TERT immortalized human fibroblast (BJhTert) [11] cells were seeded on 8-well chamber slides (LabTek, Nunc), 20,000 cells per well. The cells were allowed to attach overnight and subsequently starved in Modified Eagle Medium (MEM, Gibco)+0.1% FCS (heat-inactivated, Sigma) for 48 h. For stimulation PDGF-BB was added to a final concentration of 100 ng ml⁻¹ and the cells were incubated at 37°C for the indicated time points. If the cells were treated with drugs, the drugs Gleevec (Novartis) and 5-Iodotubercidin (Sigma), dissolved in DMSO (Sigma) to a 10 mM stock concentration, were applied at a final concentration of 10 µM in fresh MEM+0.1% FCS for 1 h at 37°C before PDGF-BB was added at the indicated time points. The drug concentration was held constant throughout the experiment.

For fixation the cells were put on ice and the medium was immediately exchanged for ice-cold DEPC (Applichem Lifescience)-treated PBS (DEPC-PBS, 900 µl per well). After a 1 min wash, the cells were fixed in 2% (w/v) paraformaldehyde (PFA, Sigma Aldrich) in DEPC-PBS for 30 min at room temperature (20–23°C). Prior to permeabilization in ice cold 70% ethanol for 30 min on ice, the cells were washed again in ice-cold DEPC-PBS. Now the silicon mask of the chamber slides was removed, the slides were dehydrated through a series of 70%, 85% and 99.5% ethanol for ∼1 min each, hydrophobic barrier pen was applied at the borders of the slide wells and finally 8-chamber Secure Seals (9 mm in diameter, 0.8 mm deep; Grace Bio-Labs) were attached to the slides.

One mg donkey anti-rabbit IgG (711-005-152, Jackson Immunoresearch) and 1 mg donkey anti-mouse IgG (715-005-150, Jackson Immunoresearch) were concentrated over Amicon ultra 0.5 ml, 10 kDa (UFC501008, Millipore) and dialyzed in 7 kD dialysis cup (Slide-A-Lyzer MINI dialysis units, #69562, Pierce) against PBS over night at 4°C. Antibodies were activated in SMCC (Pierce) added in 25-fold excess over antibody for 2 h at room temperature. Primer oligonucleotide (5′ thiol-AAA AAA AAA ATA TGA CAG AAC TAG ACA CTC TT (Eurogentec)) and blocked oligonucleotide (5′ thiol-AAA AAA AAA AGA CGC TAA TAG TTA AGA CGC TTU UU (Biomers)) were degassed at 95°C for 3 min and reduced by incubation with 25 mM DTT for 1 h at 37°C. Both, antibodies and oligonucleotides were purified using an Illustra NAP-5 column (#17-0853-01, GE Healthcare), equilibrated and eluted with PBS+5 mM EDTA pH 7.5. Afterwards anti-rabbit IgG was immediately mixed with the primer oligonucleotide and anti-mouse IgG with the blocked oligonucleotide. PLA probes were then dialyzed against PBS at 4°C overnight. Before purification by HPLC the conjugates were concentrated to 25 µl by Amicon ultra 0.5 ml, 10 kDa columns and 10 mM DTT was added to the conjugate to reduce unbound oligonucleotide dimers. The conjugates were incubated for 10 min at room temperature and subsequently injected into HPLC (Hitachi D7000 series). Chromatography was done over a Superdex 75 PC 3.2/30 column with a flow rate of 0.06 ml min⁻¹ to separate conjugates from free oligonucleotides and samples were collected at 16–20.5 min (conjugate peak).

The protocol is based on the protocol described previously in Jarvius et al. 2007 [5]. All reactions were performed in Secure Seals attached to the slides. Prior to each new step all liquids were removed from the Secure Seal chambers and for incubations exceeding 30 min the chambers were sealed with q-PCR film to avoid evaporation. All washes were performed by flushing the wells with ∼500 µl of the appropriate washing buffer.

