Abstract
We report an artifactual in situ hybridization (ISH) labeling pattern in embryonic rat tissues. It is caused by a short multiple cloning site-derived sequence incorporated into the RNA probes by in vitro transcription of templates cloned into pBluescript or its descendants. The artifact was seen in tissues in which programmed cell death (apoptosis) takes place during embryogenesis, i.e., in the mesonephric area, developing nervous system, interdigital mesenchyme of the hand plate, and permanent kidney. Labeling of the radioactive ISH with TUNEL verified the co-localization of the artifactual hybridization signal with cells at early stages of apoptosis. Even though the identity of the hybridization target in apoptotic cells remains unknown, it might be highly species-specific, because this artifact was never observed in mouse tissues.
I
The interpretation of ISH labeling is critically dependent on appropriate controls. A widely used method is to transcribe the probe sequences also in the sense direction to provide a control probe with identical hybridization parameters as in the antisense probe. Thus, the sense control is believed to reveal unspecific interactions of the probe with tissues. The template sequence is often cloned into a multiple cloning site (MCS) of a plasmid vector flanked with two distinct promoters enabling the synthesis of sense and antisense probes by different RNA polymerases. The length of the transcript is limited by linearizing the plasmid from restriction sites in MCS downstream to the insert. Short sequences resulting from transcription of the vector itself, between the RNA promoter and insert, are always incorporated into the probe. These synthetic sequences are commonly considered to be insignificant for the hybridization.
We have used radioactive ISH as a routine method, which has provided satisfactory results with more than 100 probes over several years. However, sections from rat embryos have occasionally displayed confusing signals. Therefore, we have now analyzed these experiments and identified a short sequence of the pBluescript (Stratagene; La Jolla, CA) MCS that is responsible for the artifactual hybridization pattern. In embryonic rat samples, this artifact is associated with the regions of apoptosis. By double detection of radioactive ISH and TUNEL, we were able to co-localize the cells hybridizing with the multilinker sequences and those at early stages of apoptotic death.
Materials and Methods
Animals, Tissues, and Preparation of Sections
Pregnant Sprague–Dawley rats were sacrificed and the embryos were dissected in Dulbecco's phosphate-buffered saline (PBS). The day after overnight mating was considered as the Day zero of embryogenesis (E0). E13 embryos were fixed overnight in 4% paraformaldehyde (PFA) in PBS at 4C and E13 urogenital blocks and E15 kidneys for 2 hr at room temperature (RT). Developmental stage-matched samples were also collected from CBA × NMRI mice (plug day was E0). The embryos and tissues were mounted in paraffin after ethanol series and 7-μm sections were cut to silane-coated glasses.
Plasmids
The pBluescript KS+/– and SK+ vectors were from Stratagene and pGEM3Zf(+) from Promega (Madison, WI). The MCS of pBluescript KS+ was deleted by KpnI-SacI digestion and the ends blunted with T4 polymerase (Promega) and joined via intramolecular ligation, which created a new MscI restriction site between the T3 and T7 promoter. The resulting plasmid was named pBluescript ΔKS. The pBL-KA15 plasmid (a gift from Dr. Nagata, Japan) contained a full-length cDNA of mouse FasL inserted into the XhoI site of pBluescript KS+ in antisense orientation from the T3 promoter. The pMF-1 plasmid (also a gift from Dr. Nagata) contained a full-length murine Fas cDNA inserted into the EcoRI site of pBluescript KS+ in sense orientation from the T3 promoter. Sequencing and restriction enzyme digestions verified the orientation of the template sequences and the identity of the plasmids.
