OUP user menu

Gene expression in rod shaped cardiac myocytes, sorted by flow cytometry

Claudius Diez, Andreas Simm
DOI: http://dx.doi.org/10.1016/S0008-6363(98)00189-8 530-537 First published online: 1 December 1998

Abstract

Objective: Primary cardiac myocyte cultures are usually contaminated with variable parts of different cell types, such as fibroblasts, endothelial cells and smooth muscle cells. Thus, the objective of our study was to analyse the gene expression in a pure population. Methods: To obtain an homogeneous population, cardiac myocytes from adult rats were fixed with ethanol and sorted by flow cytometry. This approach is suitable for isolating either single cells or up to several thousand cells. To measure the messenger ribonucleic acid (mRNA) expression of different genes at the level of a few rod-shaped myocytes, a cDNA library was created by polymerase chain reaction (PCR). Results: Sorting by a fluorescence-activated cell sorter (FACS) resulted in pure rod-shaped cardiac myocytes and isolated RNA from these cells is undegraded, as shown by Northern blotting. We demonstrated both the expression of housekeeping genes, such as β-actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as well as the myocyte-specific transcripts, α-cardiac myosin heavy chain (α-MHC) and β-MHC. Furthermore, we showed the induction of the immediately early gene c-fos at the level of ten sorted cells. Conclusions: This method allows one to study gene expression in different cell types within the heart, in tissue samples or to tackle the problem of heterogeneity within a cell population.

Keywords
  • Cardiac myocytes
  • Cell sorting
  • Single-cell RT-PCR
  • Gene expression
  • Rat
  • Northern blotting

Time for primary review 31 days.

1 Introduction

Most primary cultures of differentiated cardiac myocytes are contaminated with variable amounts of cells, such as fibroblasts, smooth muscle cells and endothelial cells. In order to avoid these contaminations in subsequent gene expression studies, several attempts were made to analyse the pattern of gene expression in single cells. Krown et al. [15], e.g., described a patch–clamp-based technique by aspirating a fraction of the cytoplasm without contaminating genomic DNA in a pipette for further gene expression analysis. Harbeck and Rothenberg [12]isolated single cells from a diluted cell suspension using microcapillary glass pipettes under visual control using an inverted phase contrast microscope. Another approach is the use of an ultraviolet-laser microbeam to isolate single cells from formalin-fixed and paraffin-embedded tissue samples [3]. In this article, we exploited the potential of a fluorescent-activated cell sorter (FACS). Flow cytometry allows the separation of uniform cells from a heterogeneous population. Using fluorescence-tagged antibodies, one can further characterise the cells according to their expression of specific proteins. Last but not least, one may sort either single cells or a large number of cells by FACS. Different groups used this technique to measure gene expression of sorted single blood cells, such as T- or B-cells, by polymerase chain reaction (PCR) [11, 23]. Esser et al. [10]sorted a T-cell subpopulation after antibody staining and analysed gene expression by Northern blotting.

Unfortunately, cardiac myocytes reveal a severe limitation of FACS-based cell sorting due to their susceptibility to mechanical damage and rapid RNA degradation during sorting. We recently developed a method to isolate RNA from ethanol-fixed cells and we report here on our extended procedure to sort pure rod-shaped cardiac myocytes. We are able to isolate RNA from sorted myocytes, which is suitable for Northern blotting. Additionally, after creating a cDNA library, we show the expression of different mRNA transcripts at the level of a few cardiac myocytes.

2 Methods

2.1 Materials

All chemicals were purchased from Sigma and Merck, unless otherwise indicated.