First the wells were rehydrated in 1× PBS-T (DEPC-PBS+0.05% Tween20 (Sigma)) for 10 min at room temperature. The cells were blocked in Protein-Block, serum free (Dako), 2.5 ng µl⁻¹ sonicated salmon sperm DNA (Invitrogen) and 2.5 mM L-cysteine (Sigma) for 90 min at 37°C. After blocking primary antibodies (63 ng ml⁻¹ rabbit-anti-PDGFRβ (#3169 Cell Signaling) and 6 µg ml⁻¹ mouse-anti-pY100 (#9411, Cell signaling)) were applied in blocking solution over night at 4°C. Primary antibodies were removed by flushing the chambers with ∼500 µl 1× PBS-T. Secondary PLA probes (1.6 ng µl⁻¹ donkey-anti-rabbit-primer and 5 ng µl⁻¹ donkey-anti-mouse-blocked, final concentration, see above) were incubated separately in half the final volume of blocking solution for 30 min at room temperature before they were mixed and applied to the cells for 1 h at 37°C. Subsequently the cells were washed in 10 mM Tris HCl pH 7.5, 0.1% Tween20 for 5 min at 37°C and once with 1× PBS-T. Hybridization was done with 125 nM connector oligonucleotides (5′ phosphate-CTA TTA GCG TCC AGT GAA TGC GAG TCC GTC TAA GAG AGT AGT ACA GCA GCC GTC AAG AGT GTC TA (Eurogentec) and 5′-phosphate-GTT CTG TCA TAT TTA AGC GTC TTA A (Eurogentec)) in 1× ligation buffer (10 mM Tris acetate, pH 7.5, 10 mM magnesium acetate, 50 mM potassium acetate), 0.25 µg µl⁻¹ BSA (New England Biolabs), 0.05% Tween20, 250 mM NaCl for 30 min at 37°C. After a wash in 1× PBS-T ligation was performed in 1× ligation buffer, 0.05 U µl⁻¹ T4-DNA ligase (Fermentas), 1 mM ATP (Fermentas), 0.25 µg µl⁻¹ BSA, 0.05% Tween20 and 250 mM NaCl for 30 min at 37°C. Subsequent to an additional washing step in 1× PBS-T, rolling circle amplification was done with 0.25 U µl⁻¹ phi29 DNA polymerase (Fermentas) in 1× RCA buffer (50 mM Tris-HCl, 10 mM MgCl₂, 10 mM (NH₄)₂SO₄, pH 7.5), 0.25 mM dNTP (Fermentas), 0.2 µg µl⁻¹ BSA, 5% glycerol (Sigma) 90 min at 37°C. Before detection with 100 nM detection oligonucleotide (5′ Alexa 555-CAG TGA ATG CGA GTC CGT CT 3′ (Trilink)) in 2× SSC, 0.25 µg µl⁻¹ BSA, 75 ng µl⁻¹ poly(A) (Sigma) and 0.05% Tween20, the cells were washed once in 1× PBS-T. After detection Secure Seals were removed and the slides were washed twice in TBS for 10 min each. The cytoplasm of the cells was thereafter counterstained with 2.5 µg ml⁻¹ WGA 488 (Invitrogen) diluted in 1× DEPC-PBS for 30 min at room temperature. Final washes were done 2× 10 min in TBS before the slides were spun dry and mounted in Vectashield mounting medium (Vector)+100 ng ml⁻¹ DAPI.

Primers and padlock probes were designed for the DUSP6 and ACTB transcripts using GenBank accession numbers NM_001946 (DUSP6) and NM_001101.3 (ACTB). The primers used for in situ synthesis were for DUSP6 5′-CCG CTG GCT CTT AGT GTC AAT GAA T-3′ (Exiqon) and for ACTB transcripts 5′-GTG GAC GGG CGG CGG ATC GGC AAA G-3′ (Exiqon). Bold letters represent LNA-modified bases. The padlock probes and detection oligonucleotides applied for detection of DUSP6 and ACTB transcripts were 5′ phosphate-GAG AGA GAT TCA TTG TTC CTT TTA CGA CCT CAA TGC TGC TGC TGT ACT ACT CTT CCT GGG CAG CTT CAT T (Integrated DNA Technology) for DUSP6 and 5′ phosphate-AGC CTC GCC TTT GCC TTC CTT TTA CGA CCT CAA TGC ACA TGT TTG GCT CCT CTT CGC CCC GCG AGC ACA G (Integrated DNA Technology) for ACTB. Underscored letters denotes the target complementary ends and italic letters represent the site and the sequence of the respective detection oligonucleotides. The detected DUSP6-transcripts were Cy3-labeled (MWG) for the individual time series experiments and Cy5-labeled (Eurogentec) in all other experiment. ACTB-transcripts were Cy5-labeled (Integrated DNA Technology).

