- Methodology article
- Open Access
Single-cell qPCR on dispersed primary pituitary cells -an optimized protocol
© Hodne et al; licensee BioMed Central Ltd. 2010
- Received: 13 July 2010
- Accepted: 12 November 2010
- Published: 12 November 2010
The incidence of false positives is a potential problem in single-cell PCR experiments. This paper describes an optimized protocol for single-cell qPCR measurements in primary pituitary cell cultures following patch-clamp recordings. Two different cell harvesting methods were assessed using both the GH4 prolactin producing cell line from rat, and primary cell culture from fish pituitaries.
Harvesting whole cells followed by cell lysis and qPCR performed satisfactory on the GH4 cell line. However, harvesting of whole cells from primary pituitary cultures regularly produced false positives, probably due to RNA leakage from cells ruptured during the dispersion of the pituitary cells. To reduce RNA contamination affecting the results, we optimized the conditions by harvesting only the cytosol through a patch pipette, subsequent to electrophysiological experiments. Two important factors proved crucial for reliable harvesting. First, silanizing the patch pipette glass prevented foreign extracellular RNA from attaching to charged residues on the glass surface. Second, substituting the commonly used perforating antibiotic amphotericin B with β-escin allowed efficient cytosol harvest without loosing the giga seal. Importantly, the two harvesting protocols revealed no difference in RNA isolation efficiency.
Depending on the cell type and preparation, validation of the harvesting technique is extremely important as contaminations may give false positives. Here we present an optimized protocol allowing secure harvesting of RNA from single cells in primary pituitary cell culture following perforated whole cell patch clamp experiments.
- Primary Cell Culture
- Pipette Solution
- Extracellular Solution
- Patch Pipette
- Patch Clamp Experiment
Using the patch clamp technique, we investigate the electrophysiological properties of single pituitary cells in culture , our main objective being to study the differences in electrical properties of follicle-stimulating hormone (FSH)- and luteinizing hormone (LH)-producing cells (gonadotropes) in teleost fish. Contrary to mammals, fish have separate cells producing FSH and LH, respectively [2–4]. Because a primary pituitary culture contains a mixture of endocrine cell types, it is imperative to discriminate between the cells investigated. Single-cell PCR may be used as a method to determine the identity of a cell in culture, and further to investigate its gene expression levels. Two harvesting techniques are commonly used in combination with single-cell PCR: harvesting of whole cells [5, 6], and harvesting of cytosol only [7, 8]. Harvesting of whole cells secures the entire cell content before lysis and PCR. However, there is a great risk of collecting extracellular solution during harvesting. Depending on cell culture type and conditions, the extracellular solution may contain RNA contaminations giving false results. To limit the risk of such contamination, cytosol may be harvested through a patch pipette using the tight giga seal whole cell patch clamp configuration. A drawback of cytosol harvesting is the potential loss of RNA if some cytosol is not harvested. Also, electrophysiological experiments are often dependent on using the perforated patch configuration, where the membrane patch is perforated by an agent added to the pipette solution, the antibiotic amphotericin B being one of the most commonly used agents . This ensures electrical contact between the pipette solution and cytosol without diffusional loss of organic compounds from the cytosol. However, the progression from perforated patch to complete rupture of the membrane patch necessary for cytosol harvesting is often difficult to perform because the giga seal is easily lost in the process. Therefore, it may be required to compromise by performing electrophysiological experiments and gene expression analysis on different cells [10, 11]. Fan and Palade  and Sarantopoulos et al. successfully tested a more hydrophilic perforating agent, a saponin derived from the horse chestnut tree, called β-escin to reduce Ca2+ current rundown in rat neurons. They concluded that β-escin also improves the giga seal formation and gives lower access resistance compared to traditional pore-forming antibiotics. In addition, Fan and Palade  reported that perforation of the membrane proceeds more rapidly using β-escin compared to nystatin, amphotericin B and gramicidin. Thus, β-escin could potentially improve the efficiency when harvesting cytosol for single-cell PCR. In this paper, we present a novel protocol for single-cell quantitative (q) PCR based on harvesting of cytosol through the patch pipette. The present study describes a protocol that reduces the risk of RNA contamination producing false positives, and at the same time facilitates the difficult transition from a perforated patch to a complete hole in the membrane necessary for cytosol harvesting.
GH4 cell line
The GH cell lines were originally derived from a rat pituitary tumor . We have used the GH4C1 subclone, which predominantly secretes prolactin (PRL) . Cells were grown in Ham's F-10 medium (Sigma, MO, USA) supplemented with NaHCO3 (14.3 mM), penicillin (100 U/ml), streptomycin (100 μg/ml; both from Invitrogen, CA, USA), horse serum (7.5%), and calf serum (2.5%; both from Med Probe, Oslo, Norway), pH 7.4. Cells were grown in 25 cm2 tissue culture flasks and incubated at 37°C in a humidified atmosphere with 5% CO2 and 95% air. Once per week, the cultures were split by trypsin EDTA (200 μg/ml; Lonza, Belgium). Prior to the experiments, the cells were plated in 35 mm plastic petri dishes with a cell density of 20 000- 50 000 cells/cm2.