2.2 Preparation of heart muscle cells and cell culture

Preparation of cardiac myocytes was done according to the protocol of Jacobson and Piper [13], with minor modifications. Briefly, two 300–400 g male Wistar rats were anaesthetised with 60 mg/kg ketamine hydrochloride (Ketavet®, Parke-Davis, Berlin, Germany), 3 mg/kg xylacine (Rompun®, Bayer, Leverkusen, Germany). We added heparin (Thrombophob®, Nordmark, Uetersen, Germany) to the anaesthetic injection to prevent intravascular and intracardial blood clotting, which would lead to failing perfusion after surgery. The hearts were excised, mounted on a Langendorff system and rinsed with perfusion buffer (PB), containing 110 mmol/l NaCl, 2.6 mmol/l KCl, 1.2 mmol/l KH2PO4, 1.2 mmol/l MgSO4, 25 mmol/l NaHCO3 and 11 mmol/l glucose at 37°C, which was continuously gassed with 95% O2/5% CO2. Perfusion was continued by recirculating PB containing 0.25 mg/ml collagenase P (Boehringer, Mannheim, Germany) and 12.5 μmol/l CaCl2 for 20 min. The ventricles of the hearts were then cut using a tissue chopper into 0.7×0.7 mm pieces and incubated for a further 10 min in 30 ml of reperfusion buffer containing 400 mg of bovine serum albumin (BSA). The material was filtered through a 250-μm nylon mesh, centrifuged at 25×g for 3 min and the resulting pellet was washed in 20 ml of Earleś solution (116.4 mM NaCl, 5.4 mM KCl, 1 mM NaH2PO4·2H2O, 5.1 mM glucose, 0.14 mM Phenol Red, 20.2 mM Tris, pH 7.5) containing 200 μmol CaCl2 and penicillin/streptomycin (100 units/ml 100 μg/ml, Seromed, Berlin, Germany). The pellet was resuspended in 10 ml of Earle's solution with 500 μM CaCl2 and penicillin/streptomycin. 5 ml of this suspension were transferred to a fresh tube with 10 ml of Earle's solution, 1 mM CaCl2, 4% BSA and penicillin/streptomycin. After centrifugation at 15×g for 1 min, the cells were dissolved in Medium 199 (Seromed, Biochrom, Berlin, Germany). Prior to FACS analysis, one volume of cells was injected into three volumes of ice-cold ethanol containing 0.1% diethyl-pyrocarbonate (DEPC). After 10 min incubation on ice, cells were collected by centrifugation and resuspended in phosphate-buffered saline (PBS) containing 0.1% DEPC.

2.3 Cell Sorting by FACS

Cells were sorted at a flow rate of 250 cells/s on an Epics Elite ESP flow cytometer (Coulter, Krefeld, Germany) using a 100-μm sort sense tip and a 488-nm Argon laser for excitation. Rod-shaped cardiac myocytes can be sorted using the following sort parameter: cell size (high forward and sidescatter), autofluorescence (excitation at 488 nm; emission at 525 nm) and, as the most important parameter, cell length (time of flight).

2.4 Northern blot analysis

For Northern blot analysis, total RNA from 300,000 sorted myocytes was extracted according to Chomczynski and Sacchi [7]. Total RNA was denatured at 65°C in a solution containing 1.2 M formaldehyde, 35% formamide and 50 ng/μl ethidium bromide. The RNA was electrophoresed in a 1.25% agarose gel containing 6.7% formaldehyde. After transfer by vacuum onto nylon sheets (Amersham hybond N) in 6×SSC (1×SSC=150 mM NaCl, 15 mM sodium citrate, pH 7.0), the RNA was crosslinked by UV irradiation (120 mJ/cm2 for 30 s). The blot was hybridized at 60°C in 5×SSC, 10×Denhardt's solution, 0.1% sodium dodecyl sulfate (SDS) and 300 μg/ml denatured herring sperm DNA. The DNA probe was labeled with [α-32P]dCTP (Hartmann Analytic, Braunschweig, Germany) using an oligolabeling kit from Pharmacia (Freiburg, Germany). The membranes were washed in solutions containing decreasing concentrations of SSC in 0.1% SDS and 2 mM EDTA. The final wash contained 0.4% SSC. Radioactivity was detected and quantified using a phosphorimager. A 402-bp rat GAPDH fragment or a 302-bp rat α-cardiac myosin heavy chain (MHC) fragment were cloned into pCR 2.1 TOPO-vector (Invitrogen, Leek, Netherlands). Reamplification with M13 forward and M13 reverse primer revealed a 652-bp and a 552-bp fragment, respectively, which was used directly as a probe for Northern blotting.

