A simplified and rapid in situ hybridization protocol for planarians

Gene expression analysis is a foundational tool in cell and developmental biology studies. Next-generation sequencing technologies have revolutionized the breadth and resolution for examining the molecular profile of cells and how they change during development or in response to stimuli. However, validating and following up on interesting leads from sequencing projects requires low-throughput techniques. In situ hybridization (ISH) is an established method for analyzing expression patterns that provides valuable spatial information for understanding the context of gene expression. Protocols for ISH are laborious as they must accomplish several important goals. First, fixation: mRNA molecules must be locked in position and tissue morphology preserved. Second, permeabilization: staining reagents need access to target molecules. Third, hybridization: chemically modified RNA probes need to specifically hybridize to target mRNA molecules. Finally, detection: methods for visualizing the location of probe binding must provide a strong signal with minimal background staining. Numerous protocols have been developed for ISH in diverse organisms [1–4] and while there are often species-specific techniques, many of the processes and reagents are shared across model systems.

Planarians have emerged as a valuable model system for studying stem cell biology, regeneration and other foundational topics in biology [5]. Planarians present some unique challenges for ISH and immunostaining techniques: they secrete a layer of mucus that must be removed prior to fixation to allow penetration of staining reagents and many species produce body pigment that can hinder visualization in staining experiments. Previous work has focused on developing reliable protocols with improved spatial resolution and sensitivity of staining [1,6–8]. Like protocols for other model organisms, ISH protocols for planarians are laborious, multiday procedures that limit throughput and are challenging to incorporate into laboratory experiences for undergraduate students. Here, the authors focused on modifying, combining or eliminating steps of the procedure to reduce the overall time and number of contact points with samples.

Materials & methods

Animal husbandry

Asexual Schmidtea mediterranea clonal line CIW4 and Dugesia japonica were maintained as previously described [7].

Probe synthesis

Genes of interest were PCR-amplified from cDNA and cloned into pJC53.2 as previously described [9]. The DNA template for probe synthesis was generated by PCR amplification using T7 primers and purified using a DNA clean and concentrator kit (Zymo Research, CA, USA). Antisense RNA probes were synthesized by in vitro transcription with either DIG-11-UTP (Sigma-Aldrich, MO, USA) or Fluorescein-12-UTP (Sigma-Aldrich) using either the T3 or SP6 riboprobe system (Promega, WI, USA) with modifications from the manufacturer's suggested protocol. The transcription reaction volume was reduced to 10 μl and 0.5 units of thermostable inorganic pyrophosphatase (New England Biolabs, MA, USA) was included in the reaction. Following the transcription reaction, the DNA template was degraded by treatment with DNase (Promega, WI, USA) before being analyzed on a 1% agarose gel. Successful reactions were diluted 1:100 in prehybridization solution and stored at -20°C.

In situ hybridization

The protocol from King and Newmark, 2013 [7], served as a starting reference against which to compare protocol alterations. Modifications to the protocol included the addition of 20% methanol, 10% acetic acid and 5 mM ethylenediaminetetraacetic acid to the fixative; increasing the H₂O₂ concentration to 6% and adding detergent to the bleach solution; elimination of the postfixation dehydration and proteinase K steps; reduced washing times and steps; inclusion of cold-water fish gelatin in the blocking solution; room temperature antibody incubation; elimination of polyvinyl alcohol (PVA) from the development solution; and colorimetric development at 37°C.

Immunofluorescence & FISH

Samples were processed using the rapid protocol with probes visualized using tyramide signal amplification as previously described [7]. M-phase cells were labeled using 1:5000 antiphospho-histone H3 (Cell Signaling Technology, MA, USA) and 1:1000 antirabbit-HRP (Jackson ImmunoResearch Laboratories, PA, USA) and detected by tyramide signal amplification. Ciliated protonephridia were labeled with 1:1000 antiacetylated-alpha tubulin (Santa Cruz Biotech, TX, USA) and 1:500 Alexa 488 antimouse (Thermo Fisher Scientific, MA, USA). Muscle was labeled with 1:50 6G10 antibody (Developmental Studies Hybridoma Bank, IA, USA) and 1:1000 antimouse-HRP (Jackson ImmunoResearch Laboratories) followed by tyramide signal amplification. For BrdU labeling, animals were soaked in 20 mg/ml 5-bromo-2′-deoxyuridine with 3% dimethyl sulfoxide in a 6 g/l Instant Ocean salt solution for 2 h followed by a 4-h chase prior to processing using the rapid protocol. BrdU incorporation was detected using 1:25 anti-BrdU antibody (BD Biosciences, NJ, USA) and 1:1000 antimouse-HRP (Jackson ImmunoResearch Laboratories, PA, USA) followed by tyramide signal amplification.