Cod pituitary primary cell culture
For making primary cell culture  we used maturing male and female Atlantic cod (1-3 kg) captured in the Oslo fjord and kept in the aquarium facilities at the University of Oslo. The holding tanks were continuously perfused with seawater with salinity of 28 ppt and a temperature of 7-8°C, and the fish were fed shrimps while in captivity. Immediately after decapitation, pituitaries (3-4 per culture) were transferred to ice-cold L-15 medium (Invitrogen) with penicillin (25 U/ml) and streptomycin (25 μg/ml; both from Invitrogen), 1.8 mM glucose, and pH adjusted to 7.8. Subsequently, pituitaries were washed in ice-cold phosphate-buffered saline (PBS; Invitrogen) with penicillin (25 U/ml) and streptomycin (25 μg/ml), at pH 7.8, chopped in approximately 1 mm3 pieces with a scalpel and washed again. Tissue fragments were treated with trypsin (type II-S, 1 mg/ml; Sigma) in modified PBS for 45 min in a shaker water bath at 18°C. The tissue was incubated in the water bath for another 20 min with trypsin inhibitor (type I-S, 1 mg/ml; Sigma) in modified PBS. Tissue fragments were mechanically dissociated in ice-cold PBS using a plastic pipette. The cell suspension was filtered through a nylon mesh and centrifuged at 100 g at 4°C for 10 min before the cells were resuspended in growth medium (L-15 with 10 mM NaHCO3, 1.8 mM glucose, penicillin (25 U/ml), and streptomycin (25 μg/ml), pH 7.8). All solutions were adjusted to 320 mOsm using NaCl. The cells were plated out in 35 mm plastic petri dishes coated with poly-L-lysine (Sigma) with cell densities between 150 000 and 200 000 cells/cm2. The total yield from 3-4 pituitaries was 10-12 dishes. The dishes were kept at 12°C in a humidified atmosphere, with 0.5% CO2 and 99.5% air. The growth medium was changed after 24 hours to remove damaged and detached cells.
All chemicals were purchased from Sigma Aldrich and certified ultrapure or RNase-free. During the experiments, the growth medium was substituted with standard RNase-free extracellular solution (mM): 150 NaCl, 5 KCl, 2.4 CaCl2, 1.3 MgCl2, 10 (GH4C1 culture) or 1.8 (cod pituitary primary culture) glucose, 0.1% bovine serum albumin (BSA), 10 HEPES/NaOH, pH 7.4 (GH4C1 culture) or 7.8 (cod pituitary primary culture), adjusted to 320 mOsm. The patch pipette solution consisted of (mM): 120 CH3O3SK, 20 KCl, 20 sucrose, 10 HEPES/NaOH, pH 7.2, adjusted to 300 mOsm. The pipette solution for harvesting whole cells consisted of RNase-free extracellular solution where NaCl replaced BSA and glucose.
Two perforated patch methods were compared, using either amphotericin B (Sigma)  or β-escin (Sigma)  as perforating agents. Amphotericin B (600 μg/ml) was added to the pipette solution from a stock solution (60 mg/ml DMSO) and sonicated for 1 min. Fresh stock was prepared daily, pipette solution every hour. For β-escin, we used the procedure described by Sarantopoulos et al.. In brief, β-escin (50 μM) was added to the pipette solution from a stock solution (25 mM in de-ionized nuclease-free water (Ambion)) and vortexed for 1 min. The stock was kept at -20°C for up to two weeks, the pipette solution prepared daily. The final pipette solution and stock was protected from direct light using aluminum foil.
Two harvesting techniques were compared in this work; harvesting of whole cells using a second harvesting pipette, and harvesting of cytosol using the patch pipette. Both techniques were applied following electrophysiological experiments using the patch clamp technique in the perforated patch configuration.
Harvesting of whole cells
Harvesting of single, whole cells was performed using a glass pipette with a tip diameter of 3-5 μm (1/3 to 1/2 of the cell diameter), made by a horizontal puller (P-97 Sutter Instrument CO, CA, USA). The pipette was filled with modified RNase-free extracellular solution, connected to an oil-filled cell-extractor (CellTram Oil, Eppendorf, Hamburg, Germany), and controlled by a hydraulic micromanipulator. During harvesting, the cells were observed using an inverted phase-contrast microscope. To avoid aspiration of possibly contaminated extracellular solution, the pipette was initially under positive pressure, created manually by regulating the cell-extractor. When the pipette was placed close to the cell to be harvested, the pressure was released and the cell trapped inside the pipette by applying moderate suction with the cell-extractor. The tip of the pipette containing the cell was passed several times through the air-fluid interface of the extracellular solution to remove potential debris adhering to the glass. The cell was then eluted into an RNase-free 1.5 ml tube (Ambion) with 10 μl lysis solution (Invitrogen) by applying a slight positive pressure. cDNA synthesis followed directly after harvesting, using the lysed cell solution as template.
Harvesting of cytosol using the patch pipette
As with the cell harvesting pipettes, the patch pipettes were made using the P-97 horizontal puller, with a tip diameter of approx 1 μm. The patch pipettes were silanized by briefly dipping the tip in Sigmacote (Sigma), a chlorinated organopolysiloxane, diluted in heptane, before being fire polished. The patch electrodes were filled with 3-4 μl of either amphotericin B- or β-escin-containing RNase-free intracellular solution. The pipette resistance was between 3.5 and 5.5 MΩ.