2.5 Reverse transcription-PCR

The cDNA was prepared according to Toellner et al. [23]. Briefly, three or ten, respectively, single cardiac myocytes were sorted into a 0.2-ml PCR tube containing 4 μl of first strand buffer (50 mmol/l Tris–HCl (pH 8.3), 75 mmol/l KCl, 3 mmol/l MgCl2, 10 mmol/l dithiotreitol (DTT; Life Technologies, Eggenstein, Germany), 0.5% Nonidet P40 (Pierce, Rockford, IL, USA), 300 U/ml RNAGuard® (Pharmacia, Freiburg, Germany) and 10 pmol/μl of each deoxy nucleotide triphosphate, dNTP (Boehringer) as well as 0.027 pmol/μl oligo-dT24 primer (ARK Scientific, Darmstadt, Germany). After collecting the cells by centrifugation at 14,000×g for 5 min, the tubes were kept on ice for 45 min, followed by heating at 65°C for 1 min and cooling to 20°C for 3 min. First strand synthesis was performed at 37°C for 15 min using 100 U of reverse transcriptase (SuperScript®, free of RNase H activity, Life Technologies). The enzyme was heat-inactivated at 65°C for 10 min and the samples were rapidly cooled on ice. Subsequently, 4 μl of 2× tailing buffer [40% TdT buffer (Life Technologies), 2 mmol/l dATP] and 0.4 μl of terminal transferase (Boehringer) were added. Tailing was done at 37°C for 20 min and terminated by a 10-min incubation at 65°C. The subsequent PCR with oligo-dT primer was performed in 100 μl of reaction mixture containing Taq reaction buffer (10 mmol/l Tris–HCl (pH 8.8), 50 mmol/l KCl, 2.5 mmol/l MgCl2), 10 U of Taq-polymerase (AGS, Heidelberg, Germany), 0.1 pmol/μl of odT24 primer and 1 mmol/l of each dNTP. PCR conditions were as follows: 1 min at 94°C, 2 min at 42°C and 6 min plus 10 s additional extension per cycle (Gene Amp System 2400, Perkin-Elmer). After 25 cycles and a further addition of 5 U of Taq-polymerase, 25 further cycles were performed.

A 5-μl volume of this solution served as a template for the sequence-specific second PCR. The 50 μl final reaction volume contained 200 μmol/l of each dNTP, 5 U of Taq-polymerase and 5 pmol of cDNA-specific sense and antisense primers (Table 1). The amplification included the following steps: 1 min at 94°C, 1 min annealing (for temperatures, see Table 1) and 1.5 min at 72°C without additional extension. A reaction mixture containing water instead of template served as the negative control. To confirm the identity, all products were sequenced by the fluorescent di-deoxy terminator method of cycle sequencing on a Perkin-Elmer, Applied Biosystems Division, 373A automated DNA sequencer.

