When tested together, twenty nine artificially synthesized, cloned, homozygous cystic fibrosis controls streamline test quality control by minimizing control assay number, cost, and assay time. Any rare or unavailable reported sequence can be PCR amplified from total genomic DNA or RNA template using a synthesized primer incorporating the mutant, polymorphic or variant sequence and a paired upstream or downstream primer. These synthesized upstream and downstream fragments overlap at the modified site and are amplified together to obtain a homozygous mutant control. The 5′ and 3′ flanking sequences were selected to span all the PCR sites in the test platforms we were using. In this manner 1 to 4 homozygous mutations were artificially synthesized into each of 17 fragments 433 bp-933 bp long by PCR amplification on total normal human genomic DNA template. Then one to three synthesized fragments were blunt end cloned into 9 vectors and validated by sequencing. Together these mutations included (1) 7 mutations on two joined fragments cloned into a single vector's multiple cloning site and (2) all 25 homozygous mutations originally recommended by ACMG for the core cystic fibrosis panel which are sufficiently long to be amplified by the PCR primer pairs in most commercial platforms. Together these cloned mutant sequences provide controls for each PCR assay to optimize multiplex test reliability for all recommended ACMG mutations.
Key words: controls: homozygous, synthetic, PCR amplified, multiplex
Achieving a reliability of 99% for any 25 mutation test requires an individual mutation test reliability of 99.96% (1 incorrect result among 2500 reported results). In contrast, achieving a test reliability of 99% for a 100 mutation test requires an individual mutation test reliability of 99.99% (1 incorrect among 10,000 reported results). When applying ever larger multiplex tests, simultaneously testing controls for all target sequences is required to maintain the most reliable molecular genetic analyses. Thus the College of American Pathologists' Molecular Genetics Committee dictates that whenever possible a control needs to be included in an independent control reaction for each molecular assay and that the best available control is to be used. At the same time, obtaining controls for each reported gene and its common mutations is typically the most difficult hurdle required to introduce any new assay into a Clinical Molecular Genetics laboratory's test menu. The following protocol demonstrates the relative ease required to synthesize control sequences of standard PCR amplifiable lengths for use by multiple laboratory specific and commercial multiplex cystic fibrosis test platforms.
The cystic fibrosis transmembrane receptor (CFTR) gene is mutated in both alleles in patients with cystic fibrosis (CF [MIM 21977]) and congenital bilateral aplasia of vas deferens (CBAVD [MIM 277180]). Cystic fibrosis is the most common lethal genetic disease in Caucasians, affecting about 1 in 2500 newborns while CBAVD renders about 1 in 5,000 males sterile with about half of these patients exhibiting cystic-fibrosis like symptoms and a few have full-blown cystic fibrosis. Over 1000 mutations and 50 polymorphisms have been reported worldwide throughout a large portion of this 24 exon CFTR gene with its 4600 basepair coding sequence that spans 188 kb of genomic DNA (Cystic Fibrosis Mutation Database.)
The ACMG 25 mutation cystic fibrosis panel was selected by the Cystic Fibrosis Committee of the American College of Medical Genetics for screening pregnant Caucasian women who are at a more substantial risk of having a cystic fibrosis fetus than women from other races (Grody et al. 2001). These selected 25 CFTR mutations included all mutations with a frequency of at least 0.1% in cystic fibrosis patients. This was considered an attainable goal for a large number of laboratories to offer routinely and at the same time charge a reasonable fee to encourage third party payment for their laboratory service. Currently thousands of pregnant Caucasian patients are tested each month by dozens of laboratories as a first step in detecting most fetuses at-risk for cystic fibrosis.
Although initially developed for population carrier screening, the 25 mutation core panel can also be used to screen patients with suspicious symptoms because over 90% of all affected Caucasian patients will test positive for at least one mutation including over 99% of affected Northern European Caucasian patients (Lebo and Grody, unpublished data). We incorporated the 5T allele into the select group of 25 most common mutations tested in symptomatic patients because (1) the 5T allele with the severe ΔF508 mutation and the 5T allele with decreased penetrance are the two most common alleles resulting in cystic fibrosis-like symptoms including CBAVD, (2) 0.17% to 3.4% of cystic fibrosis patients with severe symptoms have one allele with the 5T sequence without any other detected mutation, and (3) about 40% of CBAVD patients are compound heterozygotes for 5T and another CFTR gene mutation (Claustres et al. 2000; Kerem et al. 1997; Lebo and Grody, unpublished data).