As the electrophysiological recordings were performed using perforated patch, a complete hole had to be made before harvesting of cytosol. A standard 20 ml syringe was used to generate the sub-atmospheric pressure needed to make the hole without breaking the giga seal. The syringe was also used for harvesting cytosol, by applying suction. In the initial experiments, the harvesting was terminated immediately after the giga seal broke. However, after the conditions had been optimized, harvesting was finalized only when the shrinkage of the cell halted, usually after about 3 minutes of harvesting. The pipette was then passed several times through the air-fluid interface of the extracellular solution before the content was transferred to an RNase-free 0.5 ml tube (Ambion) containing 7 μl of a sodium citrate EDTA- and EGTA- free solution (Ambion). This RNA storage solution also prevents potential inhibition of reverse transcription. To inactivate RNA-degrading RNases, we used 1 μl (40U/μl) RNasin Plus RNase Inhibitor (Promega, WI, USA). Positive pressure was applied while only the first few micrometers of the patch pipette was in contact with the solution in the tube, allowing the RNA situated in the first part to be ejected. While still maintaining positive pressure the patch pipette was then gently crushed against the bottom and the rest of the (RNA) solution in the patch pipette was emptied into the tube. For details see Sakmann and Neher .
In order to optimize the qPCR assay for identification of rat PRL mRNA in the GH4 cell line, 5000 cells per μl and a total of 10 000 cells were collected in 1.5 ml tubes and centrifuged at 10 000 g for 1 min, following rapid freezing in liquid nitrogen and stored at -80°C until further use. Cells were lysed in a total volume of 10 μl using CellsDirect's (Invitrogen) lysis enhancer (1 μl) and resuspension buffer (10 μl), before being incubated at 75°C for 10 min, and immediately placed on ice until cDNA synthesis. The same method was used also for lysis of single cells. cDNA synthesis was carried out directly after cell lysis.
To optimize the qPCR assays for cod FSHβ, LHβ, and elongation factor 1α (EF1α) RNA extracted from intact cod pituitaries were used. The pituitaries were collected immediately after decapitation and stored on RNAlater (Ambion) at -20°C until further processing. Two or three pituitaries were homogenized in 1 ml TRIZOL Reagent (Invitrogen), using a Fast Prep FP 120 machine (QBiogene, CA, USA). The extraction was performed using the manufactures protocol. The extracted total RNA was dissolved in water or RNA storage solution (Ambion) and stored at -80°C until cDNA synthesis.
The quality and quantity of the total RNA was evaluated using a NanoDrop spectrophotometer (NanoDrop Technologies, DE, USA). Only pure RNA with OD 260/280 between 1.8 and 2 was accepted. Except for RNA stored on RNA Storage Solution (Ambion), we also regularly checked that OD 260/230 was between 2 and 2.2.
Reverse transcription (RT)
For first strand cDNA synthesis, 1 μl random hexamer primers (50 ng/μl) and 1 μl 10 mM dNTP mix (both from Invitrogen) were added to either the 10 μl of GH4 cell lysis solution from 10 000 cells and the single harvested cell solution, or to 3 μg of cod pituitary total RNA, or to the 11 μl of solution containing the harvested cytosol. The volumes were always adjusted to 13 μl using RNase free water (Ambion). The mixture was incubated for 5 min at 65°C, then 3 min incubation on ice. Subsequently, 4 μl 5X first strand buffer, 1 μl 0.1 M dithiothreitol (DTT), 1 μl RNase-out (40 U/μl), and 1 μl SuperScript III reverse transcriptase (200 U/μl; all from Invitrogen) were added, and the mixture was briefly centrifuged before incubation at room temperature for 5 min. The mixture was then incubated for 60 min at 50°C, before the enzyme reaction was inactivated by heating for 15 min at 70°C. The cDNA was stored at -20°C until qPCR. The same protocol was performed for both harvesting techniques, except that RNase-out was omitted for cytosol harvesting. After initial testing, the cDNA was also precipitated for removal of inhibitory factors directly following cDNA synthesis as described by Liss .
The qPCR primers where targeted to exon-exon borders, or alternatively, in the case of rat PRL where the intron was much longer than the amplicon, by placing the two primers on two different exons.
qPCR primers used in this study. Fw, forward; Rv, reverse.
5'-CAT CAA TGA CTG CCC CAC TTC-3'
5'-AGC CGC TTG TTT TGT TCC TCA-3'
5'-GAA CCG AGT CCA TCA ACA CC-3'
5'-GGT CCA TCG GGT CCT CCT-3'
5`-GTG GAG AAG AAG GGC TGT CC-3
5`-GGA CGG GTC CAT GGT G-3`
5'-CCT TCA ACG CCC AGG TCA T-3'
5'-AAC TTG CAG GCG ATG TGA G-3'
Calibration curve for cod FSHβ and LHβ primer pairs
To assess the sensitivity of the primer pairs specific for cod FSHβ and LHβ, a qPCR calibration curve with defined numbers of molecules was made. Conventional PCR with primers defining the whole coding region was used , followed by gel electrophoresis. The PCR products were purified (Wizard PCR preps DNA purification system; Promega) and the concentrations were measured using a NanoDrop spectrophotometer. The number of molecules was calculated using the procedure described by Sambrook et al.. Series of dilutions were made with 10- or 2- fold increments, ranging from 1-2 up to 105 molecules and subjected to qPCR analyses (see below). Detection limit was based on Cq values and on the quality of the melting curve for each qPCR reaction.