Multicolor FISH

Samples were hybridized with digoxigenin and fluorescein-labeled probes for the genes of interest. Sequential detection using two rounds of tyramide signal amplification was performed as previously described [7] with a 100 mM azide quenching step between rounds of amplification. Simultaneous detection was performed by incubating with both 1:2000 antidig-AP (Sigma-Aldrich) and 1:2000 antifluorescein-POD (Sigma-Aldrich), followed by tyramide signal amplification and development using the fluorescent alkaline phosphatase substrate Fast Blue BB (Sigma-Aldrich) as described [8,10].

Imaging

Samples were cleared in 80% glycerol and mounted on slides. Images of colorimetric samples were collected with an Olympus SZ61TR stereomicroscope (Olympus, Tokyo, Japan) with an AmScope MU1803-HS camera using AmLite software (AmScope, CA, USA). Fluorescent samples were imaged using an Olympus Fluoview FV1200 laser scanning confocal microscope (Olympus) and images were processed using FIJI software [11].

Results & discussion

Modification of fixation conditions eliminates enzymatic permeabilization steps for high-abundance transcripts

The sensitivity and reliability of ISH in planarians have been the focus of several studies [1,6–8]. While published protocols produce samples with good morphology and strong signal-to-background (Figure 1A & B) [7], they are lengthy, with numerous sample processing steps, and require a few days to complete. We wondered whether ISH could be optimized for time with minimal reduction in sample quality and signal strength. Two key parameters that need to be balanced for successful ISH are fixation and permeabilization. mRNA molecules must be retained in position and morphology maintained while staining reagents need to penetrate deep tissues of the sample for labeling. Fixation of planarians with 4% formaldehyde without further postfixation dehydration and enzymatic permeabilization steps resulted in a weak signal with poor staining in the prepharyngeal region (Figure 1C & D). As was shown for other model systems [12], the addition of acetic acid to the fixative modulates formaldehyde crosslinking, eliminating the need for additional permeabilization steps (Figure 1E & F). The incorporation of acetic acid in the fixative likely contributes more than simply attenuating crosslinking, as reducing formaldehyde concentration or quenching crosslinking using a Tris buffer [13] increases sample fragility with mixed effects on signal intensity (Supplementary Figure 1). The inclusion of methanol in the fixative further improves the consistency and intensity of staining (Figure 1G & H). The modified fixative also seems to increase the flexibility in timing for mucus removal using the mucolytic N-acetyl-l-cysteine (NAC) with only modest improvements for longer NAC incubation times (Supplementary Figure 2).

 Similar levels of in situ signal and background with high-abundance transcripts were found for both the rapid in situ protocol and samples fixed with 4% formaldehyde and permeabilized using proteinase K as previously described (Supplementary Figure 3A & D). However, for several low-abundance transcripts, the in situ signal was weak or absent using the rapid protocol (Supplementary Figure 3B & C). There could be several factors leading to the difference in sensitivity of the rapid protocol. The fixation conditions for the rapid protocol may not adequately fix the transcripts in place, allowing for their diffusion or degradation during processing. Alternatively, the rapid protocol may not sufficiently permeabilize the tissue, allowing adequate access to the transcripts. Signal sensitivity may vary for transcripts using the rapid protocol depending on their abundance, cell type of expression or RNA structure, such as inclusion in ribonucleoprotein granules. To increase the utility of the rapid protocol, we examined whether combining rapid bleaching and posthybridization processing with modified fixation and permeabilization steps could improve sensitivity while maintaining a significant reduction in protocol time. We found greatly improved sensitivity by performing a 7.5-min NAC treatment, fixing in 4% formaldehyde solution, dehydrating in 100% methanol, bleaching for 1 h in 6% H₂O₂ (no detergent) and performing a 10-min proteinase K permeabilization step prior to hybridization and processing using the rapid protocol (Supplementary Figure 3E & F). We also examined whether adding a proteinase K permeabilization step to the rapid protocol after bleaching improved sensitivity. To prevent samples from disintegrating during the proteinase K step, we performed the bleaching step in 6% H₂O₂ without detergent and did not agitate the samples during the proteinase K incubation. While the inclusion of the proteinase K step in the rapid protocol improved sensitivity for low-abundance transcripts compared with the rapid protocol alone, sensitivity was sometimes reduced compared with fixation in 4% formaldehyde (Supplementary Figure 3E & H), and sometimes improved (Supplementary Figure 3F & I). Based on the genes we tested, 4% formaldehyde fixation and proteinase K permeabilization combined with the hybridization and wash steps of the rapid protocol is a suitable modification for genes where the rapid protocol alone is not sufficiently sensitive.