View this table:
Table 1

Characteristics of primer

SpecificitySequencesaTm (°C)Product length (bp)PositionbAccessionReference
β-ActinTGC TGA TCC ACA TCT GCT GGA562273079–2729J00691c
GAA GTG TGA CGT TGA CAT CCG
α-MHCCTG CTG GAC AGG TTA TTC CTC A653045866–5562X15938Robbins et al.
GGA AGA GCG AGC GGC GCA TCA AGG(1990) [19]
β-MHCTTC AAA GGC TCC AGG TCT CAG GGC652025877–5675Xl 5939Robbins et al.
GCC AAC ACC AAC CTG TCC AAG TTC(1990) [19]
GAPDHTCC ACC ACC CTG TTG CTG TAG C584011001–600M17701
TGG AAA GCT GTG GCG TGA TG
c-fosAAG GAA GAC GTG TAA GCA GTG CAG C585841156–572X06769Wong et al.
GAA TAA GAT GGC TGC AGC CAA GTG C(1994) [26]
Cyclin ACAC TCA CAC ACT TAG TGT CTC TGG TGG G5912771393–116Z26580Kang and Koh
GCC GCG ATG CCG GGC ACC TCG AGG CAT TCG(1997) [14]
cdk 2TTG CGA TAA CAG GCT CCG TC56233962–729D28753Moore et al.
CAC AGC CGT GGA TAT CTG GAG(1996) [16]
cdk 4ACG CCT GTG GTG GTT ACG CT56279910–631L11007Moore et al.
CCA TCT CTG GCA CCA CTG AC(1996) [16]
  • a The upper sequence indicates the antisense primer.

  • b The binding sites listed below point to the corresponding mRNA.

  • c The numbers indicate the primer binding positions on the corresponding cDNA.

2.6 Immunohistochemical staining

We added 200 μl of myomesin monoclonal antibody (cell culture supernatant, generous gift from Prof. Eppenberger, Zürich) to 500 μl of sorted myocytes in 3% BSA/PBS and incubated the mixture for 45 min on ice. The myocytes were centrifuged at 1500×g and resuspended in 500 μl of 3% BSA/PBS. The resuspension buffer contained a Cy3+-labelled anti-mouse IgG antibody (1:200, v/v) from Dianova (Hamburg, Germany). After 30 min incubation on ice, we recentrifuged the cells and resuspended them in 500 μl of 3% BSA/PBS. Using a CytoSpin™ centrifuge (Shandon, Frankfurt, Germany), we fixed the myocytes on glass slides. Before embedding (ProLong™ Antifade Kit, Molecular Probes, Eugene, OR, USA), we washed the cells twice with PBS.

The investigation conforms with The Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1985).

3 Results

As cardiac myocytes are very susceptible to mechanical damage during passage through a cell sorter (data not shown), we used ethanol-fixed cells. To avoid degradation of the RNA, we added 0.1% DEPC to all solutions. In control experiments, we fixed murine AKR-2B fibroblasts with 70% ethanol, sorted 1,000,000 cells and prepared the RNA by the method of Chomczynski and Sacchi [7]. The isolated RNA was undegraded, as shown by Northern blotting, with a β-actin probe (Bertsch et al., submitted).

We now extended this method to isolate pure rod-shaped cardiac myocytes. The analysis by FACS is shown in Fig. 1. Freshly prepared cardiac myocytes in suspension were fixed with ethanol for 10 min and, after centrifugation, resuspended in DEPC-treated PBS. A subset of these cells supercontracted during isolation and fixation, resulting in shorter cells with an increased diameter. In order to eliminate this population and other contaminating cells, e.g., fibroblasts, the following parameters were used to reach the highest purity rather than high recovery:

Fig. 1

Sort parameters for sorting cardiac myocytes. A preparation of cardiac myocytes was fixed for 10 min in 70% ethanol, centrifuged, resuspended in PBS and analysed in a Coulter Elite ESP cell sorter. Upper panel: Cell size versus granularity, i.e., forwardscatter (log) versus sidescatter (log). Lower left panel: autofluorescence at 525 nm, excitation at 488 nm (log). Lower right panel: Cell length, time of flight (linear).

Upper panel, cell size: The dot plot diagram represents the physical scatter parameter of cells, i.e., forward scatter (cell size) on the X-axis versus sidescatter (granularity) on the Y-axis. The gated population represents the myocytes.

Lower left panel, autofluorescence: As large cells, such as myocytes, have greater autofluorescence values in comparison to small cells, such as fibroblasts, this parameter, i.e. autofluorescence at 525 nm after excitation at 488 nm, was used to obtain a second independent measurement for cell size. The marker, M1, shows the large cells, i.e., myocytes.