When the 25 mutation panel was selected by the ACMG Cystic Fibrosis Committee, the controls for each of these mutations were unavailable in any one location. Thus many laboratories expanded their CFTR mutation testing panel to 25 mutations and began offering the service without simultaneously testing all the appropriate controls. In order to meet the substantial new demand for better controls, Coriell Cell Repository, Camden, N.J., aggressively collected, transformed, and distributed human cell lines that contained all available mutations among those to be tested. In spite of substantial efforts to date, Coriell has yet to market a collection of cell lines that together contain all 25 mutations. Coriell has been marketing total genomic DNA from cell lines transformed with Epstein Barr virus with most of these represented as heterozygous mutations as a Product of Substantial Equivalence. Genzyme, which began testing about 87 mutations a decade ago, had to complete many thousands of tests over several years to identify and then incorporate a patient DNA control for each of the 87 selected mutations. Furthermore, testing all 25 cystic fibrosis mutations in the Coriell collection with each unknown set of patient samples requires a substantial investment in labor and test materials. For this reason dozens of laboratories doing 25 mutation tests have been rotating selected control patient DNAs from Coriell through their regular clinical protocol. In this fashion, all available mutation controls are only tested among several independent multiplex assays run on different days, but all mutant sequences are never tested at the same time in a single multiplex assay. Furthermore, most of the Coriell Cell lines are heterozygous at the CFTR mutation site so that these controls cannot be mixed together because the proportion of normal alleles increases each time another heterozygous genome is introduced. Therefore these heterozygous controls must be used individually and cannot be used to determine whether an assay distinguishes between a homozygous and heterozygous DNA sample, the primary deficiency in our studies (See Discussion).
No matter how rare or unavailable, all selected reported mutations, polymorphisms, and variants can be synthesized not only as homozygous but also as heterozygous controls using PCR primer pairs with mutant sequences, genomic DNA template, and editing DNA polymerase. Prior to offering a 25 cystic fibrosis mutation test at Akron Children's Hospital three years ago, our laboratory artificially synthesized and verified homozygous controls for all reported mutations by independent assays on the Innogenetics multiplex format (
Primers were selected incorporating published mutations into the mutation-specific primers and selecting the background sequences including the flanking primers from the CFTR gene sequences in The Genome Database. All primers were synthesized by Invitrogen. Pfu DNA Polymerase or Pfu Ultra High-Fidelity DNA Polymerase was used for high fidelity PCR amplification (Stratagene, La Jolla, Calif.). All PCR amplifications were performed according to recommended assay conditions and protocols for the selected enzyme (
Results
PCR Amplification with Synthesized Mutant Primers
Our goal was to design and synthesize a complete set of ACMG-recommended homozygous CFTR gene mutation controls for use in the widest variety of commercial and laboratory-specific CFTR platforms. Thus pairs of PCR primer sites were selected to span all known primer sites reported to amplify each CFTR mutation-containing region (The Genome Database;
Additional mutations were added to a synthesized mutation-carrying fragment when another homozygous mutation was desired on the same gene fragment. For example, when mutation M3 was desired on the same fragment as mutations M1/M2, it was synthesized onto Product #3 template using forward F3 and reverse R3 primers synthesized to include mutant sequence M3 (
Initially 31 of the first 34 manually selected primer pairs amplified the correct target sequences. Three flanking primer sites that initially failed to amplify a correct length unique product were moved to an even more distant location from the mutation site(s), thus assuring that the synthesized mutation-carrying CFTR fragment would also span all the known primer sites reported to amplify the gene region to be tested. All 3 additional selected primers amplified the site for which each was designed.
Our design included a plan to synthesize a small number of fragments containing all 29 selected CFTR mutations that would provide an optimal multiplex control according to the selected commercial or private format using the minimal number of control mixtures. The number of mutations that can be added to a single sequence is determined by the minimal distance between mutations required to distinguish unambiguously between normal and mutant sequences tested by the selected assay format including (1) nitrocellulose filters with slot blot designated locations, fluorescent beads, or microchips each with locations to which ASOs are hybridized uniquely, or (2) Mass Spec that hybridizes complementary nucleotide sequences adjacent to the mutations to be measured and then adds additional 3′ nucleotides until a base complementary to the single base subtracted from the reaction mixture is encountered. For instance, all our homozygous controls are tested with each assay to assure that each control unambiguously gives a homozygous mutant signal with no cross hybridization to the normal sequence.