The in situ detection-procedure was performed in a similar way as described by Larsson et al. [7]. Fixation, permeabilization and dehydration of cells were however done as described above. Subsequent to the application of the Secure Seals, the cells were rehydrated in PBS-T. Reverse transcription was performed with 1 µM of the appropriate LNA-primer in 20 U µl⁻¹ RevertAid H minus M-MuLV reverse transcriptase (Fermentas), 0.5 mM dNTPs, 0.2 µg µl⁻¹ BSA, 1 U µl⁻¹ Ribolock RNase inhibitor (Fermentas) in M-MuLV reaction buffer. Slides were incubated for 3 h at 37°C and then washed twice in PBS-T. All subsequent washes were performed in the same way. Postfixation was done in 2% (w/v) paraformaldehyde for 30 min at room temperature before the slides were washed. To make the cDNA accessible for padlock probe hybridization and to simultaneously ligate the padlock probe to its target sequence, the cells were incubated with 0.1 µM padlock probe, 0.4 U µl⁻¹ RNase H (Fermentas), 0.5 U µl⁻¹ Ampligase (Epicentre), 1 U µl⁻¹ Ribolock RNase inhibitor in Ampligase buffer (20 mM Tris-HCl, pH 8.3, 75 mM KCl, 10 mM MgCl₂, 0.5 mM NAD and 0.01% Triton X-100), 20% formamide (Merck), first for 30 min at 37°C and then for 45 min at 45°C. After an additional wash RCA was performed with 1 U µl⁻¹ phi29 DNA polymerase, 1 U µl⁻¹ Ribolock RNase inhibitor in the supplied phi29 DNA polymerase reaction buffer, 0.25 mM dNTPs, 0.2 µg µl⁻¹ BSA and 5% glycerol for 90 min at 37°C. RCPs were labeled with 0.25 µM detection oligonucleotide in 2× SSC and 20% formamide for 15 min at 37°C. After detection the slides were washed twice in PBS-T and Secure Seals were removed. The cytoplasm of the cells was thereafter counterstained and the slides mounted as described above.

The procedure for simultaneous detection of individual proteins and transcripts in single cells was based on merging the two protocols described above. The merged protocol started with in situ reverse transcription, converting mRNA to DUSP6 and ACTB cDNA. This was followed by post-fixation of the cells, digestion by RNase H and hybridization/ligation of the specific padlock probes to the target cDNA molecules. These steps were performed in the same fashion as described above for solely transcript detection. Thus, the padlock probes were locked onto their targets and the target transcripts were secured. The combined protein/mRNA protocol could subsequently proceed with the detection of phosphorylated PDGFRβ as described above except that 1 U µl⁻¹ Ribolock RNase inhibitor was added to the blocking, antibody and PLA probe mixtures and DEPC-ddH₂O was used instead of ddH₂O in hybridization and ligation. RCA was performed as described for in situ PLA except that 2.5 U µl⁻¹ phi29 DNA polymerase were used and both, padlock probes and in situ PLA signals were amplified. For detection, the fluorescence labeled oligonucleotides (see above), with a final concentration of 100 nM, were added to the detection mix described for in situ PLA.

Images were acquired using a Zeiss Axioplan 2 epifluorescence microscope, the AxioCam MRm CCD sensor and a 20×/0.8 Plan Apochromat lens together with filters for DAPI, FITC, Cy3 and Cy5. For each condition z-stacks were acquired at three positions of the well and the positions were chosen using the FITC channel (cytoplasm staining).

Image J 1.44 d was used to make maximum intensity projections (MIPs) of the image stack of each channel. In the FITC channel, ImageJ was applied to enhance the contrast, subtract background (rolling ball radius = 200 pixel) and run a median filter (radius 4 pixel). Background was also subtracted from DAPI-images (rolling ball radius = 50 pixel).

