| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Technical Briefs |
1 PharmaSeq, Inc., Monmouth Junction, NJ;
2 Department of Microbiology & Molecular Genetics, University of Medicine and Dentistry, New Jersey-New Jersey Medical School, Newark, NJ;
aaddress correspondence to this author at: PharmaSeq, Inc., 11 Deer Park Drive, Suite 104, Monmouth Junction, NJ 08852; fax 732-355-0102, e-mail mandecki{at}pharmaseq.com
Abstract
Background: We developed and evaluated a genotyping assay for detection of 50 cystic fibrosis (CF) mutations. The assay is based on small (500 µm) electronic chips, radio frequency (RF) microtransponders (MTPs). The chips are analyzed on a unique fluorescence and RF readout instrument.
Methods: We divided the CF assay into 4 panels: core, Hispanic, African-American, and Caucasian. We amplified 18 CF transmembrane regulator (CFTR) DNA fragments covering 50 mutations by use of multiplex PCR using 18 CFTR gene-specific primer pairs. PCR was followed by multiplex allele-specific primer extension (ASPE) reactions and hybridization to capture probes synthesized on MTPs. We used 100 ASPE primers and 100 capture probes. We performed fluorescence measurements of hybridized MTP kits and assay analysis using a custom automated bench-top flow instrument.
Results: We validated the system by performing the assay on 23 commercial DNA samples in an internal study and 32 DNA samples in an external study. For internal and external studies, correct calls were 98.8% and 95.7%, false-positive calls 1.1% and 3.9%, and false-negative calls 0.12% and 0.36%, respectively.
Conclusions: The MTP-based multiplex assay and analysis platform can be used for CF genotyping.
Cystic fibrosis (CF) is caused by one or more mutations in the gene encoding for the CF transmembrane conductance regulator (CFTR) protein. CF occurs when both copies of the CFTR gene function abnormally, and one functional copy is sufficient to prevent the disease. In the Caucasian population, CF is inherited with a frequency of 1:3300, making it the most lethal inherited disease of childhood, but carrier frequency and incidence of CF vary with race and ethnic group (1). A single mutation causing loss of the phenylalanine residue at position 508 (
F508) accounts for nearly 70% of all mutations observed in Caucasians with CF, but more than 1000 other mutations of the CFTR gene have been reported in all races and ethnic groups. In May 2001, the American College of Medical Genetics published a recommended panel of 25 mutations and 6 polymorphisms for population-based CF screening (2).
We describe a new platform for performing a multiplexed genotyping assay based on radio frequency (RF) microtransponders (MTPs) and provide a working example, the CF assay. CF mutations tested and the rationale are provided in Table 1
and Supplemental Data (see the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol53/issue7. We designed this assay to be ethnic-group specific, thus simplifying the design and potentially reducing costs. The mutations are grouped into 4 panels: core, Caucasian, Hispanic, and African-American. Because the assay is ethnic specific, efficiency is increased, with fewer DNA probes and reagents needed. A patient would be tested with 2 panels, the core panel and one ethnic panel; more than one ethnic panel might be used for patients with complex ethnic backgrounds.
|
|
The key element of the assay method is the MTP (Fig. 1A
), a monolithic 500 µm x 500 µm integrated circuit chip that can transmit its identification code by RF. Each chip consists of photocells, read-only memory (ROM), transmit logic circuitry, and an antenna loop. Visible light, typically red or green, is pulsed at 1.5 MHz to provide power and a stable clocking signal for the logic circuitry. The circuitry accesses the contents of ROM (the ID value) and modulates the current through the antenna in correspondence with the ID value. The resulting variable magnetic field in the vicinity of the MTP can then be measured with a nearby receiving coil and decoded to provide the ID value, which identifies the oligonucleotide immobilized on the MTP. The current MTP design uses 10 bits to encode the ID value, allowing 1024 unique values; however, the ROM contains an additional 40 unused bits, so the MTPs could be manufactured to have as many as 250 (
1015) unique ID values. Before use, the MTPs are coated with a polymer that places both hydroxy and amino groups on the surface of the chip. Such derivatized MTPs are subjected to oligonucleotide synthesis.
|
In preparation for the flow reader analysis, the MTPs are suspended in a liquid medium that prevents sedimentation of MTPs but allows flow characteristics comparable to water when sheared. The suspension is repeatedly passed through a narrow channel of the instrument, where the ID values are read and fluorescence measurements are made. The flow system is designed to support a transfer rate of up to 1000 MTPs/s. The time needed to read the ID can be as short as 300 µs and the time to read fluorescence 1–2 ms. Multiple forward and reverse passes, typically 50 total passes, of the MTPs through the flow channel are required to obtain enough data for analysis. The instrument in the present study, Tsunami IV, uses a 532-nm, 300-mW laser for both RF identification and fluorescence at a single location on one side of the flow channel. We used custom software named Retro for data analysis. Both the flow reader and MTPs are described in more detail in a recent report (3).