When heterozygous controls are desired for applications like Mass Spec, multiple approaches can be used. For instance, normal genomic sequences can be PCR amplified directly from total normal genomic template DNA using the flanking primers like F1 and R2 (
The number of homozygous mutations that can be added to any single DNA fragment that can be tested unambiguously depends upon the assay approach used by the selected CFTR test platform as well as the proximity of each mutations pair, the type and size of mutation (nucleotide substitution(s), insertion, or deletion), and the immediately flanking sequences of each mutation. For instance, the Intron 10/Exon 11 fragment spans 5 common mutation sites: 1717-1G->A, G542X, G551D, R553X, and R560T (
Following these strategies, 17 mutation-carrying fragments were synthesized and tested with the Innogenetics assay to confirm that these fragments carried the expected mutations. Some of the PCR fragment primers were designed to overlap with more than one independent fragment to allow pairs of amplicons to be spliced together in the orientation of choice by PCR. Then the 17 fragments were spliced together into 9 fragments and blunt end cloned. For instance, Product #9 (
Following the final PCR amplification and purification when required, the 9 spliced fragments were inserted into the cloning site of a single vector. Inserts isolated from individual clones were screened for the correct size after restriction enzyme digestion. Then the presence of the synthesized CFTR mutation(s) and flanking sequences were detected using the Innogenetics CFTR multiplex mutation assay and further analyzed by bidirectional sequencing of mini-prepped plasmid DNA. A 30% aliquot of a single miniprepped sample isolated from a 3 ml suspension culture provided substantially more than sufficient material for bidirectional sequencing, validating two commercial platforms, and providing complete controls for each of the thousands of cystic fibrosis samples tested at Akron Children's Hospital (Lebo et al. 2003). Bidirectional sequencing was completed by Cleveland Genomics (Cleveland, Ohio) using M13 sequencing primers hybridizing to these sites in the cloning vector and selected internal PCR primers used in the original PCR synthesis. Each of the 9 fragments carried the expected subset of 29 different cystic fibrosis mutations (
Three Insert Clone with A455E and N1303K Homologous Regions
In 15 of 17 synthesized, cloned, and sequenced fragments amplified from total genomic DNA using synthesized DNA fragments to introduce 31 other mutant sites had only the primary CFTR sequence listed for the normal gene [CFTR Mutation Database] with the exception of three normally variant single nucleotide substitutions: 622-61A->T; 1525-61A->G, and 405+46G->T. The first clone with three inserts had to be reengineered to replace the first and third segments (
Validation on Two Commercial Platforms:
The Innogenetics multiplex CFTR mutation format coamplified three combinations of these 9 clones each with unique CFTR mutation combinations, and analyzed the amplified fragments on two sets of mutation-specific ASOs bound to nitrocellulose filter paper. As part of this format validation, two different Intron10/Exon11 fragments were sequenced and tested: the first with G551D along with 1717-1G->A, G542X, and R560T, and the second with R553X along with 1717-1G->A, G542X, and R560T. When tested individually, the first fragment hybridized uniquely to the G551D amplified control site as well as to the other three mutations (1717-1G->A, G542X, and R560T), but not to the normal site R553 because the G551D mutation interferes with the binding to the normal R553 sequence on the nitrocellulose filter strip (
Then all the fragments that were cloned and verified by bidirectional sequencing (
In contrast, the TmBiosciences fluorescent bead assay simultaneously and unambiguously characterized all the mutant sequences even when all were mixed together into one control tube (
At the same time, these homozygous PCR amplified controls gave heterozygous signals for mutation #3 and mutation #9 for the first 9 homozygous mutations synthesized when tested on a multiplex cystic fibrosis tests from a third manufacturer. Although the manufacturer corrected the specificity of mutation #3 immediately upon learning of our results, we moved on to the Innogenetics test system when we found a second homozygous control with a heterozygous signal pattern among the first nine tested. This result emphasizes the importance of using homozygous controls over the heterozygous cell lines provided by Coriell. Obviously the third manufacturer had no way of knowing that a homozygous sample could give a heterozygous result without an appropriate homozygous control. At the same time, when less than a full panel of mutation controls is tested on any one analytical run, small changes in hybridization conditions might modify hybridization of one of the 29 targets without changing the other 28 results. Our mixed set of controls can detect both of these test specificity failures.
