In conjunction with microbead analysis, flow cytometry is a sensitive and versatile platform for the biochemical, immunological, or molecular biological analysis of proteins and nucleic acids (1–9). Extending this spectrum, the method described herein is based on the dependence of the melting temperature [Tm; the temperature where 50% of the DNA molecules is present in denatured, single-stranded (ss) form] of double-stranded (ds) DNA molecules on their length. This temperature can be reduced to near ambient conditions in the presence of H-bond destabilizing reagents such as DMSO or formamide (10, 11). Above its Tm, a sample containing a PCR product prepared using a pair of biotinylated and fluorescent primers and immobilized on microbeads via the biotin moiety, will dissociate into free ss DNA molecules labeled with the fluorescent tag and the bead-attached non-fluorescent, biotinylated complementary strand. The degree of denaturation, that is, ratio of the ds and ss species at equilibrium will depend on the incubation temperature relative to the Tm. Since the degree of denaturation at a given temperature is expected to depend also on the length of the DNA molecule (12, 13), measurement of the above ratio by flow cytometry may be used for the length comparison of PCR products prepared from DNA samples in diseases where this approach might be of diagnostic value. The range of related potential applications include genetic disorders where the underlying disease might involve insertions, deletions, triplet expansions, microsatellite polymorphisms. In addition, when primers with different length are used (14), single nucleotide polymorphisms (SNPs) or point mutations might also be detected. Although melting point analysis, based on the differential melting kinetics of primers discriminating between two alleles and real-time PCR or high resolution melting provide straightforward answers and are routinely used, a flow cytometric method of comparable sensitivity and multiplicity might offer a valuable alternative platform. Such a potentiality has been tested in two examples: triplet expansions predisposing for Huntington's disease, and a point mutation of the BRCA1 gene.
Expansion of certain nucleotide triplets in the genomic DNA are responsible for several neurodegenerative diseases. Among them, Huntington's chorea is caused by the expanded repetition of the CAG triplet localized in Exon 1 of the Huntingtin gene (also called IT15/HTT/HD). There are four categories based on the number of CAG-repetitions: healthy (<27 CAG), intermediate (27–35 CAG), those with reduced penetrance (35–39 CAG), and those of the diseased phenotype (>40 CAG) (15–17). For diagnosis, the length of genomic DNA fragments carrying the CAG-repeats are to be compared; typically this is performed after PCR amplification, sometimes following restriction enzyme digestion and ligation to adaptors (18) and the products are analyzed in a standard way by agarose or polyacrylamide gel electrophoresis (17, 19, 20), with capillary gel electrophoresis (17, 21, 22), or by sequencing.
The BRCA1 gene, coding for an 1863 amino acid nuclear phosphoprotein, is known to be involved in the homologous recombinational repair (HRR), non-homologous end joining (NHEJ), and transcription-coupled nucleotide excision repair (TC-NER) pathways as well as in the regulation of diverse though interrelated fundamental processes like transcription, cell cycle, and apoptosis in part through its interaction with histone deacetylase and chromatin-remodeling factors (23–25). Its point mutations predispose women to breast and ovarian cancer (26–28). From these BRCA1 mutations, the 185delAG, 5382insC, and the missense C61G are the most common (28). Several methods are available for the detection of point mutations, based on electrophoresis, primer extension, the use of mismatch binding proteins like MutS (a member of the mismatch repair system in Escherichia coli), ligation of adjacent primers at the site of the mutation (ligase chain reaction, LCR), enzymatic cleavage by cleavase (cleavase fragment length polymorphism) or chemical cleavage on chemically modified unpaired nucleotides by piperidine (29, 30), by the more demanding microarray (chip) and qPCR (Taq-man assay) technologies (23) and directly by sequencing. Application of the latter methods in the clinical routine is still somewhat limited due to their relatively high cost when only a few genetic alterations are targeted.
The above consideration explains the ambition to use a cytometric platform for the above purposes. The methodical developments demonstrated herein contribute to the realm of possibilities provided by cytometric microbead analysis.
MATERIALS AND METHODS
Experiments were performed between 2007.10.19. and 2008.09.03.
The plasmids carrying different length CAG repeats (31) were transformed into E. coli DH5α by heat shock and selected on LB plates containing 100 μg/ml ampicillin.
Genomic and plasmid DNA were prepared by the Intron G-DEX Genomic DNA Extraction Kit and DNA-midi Plasmid DNA Purification Kit (Intron Biotechnology, Seongnam-Si, South Korea) following the manufacturer's instructions.