We developed and validated qPCR assays using SYBR green I detection dye on a LC480 platform (both from Roche, Basel, Switzerland). The final reaction volume of 10 μl contained 1 μl of 5 μM- forward and reverse primer, 5 μl of SYBR green I master mix and 3 μl of cDNA. In the non-template control (NTC), cDNA was substituted with 3 μl of nuclease free water. To evaluate possible genomic DNA interference, the primers for rat PRL were tested on 3 μl of undiluted cell lysate from 10 000 cells without the RT-step. For cod FSHβ, LHβ, and EF1α, we used 3 μl of 1 μM genomic DNA extracted with DNeasy Tissue kit (Qiagen, Hilden, Germany) from cod pituitary. All qPCR samples were run in duplicate, and for each plate we applied a NTC and positive control for each primer pair to control the experimental setup.
The qPCR reactions with primer pairs for rat PRL was carried out using an initial step of 10 min at 95°C to activate Taq polymerase, followed by 42-45 cycles consisting of 10 s at 95°C, 10 s at 59°C, and elongation at 72°C for 10 s. The fluorescence was measured at the end of each cycle. Crossing point (Cq) values were calculated using the "Second Derivative Maximum Method" . To find the optimal qPCR conditions, a series of cDNA dilutions were made, and the Cq values were then plotted against the relative cDNA concentration. The qPCR efficiency was calculated using the slope of the regression line, according to the equation E = 10[-1/slope]. A slope of -3.32 gives E = 2, or a doubling of product for each PCR cycle. A melting curve analysis was performed directly following the PCR by continuously reading the fluorescence while slowly heating the reaction from 65°C to 98°C.
Due to higher melting temperature and shorter amplification sequences for the cod FSHβ-, LHβ- and EF1α primer pairs compared to the rat PRL primers, the annealing temperature was increased to 60°C and the elongation time was reduced to 5 s in the assays involving cod cDNA. The rest of the PCR protocol and melting curve analysis was the same as for rat PRL.
To confirm amplicon size, qPCR products were analyzed by electrophoresis using a 1% agarose gel for rat PRL and a 2% agarose gel for cod FSHβ, LHβ, and EF1α. All gels were stained with 0.4 μg/ml EtBr. Before loading PCR products on the gel, 1 μl 10 × BlueJuice (Invitrogen) and 3 μl of nuclease free water (Ambion) were added to 2-3 μl of PCR product. Electrophoresis was conducted using an electrical field of 5 V/cm for 40-50 min.
All qPCR products were regularly sequenced to assure specificity of the primers and to confirm that the different melting curve analyses were true positives. Unspecific, short nucleotide fragments and dNTPs were degraded from the qPCR reaction solution using exoSAP-IT (USB, OH, USA). Two μl qPCR solution, 2 μl of exoSAP-IT, and 3 μl water were mixed and incubated for 30 min at 37°C, and then inactivated for 15 min at 80°C. Forward primers (5 μM each) were added before the cDNA was sequenced at the sequencing platform (ABI 3730 high-throughput capillary electrophoresis sequencers) at the Department of Molecular Biosciences, University of Oslo.
All statistics were calculated using GraphPad Prism version 4.0 or InStat version 3.0b.
The different experimental groups were first tested for equal variance followed either by unpaired t test with Welch's correction or for multiple group comparison Kruskal-Wallis test. Post test was performed using Dunn's Multiple Comparison Test. Numbers are given as mean ± SD if not otherwise stated. The levels of significance are P < 0.05* P < 0.01**, P < 0.001 *** and P > 0.05 not significant (ns).
In the following experiments we compared harvesting of whole cells using a harvesting pipette, with harvesting of cytosol using a patch pipette. We started with optimizing the conditions for harvesting whole cells on the GH4 cell line. For harvesting whole cells, the cell lysis solution CellsDirect contained all necessary compounds for exposing the RNA. Prior to cell harvesting, the Petri dishes containing the GH4 cells were thoroughly washed with extracellular solution. Control experiments where small amounts of extracellular solution were harvested confirmed the absence of RNA contamination that might affect the results.
The dilution curve resulting from using 10 000 GH4 cells (Figure 4) showed that 3 μl of cDNA diluted 1:20 gives a Cq value of about 24. Theoretically, an increase of 3 in the Cq value corresponds to approximately 10 × decrease in cDNA concentration, while an increase of 6 corresponds to a 100 × decrease in concentration. Hence, compared to the 1:20 dilution from 10 000 cells, 3 μl of undiluted cDNA from one cell would be expected to increase the Cq value by 7-8, to between 31 and 32.
The technique for harvesting whole cells was subsequently evaluated on the primary cod pituitary culture. However, in addition to EF1α, mRNA for both FSHβ and LHβ were detected in all collected cells, contrary to the assumption that in fish, one cell synthesizes only one hormone. Control experiments by collecting 0.2-0.4 μl extracellular solution from multiple dishes, followed by qPCR, revealed extensive mRNA contamination in the bath with detection of all 3 genes.
The result of this extracellular contamination lead to a protocol for harvesting cytosol through the patch pipette. The tight seal made between the glass and cell membrane in connection with a patch clamp experiment is in the giga ohm range, leaving the inside of the patch pipette and its content unexposed to the possibly contaminated extracellular bath. For that reason, a stable giga seal is essential for both successful electrophysiological registration and for cytosol harvesting.
Finally, we also compared the two harvesting techniques, harvesting of whole cells and harvesting of cytosol, on the GH4 cell line (as this didn't produce false positives when harvesting whole cells), using rat PRL Cq values as an indicator of efficiency (Figure 7). Interestingly, no significant difference in Cq was detected between the two harvesting techniques. Mean Cq was 31.23 ± 0.88 (n = 9) for harvesting whole cells versus 30.87 ± 1.29 (n = 11) for cytosol harvesting, with mean difference of 0.35 ± 0.50. Also, there was no significant difference in variance between the two groups (F-test, P = 0.14).