Increasing bleaching strength improves permeability

Several hydrogen peroxide-based bleaching solutions have been described for bleaching body pigment in planarians [1,6,7], including a 1.2% H₂O₂/formamide bleach capable of producing greater sensitivity of staining for low-abundance genes [7]. However, this formamide-based bleaching solution requires a couple of hours of incubation to achieve full bleaching and, at shorter incubation times, not only does pigmentation remain, but the staining intensity is diminished in the absence of an enzymatic permeabilization step (Figure 2A–C). Increasing the concentration of H₂O₂ in the formamide solution to 6% reduced bleaching time and improved signal but, occasionally, samples showed weak staining in the prepharyngeal region (Figure 2D–F). Enzymatic permeabilization steps not only use proteinase K to improve staining but also detergents such as sodium dodecyl sulfate (SDS). Adding SDS and Triton X-100 to the 6% H₂O₂ formamide bleaching solution improved the consistency of staining in the prepharyngeal region (Figure 2G–I).

Optimizing blocking & wash steps

Solution changes and wash steps are one of the most laborious aspects of ISH with each processing step increasing the chances of damaging samples. Therefore, we set out to test if extensive wash steps are necessary for reducing background staining. Effective washing requires a difference in the concentration of components between the sample and the wash solution and time for diffusion of components in or out of the sample. We reasoned that equilibrium between the sample and wash solution occurs quickly for the first change to a new solution as the residual liquid left behind in the tube gets rapidly diluted in the newly added wash solution and that a few short washes followed by longer washes might be sufficient for reducing background. We tested several washing methods to identify the most efficient for reducing background staining (Figure 3). Mock hybridized planarians incubated with blocking solution lacking anti-DIG-AP antibody showed no background staining after 4 h of development (Figure 3A & B). Samples washed one time for <1 min prior to development showed a significant amount of background staining (Figure 3C & D). Samples washed with 4 solution changes for a total of 35 min showed only slightly more background staining compared with the recommended 9 solution changes over 2.25 h (Figure 3E–H) [7], indicating that wash time and steps could be substantially reduced.

Even the long washing procedure still produced background compared with the no-antibody controls, suggesting that the antibody may be nonspecifically binding. Cold-water fish gelatin has been used as a blocking reagent for immunostaining experiments in planarians [14]. We tested whether the inclusion of fish gelatin in the blocking solution could reduce background staining and found a significant reduction in background compared with the blocking solution containing horse serum and Roche Western Blotting Reagent alone (Figure 3A–H). The addition of fish gelatin to the blocking solution did not inhibit the binding of the antibody to its target, as development time and signal intensity were similar whether it was present in the blocking solution or not (Figure 3I & J). Antifluorescein-POD antibodies used for FISH show significant background staining that can be reduced by the inclusion of a Roche Western Blocking Reagent in the blocking and antibody solution [7]. We examined whether also including fish gelatin in the blocking and antibody solution would further improve background staining, but found comparable levels of background with and without fish gelatin present (Supplementary Figure 4). We also did not note significant differences in background staining with strong probes between blocking solutions with and without fish gelatin, but the results indicate the inclusion of fish gelatin may be particularly beneficial for the detection of low-abundance genes.