Cells and Clinical Samples
Genomic DNA derived from anonymous patients diagnosed with Huntington's disease, confirmed by sequence analysis and polyacrylamide gel electrophoresis using two different sets of primers, were isolated at the Department of Medical Genetics, School of Medicine, University of Pécs. Genomic DNA with BRCA1 5382insC mutation was from the Department of Clinical Biochemistry and Molecular Pathology, Medical and Health Science Center, University of Debrecen. Genomic DNA derived from peripheral blood lymphocytes of healthy volunteers were isolated in our laboratory. Jurkat cells were grown in RPMI medium supplemented with 10% FCS (fetal calf serum), 2 mM L-glutamine, 100 μg/ml streptomycin, and 100 U/ml penicillin (Sigma-Aldrich, St. Louis, MO).
PCR reactions were performed in 50 μl of 1× reaction buffer [7.5 mM Tris-HCl, 20 mM (NH4)2SO4, 0.1% Tween 20, pH 8.8] containing 2.5 mM dNTP-solution (Promega Life Sciences, Madison, WI), 200 ng template DNA, 10 pmole of each primer, 1.5 mM MgCl, 1 mM β-mercaptoethanol, 3% formamide, 0.2 mg/ml BSA (Promega Life, Sciences, Madison, WI), and 2.5 U Taq polymerase (Fermentas, Vilnius, Lithuania). The primers used for amplification of the CAG repeats in the Exon 1 of the Huntingtin gene (IT15/HTT/HD) were purchased from Integrated DNA Technologies (Coralville, IA). Forward primer, Hunt1: biotin-5′-CAT GGC GAC CCT GGA AAA GCT G-3′, reverse primer, Hunt2: Cy5-5′-GGC GGT GGC GGC TGT TGC TGC TGC TGC TG-3′ and reverse primer, Hunt3: Cy3-5′-GGC GGT GGC GGC TGT TGC TGC TGC TGC TG-3′ (17). The PCR reactions were performed as described in reference 19. The products were analyzed on 3% agarose gels run in 1× TAE (40 mM Tris, 20 mM acetic acid, 10 mM EDTA, pH 8.0).
5382insC Point mutation BRCA1
Allele-specific PCR reactions were performed in 50 μl of 1× reaction buffer [7.5 mM Tris-HCl, 20 mM (NH4)2SO4, 0.1% Tween 20 pH 8.8] containing 2.5 mM dNTP-solution (Promega Life Sciences, Madison, WI), 200 ng template DNA, 10 pmole of each primer, 1.5 mM MgCl, and 2.5 U Taq polymerase (Fermentas, Vilnius, Lithuania). The wild-type allele specific forward primer (BRCA1: Cy5-5′-AAA GCG AGC AAG AGA ATC GCA-3′), the mutation specific forward primer (BRCA2: Cy5-5′-CAG CAG CAG CAG CAG CAG CAC CTT AGC GAG CAA GAG AAT CAC C-3′, containing a 5′ tag of a 18 bp CAG repeat) and the common reverse primer (BRCA3: biotin-5′-AGT CTT ACA AAA TGA AGC GGC CC-3′) were used together in a multiplex reaction (32). The primers determine a 126 bp long fragment of the human BRCA1 gene, including the site of the 5382insC point mutation. PCR reaction was performed using the following profile: 1 × 95°C: 12 s; 37 × 94°C: 15 s; 62°C: 15 s; 72°C: 30 s; 1 × 72°C: 5 min, 4°C. The products were analyzed on 3% agarose gels run in 1× TAE (40 mM Tris, 20 mM acetic acid, 10 mM EDTA, pH 8.0).
Binding of PCR Products to Beads
Biotinylated and fluorescently labeled PCR products were added to 10 000 polymeric streptavidin coated beads (6 μm diameter, purchased from Polysciences, Warrington, PA) in 200 μl 1 × PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) and incubated at RT for 40 min in the dark. After incubation, the beads were washed three times in 1 × PBS and centrifuged at 20,000g for 10 min.