Preparation of [35S]-cRNA Probes
The plasmids used as templates were purified with a maxiprep kit (Qiagen; Hilden, Germany) according to guidelines given in the company's manual. The plasmids were fully linearized with suitable restriction enzymes overnight and after reaction the restriction mixtures were phenol–chloroform-extracted and ethanol-precipitated. The linearized plasmids were then in vitro transcribed in the presence of [35S]-UTP. The reaction mixture consisted of 1 × transcription buffer (Promega), 1 μg template plasmid/reaction, 0.2 M DTT (Promega), 2.5 mM GTP/ATP/CTP (Promega), 50 μCi [35S]-UTP (>1000 Ci/mmol) (Amersham Pharmacia Biotech; Uppsala, Sweden), 20 U of rRNase inhibitor (Promega), and 15–20 U of appropriate RNA polymerase (Promega). The reaction volume was adjusted to 20 μl with nuclease-free water (Promega). After incubation (37C for 1 hr) 0.5 U of RQ1 RNase-free DNase (Promega) was added and incubation was continued for 30 min. The transcription product was purified by a nick column (Amersham Pharmacia Biotech). From 400 μl fractions of the eluate [10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 10 mM DTT, 0.1% SDS], a 0.5 μl sample was taken and its radioactivity was measured with a scintillation counter (LKB Wallac; Turku, Finland). Only the labeling products with the peak activity in the second fraction were used for the subsequent hybridization step. We routinely discarded weakly labeled probes (total activity less than 8.0 × 107 cpm). The high-activity fraction was precipitated with 5 M NaAc (pH 5.2) and absolute ethanol overnight. The precipitated RNA was washed with 70% and absolute ethanol, air-dried, dissolved in a small amount of hybridization buffer [60% deionized formamide (Riedel-de Haën; Seelze, Germany), 0.3 M NaCl, 20 mM Tris-HCl (pH 8.0), 5 mM EDTA, 10% dextran sulfate (Sigma; Natick, MA), 1 × Denhardt's solution (Sigma), 0.5 mg/ml yeast RNA (Sigma), 100 mM DTT (Promega)] and diluted to a final concentration of 2.0 × 104 cpm/μl. During the dilution process the probes were kept on ice and denatured for 2 min at 80C before application to the slides.
In Situ Hybridization
ISH was performed as described by Wilkinson and Green (1990), with some modifications. The sections were taken to RT 30 min before pretreatment. They were deparaffinized twice for 10 min in xylene and hydrated by an ethanol series (2 × absolute ethanol, 2 × 94%, 1 × 85%, 1 × 70%, 1 × 50%, 1 × 30% ethanol, 3 min each). All prehybridization solutions were made in diethylpyrocarbonate-treated deionized water. After a 5-min PBS wash, the samples were fixed (4% PFA, 20 min), and washed twice with PBS. The sections were then treated with 7 μg/ml proteinase K (Sigma) in 50 mM Tris-HCl (pH 8.0), 5 mM EDTA for 15 min at RT. After a 5-min PBS wash, the slides were postfixed in 4% PFA for 20 min and briefly washed with deionized water. The samples were then treated with 0.1 M triethanolamine-HCl (pH 8.0)–acetic anhydride for 10 min. After two PBS washes, the glasses were dehydrated in an ethanol series (1 × 30%, 1 × 50%, 1 × 70%, 1 × 94% and 2 × absolute ethanol for 3 min, except for 5 min in 70%). The hybridization mixture was pipetted onto the air-dried glasses, covered with parafilm, and hybridized in a humidified chamber (50% formamide, 5 × SSC) at 52C for 16 hr.
After hybridization the glasses were washed with 5 × SSC, 10 mM DTT at 50C for 30 min. The high-stringency wash was performed with 50% deionized formamide, 2 × SSC, 30 mM DTT at 65C for 30 min. After three 10-min washes with NTE buffer [0.5 M NaCl, 10 mM Tris-HCl (pH 8.0), 5 mM EDTA], the samples were treated with 20 μg/ml ribonuclease A (Roche; Basel, Switzerland) in NTE buffer at 37C for 30 min and washed with buffer once more. Then the high-stringency wash was repeated. The glasses were washed with 2 × SSC and 0.1 × SSC at 37C for 15 min and dehydrated in ethanol series (30%, 60%, 80%, 95% ethanol) with 0.3 M ammonium acetate. After two washes with absolute ethanol, the glasses were air-dried.
Autoradiography
The dried glasses were dipped in NTB2 autoradiographic emulsion (Kodak; Paris, France) diluted 1 + 1 in deionized water containing 2% glycerol. After air-drying of the emulsion, the glasses were packed in light-tight boxes with silica gel and exposed at 4C for 2–3 weeks. The slides were developed with D-19 developer, fixed with sodium fixative (Kodak), and counterstained with hematoxylin (Shandon; Pittsburgh, PA), then dehydrated and mounted in Mountex (Histolab Products; Västra Frölunda, Sweden).