With the maintained quality and time savings found by reducing postantibody wash steps, we tried reducing other wash steps, particularly posthybridization. We reasoned that any unhybridized probe remaining in the samples would be degraded by nucleases in the blocking solution at later steps and found that shortening these washes still generated satisfactory results (most figures show samples with reduced posthybridization washes). Additionally, quality staining can be obtained with reduced prehybridization and preantibody blocking times as well as 2-h room temperature antibody incubation times (most figures show samples with shortened blocking and antibody incubation times). These modifications provide substantial time savings, eliminating a full day off protocol time and reducing hands-on sample processing time.

Incubation at 37°C increases rate of development

Colorimetric development using the alkaline phosphatase substrate NBT/BCIP can take a few hours to complete. The rate of development has been shown to increase with the addition of macromolecular crowding agents such as PVA [1]. However, when testing several PVA concentrations, a significant difference in development rate between reactions containing PVA or not was not observed (Supplementary Figure 5). The viscosity of a development solution containing PVA increases the challenge of removing the solution without pipetting up or damaging samples. Furthermore, failure to remove all PVA prior to ethanol washing causes any remaining PVA to precipitate. Given the minimal benefits and additional challenges, we found it more efficient to exclude PVA from the development solution. Next, we examined whether increasing the temperature of development would reduce development time without increasing background staining. Samples incubated at 37°C showed significantly faster development, reaching an ideal staining intensity in a shorter time with no difference in signal-to-background staining (Figure 4).

Rapid protocol has broad compatibility with common stains

We next tested the versatility of the rapid ISH protocol for other common analyses. Double FISH is typically a 4-day protocol and can require substantial optimization to achieve consistent results [7]. Using the rapid protocol modifications, we were able to reduce protocol time to 3 days while maintaining strong labeling with minor background staining (Figure 5A). Immunofluorescence staining can be challenging, as fixation and processing conditions can greatly impact epitope accessibility [15,16]. We tested four commonly used antibodies for S. mediterranea (the M-phase marker phospho-histone H3 [17], the muscle marker 6G10 [16], the cilia marker acetylated-α-tubulin [18] and an antibody for detection of incorporated BrdU [14]) and found that all showed normal staining in combination with FISH using the rapid protocol for fixation and processing (Figure 5B–F). In addition, we examined whether the rapid protocol also worked for D. japonica. With slightly reduced NAC treatment concentration and time, D. japonica samples remained intact through the protocol and showed strong labeling compared with samples where mucus removal was accomplished using a 2% HCl treatment [19] (Supplementary Figure 6).

To further reduce protocol time, we examined the feasibility of the simultaneous detection of two genes through a combination of tyramide signal amplification and a fluorescent alkaline phosphatase substrate, Fast Blue BB [8,10]. Using this approach, we were able to incubate with both antidig-AP and antifluorescein-POD antibodies simultaneously, allowing us to reduce the protocol time to two days (Figure 6).

Conclusion

ISH protocols can be substantially shortened with minimal impact on staining quality through modification of components and optimization of procedures. Permeabilization can be incorporated into fixation and bleaching steps through the addition of acetic acid and methanol to the fixative and the inclusion of detergents in the bleaching solution, thereby eliminating the need for a separate enzymatic permeabilization step. Numerous long washes are unnecessary for all but possibly the very largest samples, and wash time and labor can be saved by optimizing wash conditions and by including fish gelatin in the blocking solution. Colorimetric development time can be reduced by incubating samples at higher temperatures. In combination, these enhancements eliminate 20 sample handling steps and a full day of protocol time compared with previous methods (Table 1) [7].

Future perspective

ISH has been a crucial tool in planarian research for analyzing gene expression. This study provides researchers with a simplified and time-efficient method for ISH using planarians. Not only will this protocol help researchers to process more samples with less labor and a quicker time to results, but it will also increase the feasibility of introducing this method in academic teaching laboratories that typically do not have schedules conducive to carrying out the traditionally lengthy protocols. Additionally, while planarians provide some unique challenges for ISH, the reagents and procedures used for planarians are generally shared with other systems and it is likely that the modifications presented here may be applicable to other model organisms.