Heat Treatment in Formamide
The beads carrying PCR products on their surface were treated with a concentration series of formamide diluted in ddH2O. In the experiments related to Huntington's disease the concentration of formamide was titrated between 65 and 75 v/v % to find the optimal concentration where the PCR products with different length can be distinguished. In the case of the BRCA1 5382insC point mutation, formamide was titrated between 57 and 62 v/v %. Heat treatment was performed in a total volume of 200 μl, in a PCR tube containing ∼ 10,000 beads. The beads were incubated with formamide at 40°C in the dark for 3 min, washed in 600 μl of 1 × PBS three times and analyzed by flow cytometry.
Flow cytometric measurements were performed using a Becton Dickinson FACSArray instrument (San Jose, CA). Fluorescence signals were collected in the logarithmic mode and the data were analyzed with the BD FACSArray system software. Fluorescence signals were detected in the PI/Yellow and Far Red channels through the 585/42 and 675 LP interference filters of the instrument, respectively. Every sample was gated with the same protocol on FSC–SSC plots.
We used appropriate positive and negative controls, including samples representing background signals, for every flow cytometric measurement. Instrument performance was checked on a weekly basis using Spherotech 8-Peak Validation Beads.
Results of the measurements were presented as bar diagrams where the error bars indicate the standard deviation of three independent measurements (n = 3). Mean intensities of the fluorescence distribution histograms have been determined by the above software.
List-Mode Data Files
FCS data file can be obtained by contacting László Imre Table 1.
Table 1. Fluorescence reagent description
Excitation peak (nm)
Emission peak (nm)
The principle of the experimental set-up is shown in Figure 1. A pair of short and long PCR products are prepared using a fluorescent dye-conjugated and a biotinylated primer to label the two ends. The two samples of the PCR products are immobilized on streptavidine coated microbeads and heat-treated at a temperature between the melting points of the long and the short PCR products. This temperature range can be down-shifted using formamide, an H-bond forming agent (10), to avoid damage of the beads and evaporation of the buffer. The mean fluorescence intensity of the beads carrying the shorter PCR products is expected to decrease relative to the beads with the longer ones. To demonstrate the feasibility of this strategy, we first tested if triplet expansions of different length can be distinguished in the case of Huntigton's disease.
The model system used included a set of plasmids containing Exon 1 of the Huntingtin (IT15/HTT/HD) gene harboring different length CAG repeats (31). In the experiments shown in Figure 2, the DNA region of the CAG repeats was amplified using the plasmids carrying 27 and 51 CAG-repetitions. The Cy5 and Cy3 labeled reverse and biotin labeled forward primers were positioned as shown in Figure 2B. A 1:1 mixture of the PCR products was immobilized on streptavidinated microbeads, then heat treatment was performed at 40°C in the presence of formamide. The concentration of formamide was titrated in a range encompassing low and high concentrations, where both PCR products are expected to stay annealed, and where they may be differentially denatured, respectively (see Fig. 2C). At 75% formamide, 96% of the short PCR products were denatured, that is, their fluorescent-labeled strands dissociated upon incubation at 40°C, whereas the intensity of the beads carrying the longer products decreased only by 26%, which did not depend on the fluorescent label carried by the PCR products (data not shown).
As shown in Figure 2D, PCR products amplified from the plasmids containing intermediate length CAG repetition could also be clearly distinguished from those amplified from the 27 CAG plasmids representing the upper limit of normal length. Since the patients suffering from Huntington's disease are mostly heterozygous (33) for the pathological triplet-expansion, the measurement was also performed in such a way that the samples contained a 1:1 mixture of the PCR product of the >27 CAG plasmid and the PCR product derived from the 27 CAG plasmid. The assay could clearly distinguish between the templates representing normal and pathological triplet expansions in the case of this model experiment representing heterozygotes, irrespective of the fluorescent labels used. Next, genomic DNA isolated from lymphocytes of healthy volunteers, the Jurkat cell line expected to exhibit normal CAG repetition at the locus, and anonymous patients suffering from Huntington's disease with predetermined number of CAG repeats were amplified and treated as before (Fig. 3). All the samples with CAG repeat numbers in the healthy range gave low average fluorescence values as compared to the samples derived from people afflicted with the disease.
