the probe concentration was increased to 20 nM and probe incubation was performed for 36–60 hours. HCR v3.0 samples that served as controls for comparison with HCR-Cat were kept in 5× SSCT at 4°C in the dark until the additional steps for HCR-Cat were completed. Labeled brains were dehydrated at room temperature through a graded methanol series (30 min to 1 h each in 20%, 40%, 60%, 80%, and 2×100% methanol), followed by overnight incubation in 100% methanol at room temperature. Samples were then incubated for 3 h at room temperature in 66% dichloromethane (DCM, 270997, Millipore Sigma) in methanol and then washed twice in 100% DCM for 15 min without shaking. Samples were then stored in dibenzyl ether (DBE, 33630, Millipore Sigma) at 4°C in the dark until they were imaged.

HCR-Cat staining for adult zebrafish brain samples

For HCR-Cat staining of adult zebrafish brains, the same protocol was used as for HCR v3.0 adult zebrafish brains until and including the step for washing off the excess hairpins 4 times for 15 min each with 5× SSCT at room temperature, followed by washing the samples three times for 5 min each with DPBSTx at room temperature. Samples were then left in DPBSTx for 1–2 h at room temperature, followed by incubation in Roche blocking solution for 2 h at room temperature, and then incubation in 1:400 anti-Flu-POD Fab in blocking solution at 4°C for 36–60 hours. Samples were then washed 6 times for 15 min each with DPBSTx at room temperature. The signal was developed by replacing DPBSTx from the last wash with freshly made 1:200 tyramide solution (300 μL or enough to cover the samples) in the dark at room temperature for 45 min. TSA Fluorescein was used for hairpins that were labeled with FITC to match its endogenous fluorescence. Tyramide solution was removed by a quick wash followed by washing the samples 5 times for 5 min each with DPBSTx, and the samples were left in DPBSTx overnight at 4°C to reduce background. Labeled brains were then dehydrated and cleared as described for HCR v3.0 for adult zebrafish brains.

Imaging of adult zebrafish brain samples

Adult zebrafish brains were imaged using a La Vision ultramicrosope (La Vision Biotec Ultramicroscope II Light Sheet). A drop of ultraviolet cured resin adhesive 61 (Norland Optical) was used to mount and solidify first on the imaging platform. Another drop of resin was added on top of the cured one, and the brain was added to this drop for imaging in the horizontal plane. The orientation of the brain was adjusted using forceps via the spinal cord until the resin surrounding the brain was cured, which held the brain in place during imaging. Brain samples were imaged in DBE and images were acquired at 2× magnification.

Mice

Mouse husbandry

Mouse husbandry and all procedures involving mice were performed in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Institutional Animal Care and Use Committee (IACUC) and by the Office of Laboratory Animal Resources at the California Institute of Technology. Eight-week-old, male C57BL/6J (strain #: 000664) and Thy1-YFP-H (strain #: 003782) mice were obtained from the Jackson Laboratory. Mice were housed 3–4 per cage, on a 12 h light/dark cycle, and had ad libitum access to food and water.

Mouse tissue processing

At the time of tissue collection, mice were euthanized via i.p. injection of 100 mg/kg euthasol, then transcardially perfused with 30 mL of ice-cold 4% PFA (15714-S, Electron Microscopy Sciences) in 1× DPBS. Whole intact skulls with brain inside were post-fixed for at least 48 h with 4% PFA in 1× DPBS at 4°C. Skulls were washed three times for 1 h each with 1× DPBS at room temperature, then brains were extracted and stored at 4°C until further processing. For 100 μm sections, brains were cryoprotected through immersion in 30% sucrose in 1× DPBS. Once the tissue had sunk, it was flash-frozen in O.C.T. Compound (4586, Scigen) using a dry ice-ethanol bath and kept at −70°C until sectioning. Coronal sections were obtained using a cryostat (Leica Biosystems), collected in 1× DPBS, and stored at 4°C until use. For thick brain sections, brains were sectioned coronally with a vibrating microtome (Leica Biosystems) at 600 μm. Thick sections were collected in ice-cold 1× DPBS and stored at 4°C until use.

PACT tissue clearing

Thick (600 μm) coronal sections of mouse brain were PACT-cleared using published protocols^6,7. Briefly, slices were incubated in a hydrogel solution consisting of 1× DPBS, 4% acrylamide (1610140, Bio-Rad), 1% PFA, with 0.25% VA-044 polymerization thermal initiator (VA-044, Wako), overnight at 4°C, then oxygen was removed by bubbling nitrogen gas through the solution for 5 min. The hydrogel was allowed to polymerize for 2 h at 37°C. Excess hydrogel was removed by briefly washing in 1× DPBS, then tissue was delipidated through a 16 h incubation in 1× DPBS with 8% SDS. Samples were thoroughly washed with 6 washes in vast excess (i.e. 50 mL per wash) of 1× DPBS, for 1 h per wash.