Labeling of Radioactive ISH with TUNEL
The sections hybridized as described above were developed, fixed, and TUNEL staining by the ApopTag In Situ Apoptosis Detection Kit (Intergen; Purchase, NY) was performed according to the manufacturer's instructions. The slides were then washed with PBS twice for 15 min and treated with equilibration buffer for 1 min. Then the reaction mixture was pipetted onto the sections and covered carefully with a plastic coverslip. The terminal transferase reaction was performed at 28–30C for 3 hr to avoid melting of the autoradiographic emulsion. The reaction was stopped by a wash solution for 30 min at RT. The slides were washed twice with PBS and mounted in Immumount (Shandon). The samples were photographed with an Olympus Provis AX70 microscope (Olympus Optical; Tokyo, Japan) with bright-field, darkfield, and phase-contrast optics and fluorescence microscopy.
Results
A distinct ISH pattern in embryonic rat tissues was found to be artifactual because it was not dependent on the orientation of the template sequences but rather on the identity of the plasmid used and the direction of the transcription. Furthermore, this hybridization pattern was never detected in embryonic mouse tissues of matching age. When the probe sequences were transferred to the pGEM3Zf(+) vector (Promega), the signal artifact was no longer seen (not shown).
To ascertain that the hybridization signal was derived from pBluescript sequences, we performed a series of ISH experiments with probes transcribed from the T3 promoter of “empty” pBluescript vectors linearized with BamHI (Figure 1). The artifactual signal was detected with pBluescript KS+ and KS– (MCS in KpnI → SacI orientation between T3 and T7), but not with pBluescript SK+, in which MCS is in the SacI → KpnI direction between T3 and T7. In addition, the MCS-deficient pBluescript ΔKS did not produce any signals. The hybridizing sequence was therefore localized to the KpnI end of the MCS. A probe produced from the T3 promoter of pBluescript KS+ linearized with HindIII (not shown) further limited the sequence between KpnI and HindIII restriction sites in the pBluescript MCS.
We searched the vector database (www.ncbi.nlm.nih.gov/VecScreen) with the 36-base KpnI-HindIII oligonucleotide sequence using BLAST (Madden et al. 1996) to identify other vectors, that could give rise to similar artifactual in situ signals when used as a template for the transcription of riboprobes. A 100% match was obtained with 13 vectors, all of which were from Stratagene. The KS and SK versions of pBluescript II, pBC, and Phagescript harbor the same KpnI-HindIII sequence adjacent to T3 and T7 promoters, respectively. The problematic sequence is also included in the cloning sites of the PCR-cloning vector pPCR-Script and the eukaryotic expression vector pOPRSVI/MCS, where it resides in the T7 promoter side of the multilinker sequence.
The artifactual ISH signal was observed in the E13 rat embryo in tubular structures and stromal cells of the posterior mesonephros (Figure 2A and 2B). In addition, the mesenchyme surrounding the notochord (Figure 2C and 2D) and single cells in the developing nervous system were labeled (Figures 2E–2G). At E15, the artifactual hybridization signal was seen in the forelimbs at the distal and central parts of the inter-digital mesenchyme (Figures 2H–2J) and in the meta-nephros, where strong punctate signals were observed in the renal medulla (Figure 3A and 3B). Routine hematoxylin counterstaining revealed typical apoptotic cells, with nuclear pyknosis and fragmentation, in all the above-mentioned locations (Figure 2G). By combining TUNEL staining with ISH, the artifactual in situ signal in the metanephric kidney was shown to co-localize with TUNEL-positive cells with large, diffuse nuclei (Figures 3A–3F).