The highly robust and sensitive method of qPCR has facilitated gene expression studies and has given new insights into single cell gene expression. The lognormal variation of copy number within single cells in a population  reflects the stochastic nature of how a cell continuously turns on and off its different genes [25–27]. In addition, increased demands for new sensitive and reliable techniques in clinical investigations using limited amount of sample [28–30] makes single-cell qPCR an attractive strategy. Therefore, it is essential to have robust and reliable protocols that discriminate between true negative and true positive results. The numerous protocols for securing single cells or cell content for gene expression studies provide possibilities for having different cell or tissue preparations [6–8, 18, 31–33]. The literature also thoroughly documents important optimization steps from cell lysis, to reverse transcription and qPCR [5, 18, 23, 24, 34–38].
However, we found that the literature provides rather limited information dealing with often unavoidable contaminations that may result in false positives . A common solution to avoid false positives is to reduce the number of PCR cycles , but in our opinion, such a strategy may affect the amplification efficiency of the gene of interest. As such, it treats the symptoms rather than the root of the problem. Furthermore, the number of harvesting techniques suitable for securing the RNA is limited following electrophysiological experiments; harvesting of whole cells [23, 24], or harvesting of only cytosol through the patch pipette [7, 8]. Because perforating agents commonly used in patch clamp studies, often disrupt the giga seal during the necessary transition from a perforated membrane to a complete whole in the cell membrane, the latter method often excludes the possibility of combining perforated whole cell recordings with single-cell qPCR on the same cell.
In the present study, we have established a reliable method for gene expression analysis of single cells in primary culture following patch clamp experiments, focusing on FSHβ- and LHβ- producing cells from Atlantic cod. We evaluated the two harvesting strategies suitable in combination with the patch clamp technique; harvesting of whole cells and harvesting of cytosol through the patch pipette.
The novelty of our single-cell qPCR protocol is the combined use of the perforating agent β-escin and glass-pipette silanization allowing single-cell qPCR and perforated patch clamp experiments on the same cell. Together, this improves the harvesting efficiency and reduces the incidence of false positives.
The GH4 cell line is easily maintained and only needs gentle trypsination between cell passages. The cells also firmly attach to the dish surface even without coating. This is in contrast to the primary cell culture from fish pituitary, which needs both longer trypsination and mechanical dissociation for proper dispersal. Although we should expect variation in gene expression also in the homogenous GH4 cell line, this variation is less than what we find in a primary cell culture. The GH4 cell line is thus better for detecting and analyzing poor assay design and interassay variation. The initial work on the GH4 cell line proved valuable as it allowed analyzing and validating in detail the harvesting step and also the different reagents used. In addition, it allowed comparing the impact that different culture conditions have on the assay performance.
Following harvesting of the whole cell, the cell was transferred to a cell lysis solution to access the RNA. CellsDirect (Invitrogen) has been successfully used for reliable qPCR down to single cells in earlier studies [34, 38], and was therefore chosen as a basis for cell lysis. Cell lysis and cDNA synthesis in the present study turned out to be readily transferable from 10 000 cells down to a single cell and no de-escalation of additives or additional steps was needed.
After cell lysis we reverse-transcribed the RNA to cDNA. The priming strategies of the reverse transcription between using oligo (dT), random primers, specific primers or combinations of these seem to be assay- and transcript-dependent [37, 40–42]. Also, a newly developed method for single-cell gene expression analysis uses oligo (dT) immobilized beads to obtain mRNA . However, one important drawback of using oligo (dT) is that mRNA with degraded poly(A) tail will not be reverse transcribed. Instead, random primers can potentially bind to all regions of an RNA fragment increasing the chance of converting RNA fragments complimentary to the designed PCR primers. Interestingly, combining random primers and oligo (dT) have been successfully used for single-cell analysis [5, 37]. In future studies the use of both random primers and oligo (dT) will be further evaluated, especially for low abundant transcripts like receptors and ion channels.
As with priming strategies, great variation of the enzyme performance of the reverse transcriptase on different genes is observed . Therefore, for quantitative measurements the same enzyme should be used between every sample. In our laboratory we have tested different types of reverse transcription enzymes and support earlier results  that SuperScript III reverse transcriptase gives high efficiency even when having low transcript levels.
All primers were tested and confirmed insensitive towards genomic DNA. We could therefore omit the extra step of DNase treatment. The primers were further evaluated using a cDNA dilution curve. For FSHβ and LHβ, it proved crucial to have relatively short amplicons for an optimal assay with high specificity and efficiency, but also it allowed us to reduce the elongation time of the PCR thereby reducing chances of amplifying genomic DNA . We also observed that small variations in primer annealing temperature greatly affected the qPCR assay efficiency. In general, increased PCR efficiency and sensitivity are obtained by reducing the primer annealing temperature. However, too low annealing temperature can potentially lead to unspecific binding of primers and create unspecific or multiple peaks in the melting curve, and thus detection of false positives.
Several factors in the reverse transcription mixture, including reverse transcriptase, inflict with the PCR and inhibit the polymerase enzyme [22, 23]. To overcome the inhibitory effects, two strategies have been tested and validated down to single cells: either to use as low concentration of the reverse transcription agents as possible , or to remove the inhibiting reagents using ethanol precipitation adapted to low transcript levels . The present study substantiates the importance of ethanol precipitation for reliable gene detection in single cells, significantly stabilizing and reducing the mean Cq values (Figure 7). Furthermore, ethanol precipitation of cDNA may help to remove any distorting salt compounds found in the RNA storage solution and pipette solution.