HCR v3.0 and HCR-Cat for mouse tissue

Split initiator probe sets against endogenous genes and reporter transcripts were designed according to a published protocol¹⁰, based on published amplifier sequences¹, ordered from Integrated DNA Technologies, and diluted in Ultrapure water before use. All incubations were at room temperature unless otherwise stated. For 100 μm sections, tissue was first mounted on a glass slide (2575-plus, Brain Research Laboratories), then permeabilized in DPBTx (1× DPBS with 0.1% Triton X-100) for 1 h at room temperature. PACT-cleared brain sections were stained free-floating. Subsequently, both 100 μm sections and PACT-cleared 600 μm sections were stained the same way. Samples were first incubated for 1 h at 37°C in a probe hybridization buffer consisting of 2× SSC, 10% ethylene carbonate (E26258, Sigma-Aldrich), and 10% dextran sulfate (3730, Sigma-Aldrich). Following this incubation, the samples were incubated for 12–16 h at 37°C in pre-warmed hybridization buffer with 2 nM of each probe. Samples on the slides were covered with a coverslip and incubated in a humidified chamber to prevent drying of the tissue. After probe hybridization, samples were washed twice for 30 min in stringent wash buffer (2× SSC, 30% ethylene carbonate) at 37°C and then twice for 30 min in 5× SSCT (5× sodium chloride sodium citrate + 0.1% Tween 20) at room temperature. The stringent washes were done by pipetting 1 mL of pre-warmed stringent wash buffer onto the slides and the 5× SSCT washes were done in Coplin jars using about 60 mL of buffer per wash. For 600 μm cleared sections, both the stringent and 5× SSCT washes were done in 5 mL per wash for up to 3 cleared sections per tube. Samples were then incubated in the HCR amplification buffer (2× SSC, 10% ethylene carbonate) for at least 1 h at room temperature. Alexa Fluor 488 (for HCR v3.0) and FITC-conjugated HCR hairpins (for HCR-Cat) were heated to 95°C for 90 s, then cooled to room temperature for at least 30 min in the dark. Tissue was then incubated overnight in amplification buffer with 60 nM hairpins in the dark. Samples on the slides were covered with a coverslip and incubated in a humidified chamber to prevent drying of the tissue. Following HCR amplification, the samples on the slides were washed twice for 30 min each in 5× SSCT at room temperature using excess volume in Coplin jars. Samples for HCR v3.0 were kept in 5× SSCT at 4°C, while samples for HCR-Cat were washed three times for 15 min each with DPBTx. Tissue was blocked in TNB blocking solution (FP1012, Akoya Biosciences) for 1 h, then incubated overnight in TNB blocking solution with 1:100 sheep anti-Flu POD Fab fragments at 4°C. Samples were then washed three times for 30 min each with DPBTx, and signal was developed with an AlexaFluor 647 or AlexaFluor 488 Tyramide SuperBoost kit (Thermo Fisher B40936 or B40932), according to the manufacturer’s protocol. After 10 min of signal development, the reaction was terminated and sections were washed 4 times for 30 min each with DPBTx, followed by an overnight incubation in DPBTx. Samples for both HCR v3.0 and HCR-Cat were then stained with 1:10,000 Hoechst 33342 (H3570, Thermo Fisher) in 1× DPBS for 20 min. Samples were then washed three times for 10 min each with 1x DPBS. PACT-cleared tissue underwent an additional wash in 0.02 M phosphate buffer. For 100 μm sections, a coverslip was mounted over the tissue with Prolong Diamond (P36970, Thermo Fisher). PACT-cleared sections were first incubated in 5 mL of a refractive index matching solution (RIMS), consisting of 40 g of iohexol (Histodenz; D2158, Sigma-Aldrich) dissolved in 30 mL of 0.02 M phosphate buffer (19340–72, Electron Microscopy Sciences), with pH 7.4 and refractive index of 1.46. Once sections appeared transparent, the tissue was mounted onto a glass slide in RIMS, using an iSpacer mounting chamber (IS011, SunJin Lab), and a coverslip.

Imaging and analysis of mouse samples

For imaging of stained mouse sections, a Zeiss LSM 880 was used with a 10×, 0.45 NA air objective. Imaging settings were chosen to capture the full dynamic range of the signal without saturated pixels. When possible, laser power was adjusted before adjusting detector gain. Fields of view were chosen while imaging non-experimental channels (e.g. Hoechst or YFP). For analysis of PACT-cleared mouse tissue, 600 μm confocal stacks were subdivided into thirty 23 μm blocks, and every second block was analyzed. YFP+ cell bodies were segmented using the Weka Trainable Segmentation plugin for the Fiji distribution of ImageJ²³. The resulting ROIs were used to measure cell body RNA signal intensity, using Fiji.