Subsequently, CellProfiler v.10415 was used to identify cell nuclei (thresholding method: two-class thresholding “Otsu PerObject”) in previously background subtracted DAPI images. The nuclei were then combined with background corrected FITC images to identify the cytoplasm of individual cells (method to identify objects: “Watershed - Image”, thresholding method: two-class thresholding “Otsu Adaptive”). Signals in Cy3 (PDGFRβ) and Cy5 (DUSP6 or ACTB) were enhanced (feature type: “speckles”, feature size: 10), counted (diameter: 2–20 pixel, thresholding method: “RobustBackground Global” (for Cy3), diameter: 1–20 pixel, thresholding method “RobustBackground Adaptive” (for Cy5), threshold correction factor: 0.8) and related to the individual cells. Subsequently, if images did not contain signals, the amount of signals had to be manually corrected, since then the adaptive enhancement of Cy3 and Cy5 signals gave rise to false positive signals. Graphs were done in R [12]. In Figure 1B images were for better visualization treated in ImageJ as described here: One slice of the DAPI z-stack was selected, background was subtracted in ImageJ and thresholded for brightness and contrast. Cy3 and Cy5 z-stacks were projected to MIPs, the background was removed, brightness and contrast adjusted and a maximum filter (size: 1 pixel) was applied.

To determine the kinetics of PDGFRβ phosphorylation and DUSP6 expression upon stimulation with PDGF-BB, BJhTert cells were starved for 48 h prior to treatment with 100 ng ml⁻¹ PDGF-BB. The cells were stimulated for different length of time (0 min–240 min) and subsequently fixed. Thereafter the phosphorylation of PDGFRβ was investigated using in situ PLA with one antibody directed against the receptor and one against phosphorylated tyrosine residues. In parallel, the expression on DUSP6 transcripts was monitored by padlock probes on separate slides (Figure 2). As expected, the phosphorylation of PDGFRβ first increased substantially during the initial time points, peaking at 5–10 min after stimulation, before it gradually decreased again towards background levels due to internalization of the receptor [9]. In accordance with what has been reported before [10], DUSP6 transcript levels remained constant for the first 30 min of stimulation, and then showed an increase until 3 h. Interestingly, the distribution in numbers of detected molecules per cell in the cell population is quite large, especially for PDGFRβ phosphorylation in response to stimulation. Regarding DUSP6 expression, not all cells are responding to stimulation with PDGF-BB to the same extent. It rather appears that, for the time investigated, some cells burst with transcription while in others the signal is not propagated to the induction of DUSP6 expression.

Next, we tested the combined detection of protein modifications and single transcripts (Figure 1). For this purpose phosphorylation of PDGFRβ and ACTB mRNA molecules – an abundantly expressed housekeeping gene to enable a more detailed analysis of detection efficiency – were detected either individually or simultaneously. The cells were fixed and permeabilized before cDNA synthesis and padlock probe hybridization and ligation were performed. After the ACTB cDNA had been targeted, the cells were blocked and stained with in situ PLA reagents required to detect phosphorylated PDGFRβ. Both, reacted padlock probe and in situ PLA probes were amplified in the same reaction by RCA and detected by hybridization of fluorescence labeled detection oligonucleotides. We investigated whether the combined detection shows similar results as when the two detection reactions were run separately, and whether there is a decrease in detection efficiency upon combining the two. BJhTert cells were stimulated with PDGF-BB for 0 min, 5 min or 180 min. The results of the combined detection were consistent with the results of individual staining and the levels of ACTB transcripts were unaffected by stimulation with PDGF-BB, as expected. The combined protocol leads to a decrease in detected signals for both in situ PLA and padlock probes of about 50% compared to when performed in separate reactions (Figure 3).

Having developed a method that enabled measurement of both protein activity and mRNA expression in single cells, we then tested if the assay would be suitable to detect perturbation of the PDGFRβ-signaling pathway using drugs that affect the pathway at different levels. In that way we were able to investigate if the assay could help in zooming in on the action of a drug. As proof of principle, we examined the effect of Gleevec (a tyrosine kinase inhibitor [13] targeting the PDGFRβ directly) and 5-Iodotubercidin (an inhibitor of ERK2 activation [14], located downstream of PDGFRβ but upstream of the induction of DUSP6 transcription) on phosphorylation of PDGFRβ and DUSP6 expression. As expected, treatment with Gleevec eliminated the increase in signals for both molecules, since it inhibits the activation of the PDGFRβ and thus all downstream events. 5-Iodotubercidin, on the other hand, did only eliminate the upregulation of DUSP6 but not phosphorylation of the PDGFRβ, since it affected the signaling downstream of the receptor but upstream of DUSP6 (Figure 4).