During the optimization of the different harvesting strategies, we revealed some interesting differences between the initial GH4 cell line test model and the primary dispersed cell culture from fish pituitary. The main difference being the prolonged trypsination and mechanical dissociation needed to disperse the cells from fish pituitaries. This treatment thus leads to increased contamination of RNA in the bath. In addition, we had to minimize washing of the Petri dishes with extracellular solution because the primary pituitary cells detached, making it impossible to conduct electrophysiological experiments. Because some extracellular fluid inevitably is collected together with the harvested cell, we had to exclude harvesting of whole cells as a method for obtaining RNA.
Surprisingly, we still experienced false positives after changing to cytosol harvesting through the patch pipette. The most probable cause is the ability RNA has to attach to charged residues on the glass surface. Based on these assumptions we therefore silanized the glass using Sigmacote. This proved to be crucial for avoiding possible interference from unwanted RNA. Moreover, we could dilute the Sigmacote still excluding RNA from attaching, but significantly improving the giga seal formation. The improved seal frequency and tightness after diluting Sigmacote has also been described by Przysiezniak and Spencer  using neurons from a hydrozoan jellyfish.
One very important implication concerning cytosol harvesting through the patch pipette is the difficulties of maintaining a stable giga seal during harvesting. In most cases we experienced seal breakage after a few tenths of a second into the harvesting when using amphotericin B as the perforating agent. To overcome this limitation we tested a fairly new agent in the context of electrophysiology [12, 13] named β-escin. This agent has been validated for use as a pore forming agent in patch clamp experiments but never for cytosol harvesting. In agreement with earlier studies [12, 13] we found β-escin to easily perforate the cell membrane giving low access resistance (< 15 MΩ). Furthermore, β-escin stabilized the giga seal facilitating the difficult transition from a perforated patch to a complete hole in the membrane followed by harvesting of cytosol. Using EF1α, a more stably expressed gene compared to FSHβ and LHβ allowed us to evaluate harvesting efficiency (Figure 10). The significant difference in variance between using amphotericin B and β-escin we believe reflects the unstable giga seal using amphotericin B. This instability in most cases leads to a reduced harvesting because the harvesting procedure has to be aborted as soon as the giga seal breaks to prevent extracellular contamination.
For quantitative gene expression studies using single cells it is of course plausible to favor harvesting of the whole cell, as this in most cases assures that all the RNA is collected. However, when comparing the two harvesting methods for obtaining RNA on the GH4 cell line, we found no significant difference in Cq values (Figure 7). This shows that equal amounts of RNA were obtained when harvesting only cytosol compared to harvesting the whole cell.
In conclusion, we have developed a single cell qPCR assay for reliable detection of FSHβ and LHβ from primary dispersed fish pituitary cells. The use of silanized patch pipettes for cytosol harvesting reduces the risk of contaminating the samples with RNA from the extracellular solution. In addition, using β-escin as the perforating agent in perforated patch clamp experiments allows us to progress from a perforated patch to the traditional whole cell configuration followed by cytosol harvesting without losing the giga seal, which is necessary for reliable harvesting.
We thank Olav Sand for valuable discussions and advice, Sylvie Dufour for help with establishing the primary cultures, Birgitta Norberg and Christian Mittelholzer for providing at the time unpublished sequences for FSHβ, LHβ and EF1α from cod, and Kirsten Ore, Jonathan Coleman, Hans Erik Karlsen and NIVA for technical assistance. This study was supported by the Research Council of Norway, grants no. 184851 (to F-A.W.) and 191825 (to T.M.H.).
- Haug TM, Hodne K, Weltzien FA, Sand O: Electrophysiological properties of pituitary cells in primary culture from Atlantic cod (Gadus morhua). Neuroendocrinology. 2007, 86: 38-47. 10.1159/000103867View ArticlePubMedGoogle Scholar
- Nozaki M, Naito N, Swanson P, Miyata K, Nakai Y, Oota Y, Suzuki K, Kawauchi H: Salmonid pituitary gonadotrophs. I. Distinct cellular distributions of two gonadotropins, GTH I and GTH II. Gen Comp Endocrinol. 1990, 77: 348-357. 10.1016/0016-6480(90)90224-AView ArticlePubMedGoogle Scholar