Lower right panel, length: Intact myocytes can be characterised by their rod shape (cell length). Using flow cytometry, the length of cells is measured by the parameter time of flight, i.e., the time that a particle needs to cross the laser beam. The marker, M2, represents long cells.

The population after cell sorting using the combined parameters as sort gates (cell size, autofluorescence and length), as described above, is presented in Fig. 2. Panel 2A represents a crude ethanol-fixed cell population, which was used for cell sorting. A mixture of rod-shaped and supercontracted myocytes, cell debris and contaminating small cells can be seen. The resulting cell population after cell sorting is shown in panel B. The myocyte population has a purity of more than 99%. In fact, we did not detect small fibroblast-like cells in our sorted population. This can be substantiated using a DNA-specific stain, Hoechst 33258. Panel 2C shows an ethanol-fixed sorted myocyte population. The same population is shown in panel D using the DNA stain Hoechst 33258. Indeed, all stained nuclei are associated with myocytes. According to published data, nearly all myocytes from adult rats have two nuclei.

Fig. 2

Sorting of rod-shaped cardiac myocytes by flow cytometry; panel A: Fixed cell population comprising myocytes and non-myocytes, such as fibroblasts and debris, which was used for subsequent sorting; panel B: Sorted cardiac myocytes representing a pure cell population; panel C: Sorted cardiac myocytes; panel D: staining of the nuclei with 5 μg/ml of the DNA-specific stain, Hoechst 33258, from the same cells as in panel C; panel E: sorted cardiac myocytes were stained with mouse anti-myomesin antibodies and subsequently with Cy3+ anti-mouse antibody. The picture was taken using a Leica TCS confocal laser scanning microscope.

To further confirm the identity of the sorted cells, we immunohistochemically stained sorted myocytes for myomesin, a protein of vertebrate sarcomeric M bands. This protein is exclusively expressed in striated myofibrils [21]. Panel 2E represents its intracellular localisation within a cardiac myocyte. The typical banding of the cell can be seen clearly.

Thereafter, we isolated the RNA from FACS-sorted rod-shaped cardiac myocytes. A myocyte preparation of two rat hearts was fixed with ethanol. We isolated only 8 μg of total RNA after sorting all of the cells of one heart (one half of the preparation), which was completely taken for Northern blotting. Fig. 3A demonstrates the integrity of the 18S and 28S rRNA by ethidium bromide staining of the gel. Besides the RNA from sorted myocytes (after-sort), 10 μg of total RNA extracted from the same cell suspension that was used for sorting (kept in DEPC-treated PBS for 3 h after fixation) served as a control (pre-sort). To further prove the quality of the RNA, we performed a Northern blot with a GAPDH and a α-myosin heavy chain (α-MHC) probe. Fig. 3B displays a single sharp band of 1.2 kb without any signs of degradation, which was expected for GAPDH. The myocyte-specific α-MHC mRNA could also be detected, as shown in Fig. 3C. In summary, this FACS-based approach yielded undegraded RNA that was suitable for Northern blotting.

Fig. 3

Northern blot analysis of FACS sorted cells. Total RNA of ethanol-fixed and FACS-sorted cardiac myocytes (after-sort) or from fixed control myocytes (pre-sort) was prepared as described in Section 2, electrophoresed, transferred onto nylon membranes and hybridised with a GAPDH and an α-MHC probe. (A) The isolated RNA stained with ethidium bromide, (B) the Northern blot with the GAPDH probe and (C) the Northern blot with an α-MHC probe are shown.

We now extended the method to measure gene expression at the level of only a few cells. A recovery ratio of about 50% and the critical step of targeting the small deflected droplets containing large cells (60 to 100 μm in length) led to the decision to sort at least three cells directly into PCR tubes containing lysis buffer.