The principle for detecting mutations is allele-specific primer extension (ASPE) (4)(5) on the PCR-amplified CFTR DNA; the schematic of the assay is shown in Fig. 1B
. For each mutation, 2 primers were prepared, one specific to the wild-type allele and another specific to the mutant allele. The sequence differences between the 2 primers are the tag sequence at the 5' end and a single nucleotide at the 3' end (an "anchor" sequence). Thus, for any particular allele, only 1 primer was extended in a reaction with DNA polymerase in the presence of 4 dNTPs. After the ASPE reaction, the target DNA was hybridized to a capture probe, the sequences of which are complementary to tag sequences commonly referred to as universal tags (4)(6)(7)(8). The capture probe is a synthetic oligonucleotide (24 nucleotides) covalently bound to the MTP surface. Because the tag sequence was present at the 5' end of allele-specific primers, the capture reaction was very specific. In the ASPE reaction, we used biotin-labeled dCTP in place of dCTP. Thus, the ASPE target typically contained several biotin moieties, which were subsequently visualized as a result of a binding reaction with phycoerythrin-labeled streptavidin (phycoerythrin is a pigment-carrying, highly fluorescent protein from red algae). The fluorescence was then read in the flow reader instrument.
Eighteen pairs of PCR primers were designed using an in-house custom software package, SimuPlex (9). SimuPlex accepts a sequence file comprising all the loci for which PCR primers are to be designed. The SimuPlex then identifies all qualified primers that meet the criteria defined in the parameter file and executes a local BLAST (Basic Local Alignment Search Tool) search that filters out undesired primers. Once the candidates are filtered, SimuPlex creates a "seed" multiplex set and then uses a simulated annealing algorithm to search the surrounding solution space for even better multiplex sets. The melting temperature (Tm) and free-energy calculations are based on the most accurate and up-to-date formulas and thermodynamic data sets (10)(11). The results of a multiplex PCR reaction for which primers were designed using SimuPlex software is shown in Fig. 1C
.
Allele-specific primers containing 24-nt tag sequences were designed using 3 custom programs written in Python: ExtractProbes.py, FindOptProbe.py, and Tags2Probes.py. The tag sequences were at the 5' end of the primers, and allele-specific sequences were at the 3' end. The allele-specific sequences varied in length, but always possessed a Tm of 50 °C. For each biallelic single nucleotide polymorphism analyzed by ASPE, 2 allele-specific primers (ASP) were synthesized, with each ASP differing in the tag sequence and in the polymorphic nucleotide contained at its 3' terminus.
We conducted 2 separate series of experiments to validate the performance of the assay. In the internal study, we used 23 standard (Coriell) genomic DNA samples. In addition, we used 32 coded genomic DNA samples in an external study completed in Dr. Dermody’s laboratory at the University of Medicine and Dentistry of New Jersey. During the course of this project, >100 CF reagent sets were prepared. Each reagent set consisted of a vial containing derivatized MTPs that compose the mutation panel and the assay file on electronic media. Typically, 3 MTPs were used for each probe to achieve multiple readouts for statistical accuracy.
The results from the internal study (Table 1
) indicate that correct calls were 98.8% of all determinations (807 total calls), and false-positive and false-negative calls were 1.1% and 0.12%, respectively. The results from the external study are shown in the Supplemental Data. Correct calls were 95.7% of all determinations (1086 total calls), and false-positive and false-negative calls were 3.9% and 0.36%, respectively. In addition, 27 synthetic 60-nt oligonucleotides were designed to simulate DNA mutations not present in the Coriell DNA samples. Assays performed on the synthetic samples resulted in 100% correct calls of homozygous mutation.
We are generally pleased with the results obtained in both the internal and external testing. The overall percentage of correct calls was high: 98.8% and 95.6%, respectively. Especially encouraging were high fluorescence ratios (wild type-to-mutant oligo probe), approaching 100 in many cases, indicating a high potential of the assay for DNA testing, and in particular CF testing. The wrong calls seem to be clustered for specific mutations, suggesting difficulties with certain oligonucleotides or PCR products.
In summary, MTPs were used as solid phase in a CF assay. Although the reported rate of false positives (1%–4%) is higher than in commercially available CF assays [Luminex (12), Roche, and Applera (13)], we are confident that it can be improved, because the biochemical basis of the assay is well understood and the biochemical principle is similar to that implemented in the Luminex CF assay (12). The main advantage of the MTP platform is the large number of ID codes available, currently 1024 but readily expandable. The expansion might be justified if the number of mutations being tested for increases, or if other types of assays require a higher multiplex level.
Acknowledgments
Grant funding/support: This work was funded by National Institutes of Health Grant HL074607.
Financial disclosures: The authors associated with PharmaSeq, Inc., have equity interest in the company. J.D. is a consultant to the company.
Acknowledgments: We thank Richard G. Morris and Marvin Schwalb for helpful discussions and encouragement.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