- Weltzien FA, Kobayashi T, Andersson E, Norberg B, Andersen O: Molecular characterization and expression of FSHβ, LHβ, and common alpha-subunit in male Atlantic halibut (Hippoglossus hippoglossus). Gen Comp Endocrinol. 2003, 131: 87-96. 10.1016/S0016-6480(02)00606-8View ArticlePubMedGoogle Scholar
- Weltzien FA, Andersson E, Andersen O, Shalchian-Tabrizi K, Norberg B: The brain-pituitary-gonad axis in male teleosts, with special emphasis on flatfish (Pleuronectiformes). Comp Biochem Physiol. 2004, 137: 447-477. 10.1016/j.cbpb.2003.11.007. 10.1016/j.cbpb.2003.11.007View ArticleGoogle Scholar
- Bengtsson M, Hemberg M, Rorsman P, Ståhlberg A: Quantification of mRNA in single cells and modelling of RT-qPCR induced noise. BMC Mol Biol. 2008, 9: 63-74. 10.1186/1471-2199-9-63View ArticlePubMedPubMed CentralGoogle Scholar
- Durand G, Marandi N, Herberger S, Blum R, Konnerth A: Quantitative single-cell RT-PCR and Ca2+ imaging in brain slices. Pflugers Arch. 2006, 451: 716-726. 10.1007/s00424-005-1514-3View ArticlePubMedGoogle Scholar
- Lambolez B, Audinat E, Bochet P, Crepel F, Rossier J: AMPA receptor subunits expressed by single purkinje-cells. Neuron. 1992, 9: 247-258. 10.1016/0896-6273(92)90164-9View ArticlePubMedGoogle Scholar
- Monyer H, Lambolez B: Molecular biology and physiology at the single-cell level. Curr Opin Neurobiol. 1995, 5: 382-387. 10.1016/0959-4388(95)80052-2View ArticlePubMedGoogle Scholar
- Rae J, Cooper K, Gates P, Watsky M: Low access resistance perforated patch recordings using amphotericin B. J Neurosci Methods. 1991, 37: 15-26. 10.1016/0165-0270(91)90017-TView ArticlePubMedGoogle Scholar
- Wu J, DeChon J, Xue F, Li G, Ellsworth K, Gao M, Liu Q, Yang K, Zheng C, He P: GABA(A) receptor-mediated excitation in dissociated neurons from human hypothalamic hamartomas. Exp Neurol. 2008, 213: 397-404. 10.1016/j.expneurol.2008.07.004View ArticlePubMedPubMed CentralGoogle Scholar
- Yin H, Lee KE, Park SA, Bhattarai JP, Suh BJ, Jeon JG, Kim BG, Park SJ, Han SK: Inhibitory effects of somatostatin on the substantia gelatinosa neurons of trigeminal subnucleus caudalis via somatostatin type 2 receptors in juvenile mice. Brain Res. 2009, 1304: 49-56. 10.1016/j.brainres.2009.09.070View ArticlePubMedGoogle Scholar
- Fan FS, Palade P: Perforated patch recording with β-escin. Pflugers Arch. 1998, 436: 1021-1023. 10.1007/PL00008086View ArticlePubMedGoogle Scholar
- Sarantopoulos C, McCallum JB, Kwok WM, Hogan Q: β-Escin diminishes voltage-gated calcium current rundown in perforated patch-clamp recordings from rat primary afferent neurons. J Neurosci Methods. 2004, 139: 61-68. 10.1016/j.jneumeth.2004.04.015View ArticlePubMedGoogle Scholar
- Tashjian AH, Yasumura Y, Levine L, Sato GH, Parker ML: Establishment of clonal strains of rat pituitary tumor cells that secrete growth hormone. Endocrinology. 1968, 82: 342-352. 10.1210/endo-82-2-342View ArticlePubMedGoogle Scholar
- Kineman RD, Frawley LS: Secretory characteristics and phenotypic plasticity of growth hormone-producing and prolactin-producing cell-lines. J Endocrinol. 1994, 140: 455-463. 10.1677/joe.0.1400455View ArticlePubMedGoogle Scholar
- Montero M, Le Belle N, Vidal B, Dufour S: Primary cultures of dispersed pituitary cells from estradiol-pretreated female silver eels (Anguilla anguilla L.): Immunocytochemical characterization of gonadotropic cells and stimulation of gonadotropin release. Gen Comp Endocrinol. 1996, 104: 103-115. 10.1006/gcen.1996.0146View ArticlePubMedGoogle Scholar
- Sakmann B, Neher E: Single-channel recording. 1995, Plenum Press New YorkView ArticleGoogle Scholar
- Liss B: Improved quantitative real-time RT-PCR for expression profiling of individual cells. Nucleic Acids Res. 2002, 30: e89- 10.1093/nar/gnf088View ArticlePubMedPubMed CentralGoogle Scholar
- Mittelholzer C, Andersson E, Taranger GL, Karlsen O, Norberg B: Quantification of gonadotropin subunits GPα, FSHβ, and LHβ mRNA expression from Atlantic cod (Gadus morhua) throughout a reproductive cycle. Comp Biochem Physiol B-Biochem Mol Biol. 2009, 153: 288-295. 10.1016/j.cbpb.2009.03.011View ArticlePubMedGoogle Scholar
- Sambrook J, Fritsch EF, Maniatis T: Molecular cloning. 1989, Cold Spring Harbor Laboratory Press Cold Spring Harbor, NYGoogle Scholar
- Rasmussen R: Quantification on the LightCycler. Rapid cycle real-time PCR, methods and applications. 2001, 21-34. Springer Press, HeidelbergView ArticleGoogle Scholar
- Sellner LN, Coelen RJ, Mackenzie JS: Reverse transcriptase inhibits Taq polymerase activity. Nucleic Acids Res. 1992, 20: 1487-1490. 10.1093/nar/20.7.1487View ArticlePubMedPubMed CentralGoogle Scholar