Initially, we demonstrated the expression of cardiac-specific genes, such as α- and β-cardiac heavy chain (MHC). Table 1 characterises the primers and the annealing temperatures used. The expression of β-actin as a housekeeping gene and the expression of both MHC isoforms are shown in Fig. 4. As a mean, approximately 50% of all sorted cell-probes showed a signal in the subsequent gene-specific PCR resulting in intact cDNAs. Cloning of PCR products followed by sequencing always confirmed the expected products. By quantifying gene-specific PCR products from the cDNA of three and ten sorted cells with ImageQuant™, we could show that ten cells indeed result in a higher product yield than those from three cells (mean, 2.8-fold higher; SEM, 0.39; n=10).

Fig. 4

Gene expression in sorted cardiac myocytes. Gene expression analysis of different genes in FACS sorted cardiac myocytes by PCR was performed as described in Section 2. β-Actin, α-MHC and β-MHC expression were compared in three or ten sorted myocytes, respectively. The product yields from ten cells were, on average, 2.8-fold higher than those from three cells (SEM, 0.39; n=10; quantified with ImageQuantTM). C is the negative control (without cells).

Investigating gene expression by PCR is very susceptible to cell contamination. It is known that the expression of cyclin A and cyclin-dependent kinases cdk2 and -4 decreased below detectable levels in cardiac myocytes during postnatal terminal differentiation [14]. On the other hand, the expression of these cell-cycle-specific genes is present in proliferating cells such as fibroblasts, endothelial cells and smooth muscle cells. We therefore investigated the expression of these genes in sorted cardiac myocytes and fibroblasts. As shown in Fig. 5, cyclin-A and both cdks can be detected in sorted cardiac fibroblasts. In the case of cdk2, we detected two additional PCR products that displayed a slower migration. We subsequently referred to these products as the ‘long’ (cdk2L) and ‘intermediate’ (cdk2I) isoforms of cdk2. Alignment of the sequenced PCR products indicated that cdk2L corresponded to cdk2Lhm, a form of cdk2 that had been described previously in BHK21 cells derived from golden hamsters [17]. On the other hand, although the expression of GAPDH is comparable, we could not detect the expression of cyclin A and cdk2/4 in the myocyte fraction. This is clear proof of the purity of our sorted myocytes.

Fig. 5

Cell-cycle specific mRNA-expression in myocytes and fibroblasts. Demonstration of the mRNA-expression by PCR of cyclin A, cdk2 and cdk4 within sorted cardiac fibroblasts (left) and sorted cardiac myocytes (right). GAPDH served as the control transcript. The myocyte cDNA for this experiment was the same as that for the mRNA transcript proof shown in Fig. 4. Lambda DNA (cleaved with Eco 130; MBI Fermentas, Vilnius, Lithuania) and pUC19 DNA (cleaved with MspI) were used as molecular weight markers.

To ensure that this method is useful for monitoring gene expression in a few cells, we stimulated cardiac myocytes with the phorbol ester, 12-o-tetradecanoylphorbol-13-acetate (TPA) and measured the induced c-fos expression. In cardiac fibroblasts, the expression of c-fos usually peaks after 30 min [22]. However, the c-fos mRNA expression reaches its maximum 1 h after adrenergic stimulation in cardiac myocytes from adult rats (Simm et al., submitted). We therefore stimulated the myocytes for 1 h with 1 μmol/l TPA and measured the c-fos expression from ten sorted cardiac myocytes in comparison to untreated control cells. Fig. 6 clearly confirms the induced c-fos expression (fivefold increase), whereas the expression of the housekeeping gene GAPDH remains unchanged.

Fig. 6

Induced gene expression in sorted cardiac myocytes. Prior to ethanol fixation and FACS sorting, cardiac myocytes were stimulated for 1 h with 1 μM TPA. Gene expression was investigated in ten sorted cardiac myocytes. We observed a 4.9±0.31-fold (n=10) increase in c-fos expression after stimulation and presented the data as mean±SEM relative to GAPDH expression. Control represents a cell-free GAPDH PCR reaction.