Discussion
Multiplex PCR amplification reactions typically cannot be relied upon to amplify each existing target site so that all can be visualized after a typical 106-fold PCR amplification. For instance, one of us (RVL, unpub. results) developed a 15 site multiplex PCR test for Y chromosome deletions that was assayed in 3 groups of 5 PCR target sites. When completing 105 patient assays, 12 patients were identified and reported with deletions of one or more adjacent targets that were verified by repeat individual PCR analysis at each site that failed to amplify during the initial screen. However, another 11 of the 105 samples had individual reactions that failed to amplify existing targets in two or more sites that were physically separated by amplified sites. In these samples, the non-amplifying sites could always be amplified by repeating the initially failed PCR reaction with individual primer pairs to each previously failed site with 0.1× or 10× of the total genomic DNA originally tested or by substituting repurified genomic DNA. Thus PCR amplifying all control targets together is critical in helping to assure that multiplex PCR reactions are working effectively.
Our sequence verified controls are sufficiently long to serve as internal simultaneous multiplex PCR amplification controls as well as homozygous mutant sequence controls. Sequencing is considered the gold standard of DNA sequence validation. Our cloned, sequenced controls are nested within sufficiently long CFTR gene fragments to include all the PCR primer sites save one site in eight commercial platforms (Aytay et al, Submitted). When using heterozygous cell lines, one has not confirmed that the test system can distinguish unambiguously between homozygous and heterozygous mutations. In contrast, Bajjani and Amos have prepared multiple additional cystic fibrosis controls by synthesizing and mixing 100 bp fragments. This approach does not control for multiplex PCR amplification of homozygous mutant sites, but does require adding the control sequences to control tubes after PCR amplification, typically in a different laboratory location. Furthermore, these controls almost certainly consist of a mixture of fragments with similar but not identical sequences. When synthesizing 100 basepair control DNA fragments, the proportion of identical synthesized sequences typically drops precipitously as the synthesized length exceeds 60 basepairs. Thus the concentration of the synthetic primers can only be estimated and the reliability of 100 bp synthesized controls is anticipated to be less than perfect. Furthermore, when one requires that the most robust controls are verified by sequencing, repeated bidirectional sequencing would need to be completed prior to using any newly synthesized fragments to replace exhausted control stocks. Since readable sequence typically begins about 30-40 basepairs downstream from the end of the target, then 100 bp synthesized sequences would be verified in only 1 direction. In contrast, the entire bidirectional sequence can be determined for cloned controls by using sequencing primers some distance into the vector from the cloning site or into the adjacent cloned control fragment.
Clearly, homozygous controls for each mutant site are preferable to no control and additional mutation controls prepared by this method can be used prior to developing cloned controls. However, each desired synthesized control can be readily prepared and cloned using four selected primer sites and total normal genomic DNA that give a PCR product sufficient to control for PCR amplification. We have used these controls since we first offered a 25 mutation multiplex cystic fibrosis test in our laboratory three years ago (Lebo et al. 2003). Preparing a control that works on multiple different commercial platforms required considerably more validation time prior to offering to the general genetics community (Aytay et al., 2005).
Multiple approaches can be used to synthesize heterozygous controls. For instance, normal genomic sequences can be PCR amplified directly from total normal genomic template DNA using the same flanking primers like F1 and R2 (
In summary, we have synthesized 433 bp to 933 bp sequences containing the 25 ACMG recommended mutation panel in 17 fragments inserted into 9 vector cloning sites. The entire sequences and cloning sites have been verified. These artificial mixtures can be added to test tubes adjacent to unknown samples in the PCR set up area to control for the multiplex PCR amplified cystic fibrosis test kits. Individual idiosyncrasies of test formats for which these may be employed will depend upon the precise PCR primer sites, whether these sites include normally variant sequences that interfere with PCR amplification, whether the primer pairs also amplify homologous genomic DNA sequences at other locations, and the relative amplification of each site during PCR at multiple different salt conditions found in multiple extracted DNA samples. Nevertheless, using this artificial mixture with all 25 mutations will surely improve the reliability of each cystic fibrosis assay when run as a control with every assay of unknown patient samples.
This approach provides any laboratory with a ready means to produce homozygous and heterozygous controls for any laboratory assay without requesting and sharing samples with other laboratories. When adding 32 DNA tests to the menu of the Clinical Molecular Genetics Laboratory at Boston University, my laboratory typically spent 2-4 weeks to pick, order and test PCR primers and 2-4 months to obtain the required controls from generous colleagues. Synthesizing all unavailable control DNA fragments and verifying the sequence provides a rapid means to satisfy the most stringent laboratory reviewer and further expedites test development for small groups of cooperating laboratories. Substantially larger numbers of multiplex controls can be offered through facilities with the wherewithal to develop, produce, and maintain the highest quality multiplex controls required to satisfy very large numbers of laboratories over wide geographical regions.