Drosophila

Drosophila samples

Drosophila experiments and husbandry followed standard protocols in accordance with Caltech Institutional Animal Care and Use Committee guidelines. All experiments were performed using third-instar larvae. Fly stocks were maintained at room temperature and D. melanogaster Canton S flies were used for all experiments.

HCR v3.0 staining for Drosophila samples

HCR reagents, including probes and hairpins (based on systems B1-B5¹) were purchased from Molecular Technologies. The probes can also be designed using custom published scripts^18,24. Staining was performed according to a modification of a published protocol²⁵ using home-made buffers (all in % volume - Hybridization buffer: 2× SSC, 10% ethylene carbonate, 10% dextran sulfate, 0.1% Tween 20; Stringent wash buffer: 2× SSC, 30% ethylene carbonate, 0.1% Tween 20; Wash buffer: 2× SSC, 0.1% Tween-20; Amplification buffer: 2× SSC, 10% ethylene carbonate, 0.1% Tween 20). Dissections were performed as described²⁶. Briefly, wandering third-instar larvae were dissected in 1× DPBS and the brain lobes and VNCs were removed and fixed in 4% PFA in DPBST for 30 min at room temperature, following which the samples were washed 3 times for 5 min each with DPBSTx. Ten to twelve samples were transferred into a 1.5 mL microcentrifuge tube and a volume of ~500 μL was used for all incubation steps and washes. Unless stated otherwise, all incubations and washes were performed using gentle shaking on a nutator. Incubations and washes at 37°C were done using a heat block. Samples were then pre-hybridized in ~500 μL of pre-warmed (37°C) hybridization buffer for at least 30 min at 37°C, before transferring them to pre-warmed probe solution (hybridization buffer with 16 nM probes for weakly-expressed targets and 2–4 nM for highly-expressed targets) followed by incubation at 37°C overnight (we recommend an incubation of 36–60 h for weakly-expressed targets). Samples were then washed 4 times for 15 min each with prewarmed stringent wash buffer at 37°C. Subsequently, samples were washed twice for 15 min each with 2× SSCT at room temperature. Next, pre-amplification was performed by incubating the samples in amplification buffer (pre-warmed to room temperature before use) for at least 1 h at room temperature. Hairpins h1 and h2 (30 pmol each) were individually prepared by snap-cooling 10 μL aliquots of 3 μM stock solutions. Snap-cooling was performed by heating the solutions to 95°C for 90 s, followed by passive cooling to room temperature in the dark for at least 30 min. Subsequently, the hairpins were combined and diluted into 500 μL of amplification buffer to generate the working hairpin solution. Following removal of the preamplification buffer, samples were incubated with the hairpin solution for 12–16 h at room temperature in the dark. Samples were then washed 4 times for 15 min each with 2× SSCT at room temperature. HCR v3.0 samples that served as controls for comparison with the other method were kept in 2× SSCT at 4°C in the dark until additional steps for other methods were completed. Samples were then washed once with 1× DPBS for 10 min without shaking, followed by transfer to 100% VECTASHIELD PLUS. The samples were mounted in 50–80 μL VECTASHIELD PLUS and imaged using a Zeiss LSM980 confocal microscope with either a 20× or 40× objective. Images were analyzed and processed using Fiji²³.

HCR-Cat staining for Drosophila samples

HCR hairpins labeled with haptens (FITC) were purchased from Molecular Technologies. For HCR-Cat, the HCR v3.0 protocol was followed until the step for washing off the excess hairpins, but instead of washing the samples with 1× DPBS for 10 min, they were washed three times for 5 min each with DPBSTx at room temperature. Samples were then incubated in blocking solution (1158576200z, Roche) for at least 1 h at room temperature. Blocking solution was replaced with antibody solution containing 1:400 anti-Flu-POD Fab fragments in blocking solution and samples were incubated at 4°C for at least one night. Samples were then washed 5 times for 15 min each with DPBSTx at room temperature. The signal was developed by replacing DPBSTx from the last wash with freshly-made tyramide solution (100 μL or enough to cover the samples) and the reaction was performed in the dark at room temperature. TSA Fluorescein was used to match the endogenous fluorescence of FITC. The signal was developed for 30 min using 1:200 for weakly-expressed targets and 1:300 for highly-expressed targets. Tyramide solution was removed by a quick wash followed by washing the samples 5 times for 5 min each with DPBSTx and a single wash with 1× DPBS for 10 min without shaking, followed by transfer to 100% VECTASHIELD PLUS as described in HCR v3.0 protocol. Optional: In case of too much background, samples were left in DPBSTx overnight at 4°C, followed by 1× DPBS and VECTASHIELD PLUS. The samples were mounted, imaged, and analyzed as for HCR v3.0 samples.