In summary, cardiac myocytes in suspension were fixed with 70% ice-cold ethanol for 10 min and sorted using a flow cytometer, which resulted in a pure population of rod-shaped cells. After cell lysis and first strand synthesis, the 3′-end was tailed with oligo-(dA). Amplification with an oligo-(dT24) primer led to the cDNA from a few cells. Thereafter, the expression pattern of a virtually unlimited number of different genes with sequence-specific PCR may be investigated.

4 Discussion

Two thirds of the cells that comprise the heart are non-myocytes, of which, fibroblasts represent the vast majority (over 90%) [4]. Contractile cardiac myocytes represent only one third of the heart cell number. There are several recently published reports about RT-PCR-based approaches to the study of gene expression in tissue probes from endomyocardial or ventricular bioptic material [18, 1, 2, 9]. As heart failure, for example, is frequently accompanied by fibrosis, i.e. hyperplasia of cardiac fibroblasts and accumulation of extracellular matrix components [6], it is unclear whether changed gene expression in patients with heart failure is really related to a different expression within cardiac myocytes or is due rather to a proportional increase in mRNA within cardiac fibroblasts.

We have developed a protocol to overcome the limitation of cellular heterogeneity. The method, based on a FACS, allows us to sort single or up to a large number of cells, being contingent upon the experiments planned. Therefore, single cell RT-PCR, nuclease protection assays, construction of cDNA libraries or even Northern blotting are possible subsequent experimental procedures. FACS-based cell sorting opens an avenue of possibilities, i.e. additional cellular characterisation by antibodies directed against cell-surface molecules [10]or precise cell cycle analysis. We could show that our approach allows one to detect gene expression by Northern blotting as well as at the level of a few cells.

Other groups also developed methods to measure gene expression in single cells [12, 15]. Due to the risk of contamination with genomic DNA, the authors recommend a DNAse I digestion to eliminate the DNA. However, the risk of exogenous contamination is thereby increased. Several laboratories select oligonucleotide primers spanning two exons separated by an intron to distinguish between PCR products derived from cDNA and those derived from genomic DNA. Conventional PCR approaches, e.g. nested PCR [20, 8], are hampered by the limited number of mRNA transcripts that can be investigated. Therefore, Brady et al. [5]and Trümper et al. [24, 25]introduced a PCR protocol in which only the 3′-end of all polyadenylated mRNAs is reverse transcribed, resulting in a homogeneous cDNA population of more than 600 bp. After amplification of this cDNA pool, it is possible to monitor the expression of several mRNA transcripts.

Our protocol for single cell PCR, adapted from ref. [23], does not include either DNAse I or RNAse A digestion, or oligonucleotide primers spanning exon regions separated by an intron. As previously shown by these authors, contaminating genomic DNA does not seem to cause misleading results, as genomic DNA is not amplified during the first PCRs with an oligo-dT primer annealing to the poly-A-tail of the mRNA and the oligo-dA region attached by 3′-tailing of the cDNA. We used oligonucleotide primers close to the 3′- end for the gene-specific second PCR step. Additionally, we sequenced all PCR products and always confirmed the expected gene product. In summary, this fast and reliable technique is useful for measuring gene expression at the single cell level and is helpful for investigating heterogeneous cell populations or for monitoring gene expression, being contingent upon transiently expressed proteins.

Acknowledgements

We thank Anja Scheidler for excellent technical assistance and Dr. Jürgen Hoppe for supporting these investigations. Ulrike Seitzer (Forschungsinstitut Borstel, Germany) was very helpful with the single cell PCR method. We further gratefully acknowledge Dr. Rainer Wolf for advise on the confocal laser scanning microscope and Drs. Hans Eppenberger (ETH, Zürich) and Jutta Schaper (Max-Planck-Institut, Bad Nauheim, Germany) for providing the anti-myomesin monoclonal antibody. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 355/TP C2) and the Graduiertenkolleg ‘Regulation des Zellwachstums’.

References

View Abstract