This simple strategy of differential melting point analyses was also tested as to its utility for the diagnosis of point mutations, when combined with allele specific PCR (14). For the detection of BRCA1 5382insC, a modified allele specific PCR reaction was used, applying three primers in a multiplex reaction: wild-type and mutation specific forward primers and a common reverse primer. The specificity of the forward primers was determined by the 3′ terminal nucleotide as shown in Figure 4B. Because of the presence of an extension sequence (CAG tag) at the 5′ end of the mutation specific forward primer, two different length PCR products are generated if the analyzed sample is heterozygous for the mutation; thus the mutant and the wild-type alleles can be discriminated in a cytometer. To optimize the concentration of formamide solution, the PCR products representing the wild type and mutant alleles were generated in separate amplification reactions, using Cy5 labeled forward and biotin labeled reverse primers and the genomic DNA template derived from a patient carrying one BRCA1 5382insC allele. The optimal concentration of formamide where the mean fluorescence intensity of the short products decreased almost to background levels while the long PCR products decreased only to ∼ 40% of the reference (measured at 57% formamide) was at 62% formamide, as shown in Figure 4C. The measurement could also be carried out in a fully multiplex mode, from the PCR step throughout the assay (see Fig. 4D): Using Cy5 labeled wild type specific and mutation specific forward primers and common biotin labeled reverse primers, a clinical sample with determined heterozygous BRCA1 5382insC mutation and the control DNA of Jurkat cells could be clearly distinguished. The 62% formamide concentration was somewhat below what was expected based on the calculated Tm–s as shown in Figure 4A.
We have developed a novel molecular diagnostic method applying melting temperature analysis in combination with flow cytometry. The multiplex microbead assay described can be applied for the diagnosis of several genetic disorders where the length or the sequence of a certain genomic segment is changed. These include triplet expansions, SNP-s, point mutations, microsatellite polymorphisms, and insertion/deletion polymorphisms. This method may be feasible to use if the difference between the melting points of the two different length PCR products allows their differential elution upon denaturation. This minimal difference is estimated to be around 1.6°C, the value calculated for the two PCR products used in the case of the BRCA1 point mutation. The formamide concentration where the difference is maximal is to be empirically determined in a relatively narrow concentration range. The mild conditions of denaturation achieved using formamide are expected to prevent any major nonspecific effect like molecular aggregation, nonspecific binding of DNA to the beads or adverse chemical reactions, and allow for rather simple experimental procedures.
The strategy was applied successfully both for the detection of the rare Huntington's disease and the frequent 5382insC mutation of BRCA1 gene coding for a frameshift mutation resulting in a premature stop signal at codon 1829 (34) increasing the risk of breast and ovarian cancer (26–28). The prevalence of BRCA1 mutation carriers is around 1/800 in the general population, however, it can vary significantly among different countries or ethnic groups (28, 35).
In the case of Huntington's disease, the variable length of the CAG repeat in Exon 1 of the Huntingtin gene was measured. The smallest difference that could be detected using our standard assay conditions comprised five CAG triplets, that is, 15 base pairs, present in a heterozygous manner. However, it is likely that this limit can be improved by further optimization. Data obtained through the analysis of 10 Huntington's patients with determined CAG repeats and of five healthy donors have demonstrated that the assay is remarkably reliable and can be certainly considered as an initial test in large-scale population screening programs. Although a battery of different methods has been introduced for the diagnosis of triplet expansions, neither of these appears to surpass the procedure described herein in simplicity and ease of multiplexing, considering the steps of analysis.
The approach described can also be used for the detection of other point mutations and of SNPs, applying allele specific PCR reactions to generate different length mutation/SNP specific and wild type specific PCR products. In the case of the 5382insC mutation of the BRCA1 gene tested, we added a 6-CAG tag to the mutation specific primer to achieve a length difference between the two PCR products that proved to be sufficient for length discrimination in the case of the repeat expansion disease. On the basis of the fact that the two PCR products of different CAG repeat length were shown not to interfere with each-other at immobilization to, or upon debinding of the non-biotinylated strands from the same bead (see Fig. 2, panel C), it would seem possible to develop a multiplex microbead assay for the simultaneous detection of all the 14 cancer risk associated BRCA1 mutations (34) in one single well of a 96 or a 348 well plate with the combination of three different types of fluorescent dyes linked to the PCR products and six kinds of microbeads with different size and color.
The data presented herein exemplify two areas where the above approach may be utilized. Introduction of this strategy for routine diagnostic purposes may be facilitated by the readily available instrumentation and the wealth of relevant experience already present in many laboratories. Beyond these circumstances, the relative simplicity of the method may offer an advantage over the molecular genetic approaches routinely used for such purposes. It may be offered in the case of demand for high-throughput multiplex assays for the simultaneous analysis of a limited number of genetic conditions in large populations of human and non-human biological samples.