- Chandler DP, Wagnon CA, Bolton H: Reverse transcriptase (RT) inhibition of PCR at low concentrations of template and its implications for quantitative RT-PCR. Appl Environ Microbiol. 1998, 64: 669-677.PubMedPubMed CentralGoogle Scholar
- Bengtsson M, Ståhlberg A, Rorsman P, Kubista M: Gene expression profiling in single cells from the pancreatic islets of Langerhans reveals lognormal distribution of mRNA levels. Genome Res. 2005, 15: 1388-1392. 10.1101/gr.3820805View ArticlePubMedPubMed CentralGoogle Scholar
- Spiller DG, Wood CD, Rand DA, White MRH: Measurement of single-cell dynamics. Nature. 2010, 465: 736-745. 10.1038/nature09232View ArticlePubMedGoogle Scholar
- Levsky JM, Singer RH: Gene expression and the myth of the average cell. Trends Cell Biol. 2003, 13: 4-6. 10.1016/S0962-8924(02)00002-8View ArticlePubMedGoogle Scholar
- Raj A, van Oudenaarden A: Nature, nurture, or chance: stochastic gene expression and its consequences. Cell. 2008, 135: 216-226. 10.1016/j.cell.2008.09.050View ArticlePubMedPubMed CentralGoogle Scholar
- Pantel K, Brakenhoff R, Brandt B: Detection, clinical relevance and specific biological properties of disseminating tumour cells. Nature Reviews Cancer. 2008, 8: 329-340. 10.1038/nrc2375View ArticlePubMedGoogle Scholar
- Ginsberg SD, Elarova I, Ruben M, Tan FZ, Counts SE, Eberwine JH, Trojanowski JQ, Hemby SE, Mufson EJ, Che SL: Single-cell gene expression analysis: Implications for neurodegenerative and neuropsychiatric disorders. Neurochem Res. 2004, 29: 1053-1064. 10.1023/B:NERE.0000023593.77052.f7View ArticlePubMedGoogle Scholar
- Lo Y, Hjelm N, Fidler C, Sargent I, Murphy M, Chamberlain P, Poon P, Redman C, Wainscoat J: Prenatal diagnosis of fetal RhD status by molecular analysis of maternal plasma. The New England journal of medicine. 1998, 339: 1734- 10.1056/NEJM199812103392402View ArticlePubMedGoogle Scholar
- Micke P, Ostman A, Lundeberg J, Ponten F: Laser-assisted cell microdissection using the PALM system. Methods Mol Biol. 2005, 293: 151-166.PubMedGoogle Scholar
- Kuhn DE, Roy S, Radtke J, Gupta S, Sen CK: Laser microdissection and pressure-catapulting technique to study gene expression in the reoxygenated myocardium. Am J Physiol-Heart Circul Physiol. 2006, 290: H2625-H2632. 10.1152/ajpheart.01346.2005. 10.1152/ajpheart.01346.2005View ArticleGoogle Scholar
- EmmertBuck MR, Bonner RF, Smith PD, Chuaqui RF, Zhuang ZP, Goldstein SR, Weiss RA, Liotta LA: Laser capture microdissection. Science. 1996, 274: 998-1001. 10.1126/science.274.5289.998View ArticleGoogle Scholar
- Nolan T, Hands RE, Bustin SA: Quantification of mRNA using real-time RT-PCR. Nat Methods. 2006, 1: 1559-1582. 10.1038/nprot.2006.236. 10.1038/nprot.2006.236Google Scholar
- Stahlberg A, Bengtsson M: Single-cell gene expression profiling using reverse transcription quantitative real-time PCR. Methods. 2010, 50: 282-288. 10.1016/j.ymeth.2010.01.002View ArticlePubMedGoogle Scholar
- Stahlberg A, Kubista M, Pfaffl M: Comparison of reverse transcriptases in gene expression analysis. Clinical chemistry. 2004, 50: 1678- 10.1373/clinchem.2004.035469View ArticlePubMedGoogle Scholar
- Ståhlberg A, Hakansson J, Xian X, Semb H, Kubista M: Properties of the reverse transcription reaction in mRNA quantification. Clin Chem. 2004, 50: 509-515. 10.1373/clinchem.2003.026161View ArticlePubMedGoogle Scholar
- Taniguchi K, Kajiyama T, Kambara H: Quantitative analysis of gene expression in a single cell by qPCR. Nat Methods. 2009, 6: 503-506. 10.1038/nmeth.1338View ArticlePubMedGoogle Scholar
- Sucher NJ, Deitcher DL, Baro DJ, Harris Warrick RM, Guenther E: Genes and channels: patch/voltage-clamp analysis and single-cell RT-PCR. Cell Tissue Res. 2000, 302: 295-307. 10.1007/s004410000289View ArticlePubMedGoogle Scholar
- Deprez RHL, Fijnvandraat AC, Ruijter JM, Moorman AFM: Sensitivity and accuracy of quantitative real-time polymerase chain reaction using SYBR green I depends on cDNA synthesis conditions. Anal Biochem. 2002, 307: 63-69. 10.1016/S0003-2697(02)00021-0View ArticleGoogle Scholar
- Weltzien FA, Pasqualini C, Vernier P, Dufour S: A quantitative real-time RT-PCR assay for European eel tyrosine hydroxylase. Gen Comp Endocrinol. 2005, 142: 134-142. 10.1016/j.ygcen.2004.12.019View ArticlePubMedGoogle Scholar
- Zhang J, Byrne CD: Differential priming of RNA template during cDNA synthesis markedly affects both accuracy and reproducibility of quantitative competitive reverse-transcriptase PCR. Biochem J (Lond). 1999, 337: 231-241.View ArticleGoogle Scholar
- Przysiezniak J, Spencer AN: Voltage-activated calcium currents in identified neurons from a hydrozoan jellyfish, Polyorchis penicillatus. J Neurosci. 1992, 12: 2065-2078.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.