1. PCR Synthesis of Homozygous and Heterozygous Controls. Sections I-X illustrates the incorporation of four defined homozygous mutations (M1, M2, M3, M4) into two different genomic fragments that are then spliced together for blunt end cloning. Sections XI-XV illustrate the incorporation of three defined heterozygous mutations (M1, M2, M3) into two different genomic fragments that when tested together give a heterozygous result. I. The first product to be synthesized has two mutations M1/M2 in a fragment with its upstream site defined by primer F1 and downstream site defined by primer R2 using total genomic Template DNA. II. PCR amplification incorporated mutations M1 and M2 into upstream PCR amplified Product #1 using reverse primer R1 containing the mutant sequences M1/M2 and forward upstream primer F1. III. In a separate reaction mutations M1 and M2 were incorporated into downstream PCR amplified Product #2 using forward primer F2 and reverse downstream flanking primer R2. Then Products #1 and #2 are mixed with additional F2 and R2 primers and amplified together to produce Product #3 (IV). V. Then mutation M3 was introduced into Product #3 using primers F1 and R3 to amplify upstream Product #4 (VI). Similarly Primers F3 and R2 amplify downstream Product #5 now with all three mutations M1, M2, and M3 (VII). Then Products #4 and #5 are spliced together with primers F1 and R2 to make Product #6 (VII) with mutations M1, M2, M3 and upstream and downstream sequences. Subsequently mutation M4 was introduced into another CFTR gene segment in an analogous fashion using M4 mutation-specific primers along with flanking primers R2′F4 and R5 to make Products #7 and #8 (not shown) and spliced product#9 (IX). Then Products #6 and #9 are spliced together by mixing these two products together and amplifying with F1 and R5 flanking primers to synthesize Product #10 (X) with mutations M1, M2, M3, and M4. These were blunt end cloned into a single vector (
2. Cloned homozygous CFTR gene controls. Seventeen fragments 433-933 bp long each with 1 to 4 CFTR mutations were prepared using PCR primers containing the mutant sites and flanking paired primers (See
3. Intron10/Exon11 Gene region with the sites of 5 synthesized mutations illustrated. Because the G551D and R553X nucleotide substitution sites are only separated by four basepairs, these mutations were added to different cloned fragments to avoid interference when tested in the multiple formats for which the synthesized mutations were intended.
4. Multiple Intronic Basepair Substitutions Downstream of A455E and N1303K in One Clone. The normal CFTR sequence is shown in the bottom (Subject) row and the initially analyzed cloned sequence in the top (Query) row. Top: The seven nucleotide substitutions found when sequencing the first exon9/intron 9 downstream clone were the same as the substitutions found in the chromosome 20 CFTR pseudogene. All the nucleotide substitutions between the pseudogene and the active CFTR gene are indicated in bold above the top (Query) line. Bottom: The eleven nucleotide substitutions found when sequencing the first exon 21/intron 21 clone. These eleven substitutions occurred in a 130 bp repeated sequence found 162 basepairs downstream from this site.
5. Innogenetics Nitrocellulose Filter Strip Results: Each of the clones was analyzed independently for the synthesized and cloned homozygous mutations. In every case the homozygous mutation was unambiguously distinguished (f1-f8). Note the positive test signal for the homozygous 1717-1 mutation was more intense than the three homozygous signals for G542X, G551D, and R560T which are carried by the same PCR amplified fragment (f3,left). We interpret this signal intensity difference to be a characteristic of the Innogenetics multiplex PCR amplification, but the results are readily interpretable. When these 9 clones are mixed together and tested, all the tested homozygous mutant sites gave homozygous results (mix) except the G551D locus which is mutant in the top cloned fragment (
6. Entire Mixture of Homozygous Control Clones Tested on TmBiosciences Platform. Each homozygous control gives a clearly abnormal result including the 5T locus which has 98% of the fluorescence among the 5T, 7T, and 9T beads carrying specific oligonucleotides for each of these loci (Left Panel, top 3 locations). At least 85% of the signal bound to the homozygous location at all but one tested homozygous control site. Compare this to the 8 normal site results on these fragments where the TmBiosciences test assays for additional mutations (Left) and the normal DNA which gives only normal results at each tested site including the homozygous 7T mutation.
Number | Date | Country | |
---|---|---|---|
60625863 | Nov 2004 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 10236168 | Sep 2002 | US |
Child | 11074265 | Mar 2005 | US |