Discussion
We report here that a short sequence in the multi-linker of pBluescript cloning vector hybridizes with a characteristic pattern to embryonic rat tissues. By comparing the results from specific restrictions of pBluescript vectors, we show that the region between KpnI and HindIII in the MCS is responsible for the false hybridization signal. Previously, pBluescript vector multilinker sequences have been shown to bind with the residual bodies of the stage IX and X seminiferous tubules in rat testis (Millar et al. 1994). In addition, an ISH artifact caused by pCR-Script MCS (Stratagene) has been reported from adult rat and pig brains (Blödorn et al. 1998). The sequences causing these artifactual signals have been mapped to the KpnI end of the identical MCS in the pCR-Script and pBluescript vectors, between the XhoI and BamHI (Blödorn et al. 1998) and the KpnI and XhoI sites (Millar et al. 1994). The sequence identified in our study partially overlaps with both of these regions.
The artifactual ISH signal was seen in tissues in which extensive apoptotic death is taking place during normal embryonic development. In E13 rat embryos, the vestigial embryonic kidney, the mesonephros, undergoes drastic regression starting from the posterior mesonephric elements (Torrey 1943), which were arti-factually labeled. In E15 rat, the hybridization artifact in the interdigital mesenchyme matches spatially and temporally with the apoptotic cell death beginning in mouse at E13 and ultimately leading to the separation of digits (Zakeri et al. 1994; Mori et al. 1995). In addition, in the developing permanent kidney prominent apoptosis takes place, mostly in the mesenchymal cells (Koseki et al. 1992; Coles et al. 1993). The dying cells can be seen as single cells or as small TUNEL-positive groups adjacent to the developing tubules in the medulla of E13 mouse kidney (Dudley and Robertson 1997). By performing TUNEL labeling of the ISH sections from E15 rat metanephros, we verified that the artifactual signal overlaps with the groups of TUNEL-positive nuclei.
Most cell death events during development are morphologically recognizable as apoptosis (Vaux and Korsmeyer 1999), which is characterized by the degradation of cellular DNA to increasingly smaller fragments (Walker et al. 1993). The widely used method for detecting apoptotic cells in tissue sections, the terminal deoxynucleotidyl transferase (TdT)-mediated dUTP–biotin nick-end labeling (TUNEL), is based on the tagging of 3′-OH ends of DNA generated by fragmentation (Gavrieli et al. 1992). Even though the breakdown of DNA takes place relatively late in the apoptotic cascade, it precedes actual cell death by hours (Bursch et al. 1990). Differences in the intensity of TUNEL staining might imply that the lightly stained cells represent early stages of apoptosis (Gavrieli et al. 1992). Dying T-cells have been shown to produce high levels of mRNAs at the time of first appearance of the apoptotic DNA laddering (Kerkhoff and Ziff 1995). Even though the final cell fragmentation to apoptotic bodies is accompanied by the cessation of transcription and RNase activation (Bursch et al. 1990), it is conceivable that cells undergoing earlier steps of apoptosis may produce specific messages and that these might be the source for the artifactual labeling.
(
The pattern of the artifactual ISH signal caused by the KpnI-HindIII sequence of the pBluescript MCS. (
The co-localization of the artifactual ISH signal and TUNEL-positive cells in the developing metanephric kidney. (
This ISH signal is particularly cumbersome to identify as an artifact. First, it is obviously highly species-specific and is never detected with embryonic mouse tissues. Second, the signal is immediately detected as false only if KpnI-HindIII sequence is incorporated into the sense probe. If it is part of the antisense probe, the negative sense control may lead to misinterpretation of the artifactual signal as a natural distribution of the target mRNA. The most deceptive combinatorial signals will be achieved when an antisense probe mixes the true expression pattern of the gene and the artifact. Third, the artifact is not dependent on the identity of the promoter used (T3 vs T7) per se, because the pBluescript is available as two versions containing the MCS in opposite directions (KS vs SK). Whether the KpnI-HindIII sequence is part of the sense or the antisense probe is dependent on both the version of the plasmid and the direction of the transcription. Finally, several different Stratagene vectors harbor the pBluescript MCS sequence causing the artifactual signal. Therefore, if the probes are intended for ISH of rat tissues, the pBluescript family of vectors should be avoided.
Footnotes
Acknowledgements
Acknowledgments
Supported by the Sigrid Jusélius Foundation and by the Academy of Finland.
We wish to thank Dr Shigekazu Nagata (Osaka Bio-science Institute; Osaka, Japan) for Fas and FasL probes. Ms Marja-Leena Peltonen and Ms Alla Hanninen are acknowledged for skillful technical assistance.