Human Molecular Genetics, 2000, Vol. 9, No. 18 2727-2732
© 2000 Oxford University Press
Alu-mediated PCR artifacts and the constitutional t(11;22) breakpoint
1Division of Human Genetics and Molecular Biology, The Children’s Hospital of Philadelphia, 1002 Abramson Research Center, 3516 Civic Center Boulevard and 2Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA
Received 24 July 2000; Revised and Accepted 11 September 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTS AND DISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The breakpoints of the recurrent t(11;22)(q23;q11) have recently been cloned. We identified palindromic AT-rich repeats (PATRRs) on 11q23 and 22q11 as the mechanism responsible for the rearrangement. Contradictory to our results, A.S. Hill et al. (Hum. Mol. Genet., 9, 1525–1532) suggested that Alu-mediated recombination is responsible. To clarify this discrepancy, the cloned 4.5 kb der(11) junction fragment has been completely sequenced. This sequence has been compared with that of an inverse PCR-generated der(11) junction fragment obtained by Hill et al. This reveals that the inverse PCR product has sustained a deletion between two Alu elements, such that the true breakpoint region is deleted from the PCR product. Utilizing PCR primers designed by Hill et al. to amplify across the der(11) breakpoint, we obtained a deleted PCR product even when our cloned der(11) junction fragment was used as template. Further, we find that the PCR primers that they utilized for amplification of the der(22) junction fragment are not located on the der(22). They are oriented in opposite directions within the region deleted from the der(11) PCR product, generating an artifact derived from the der(11) chromosome. Analysis of the truncated PCR products indicates a mixture of sequences from two distinct Alu elements, suggesting that the putative junction fragment described by Hill et al. is an Alu-mediated PCR artifact. These data suggest that caution should be exercised when analyzing PCR-based data, particularly when amplification is carried out in a region containing repeat structures with specific, difficult-to-amplify sequences.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTS AND DISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
PCR has proved to be a powerful tool as an alternative method to conventional library-making and cloning strategies. Many PCR-based methods for generating information about an unknown-sequence region adjacent to a known-sequence region have been established; for example, inverse PCR (1), vectorette PCR (2) or panhandle PCR (3). These methods are quite convenient, as one can avoid the difficulties of making and screening a genomic library. However, the risk of generating PCR artifacts exists, which can lead to misinterpretation of the data.
Recently, we demonstrated that palindromic AT-rich repeats (PATRRs) on chromosomes 11q23 and 22q11 are responsible for the recurrent constitutional t(11;22) (4). Hill et al. (5) reported that Alu-mediated recombination contributes to this translocation. Hill et al. obtained the der(11) junction fragment by means of inverse PCR. By sequence comparison with the normal chromosome 11 bacterial artificial chromosome (BAC), 442e11 (GenBank accession no. AC007707), they demonstrated that the breakpoint appears to be within an Alu element. Database searches with the unknown sequence adjacent to the putative breakpoint identified corresponding normal chromosome 22 sequence. Thus, they localized the breakpoint on chromosome 22 within another Alu element. They designed PCR primers flanking these Alu elements to amplify presumptive junction fragments from both the der(11) and the der(22) and confirmed that the breakpoints of all their t(11;22) samples were located in these Alu elements on both chromosomes. Their cloning strategy seemed elegant and appeared to have no inherent flaws. Nonetheless, their results seemed to completely contradict our findings. PATRRs are rare in the human genome and the presence of two of them, one on each of the involved chromosomes, was suggestive of a novel mechanism to explain the recurrent t(11;22). In contrast, although Alu elements are abundant in the human genome, there are a limited number of reports which demonstrate human chromosomal aberrations caused by Alu–Alu recombination (6). Furthermore, almost all of the reported germline rearrangements involving Alu elements appear to be intrachromosomal rather than interchromosomal events. Transfection studies designed to measure the recombination potential between DNA sequences also demonstrate that the recombination rate between two Alu elements is indistinguishable from the frequency observed for a control DNA sequence (7). Thus, the hypothesis that Alu–Alu recombination might be the mechanism for this site-specific, recurrent translocation appeared open to question and worthy of further investigation.
In contrast, we isolated the der(11) junction fragment by a more traditional phage-based cloning method. A rearranged band for the der(11) was identified in HindIII-digested DNA from a t(11;22) carrier. We constructed a genomic library with HindIII-digested DNA and obtained a 4.5 kb der(11) junction fragment. The breakpoint region of the 4.5 kb phage clone was sequenced and sequence comparison with BAC 442e11 revealed that the t(11;22) breakpoint resides within a PATRR. In subsequent experiments, the translocation breakpoints of 40 independent t(11;22) carriers have been analyzed and all of the breakpoints localize within the same PATRR (8). Amongst these 40 additional cases are several cell lines [GM04403, GM06229 and GM03372 (carrier parent of GM03371)] used by Hill et al. (5) in their experiments.
In this study, we have further analyzed the t(11;22) breakpoint region using their published PCR method. The data presented here demonstrate that the junction fragments that they obtained by inverse and traditional PCR result from Alu-mediated PCR artifacts. This report highlights the danger of utilizing only a PCR-based strategy for genome analysis and suggests that caution be exercised in the interpretation of data derived from such methods.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTS AND DISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In our previous study, the 4.5 kb der(11) junction fragment from one t(11;22) balanced carrier was isolated by hybridization screening of a phage genomic library (4). This cloned der(11) junction fragment has been completely sequenced (GenBank accession no. AF288053) and compared with the junction fragment reported by Hill et al. (5) (GenBank accession no. AF226670). Within the sequence of the cloned der(11) junction fragment, we have identified both of the Alu elements in which, according to Hill et al. (5), the t(11;22) breakpoints are located (Fig. 1a). Situated between these Alu elements shown in Figure 1a is the AT-rich region derived from fusion of the two PATRRs located on chromosomes 11 and 22. We have previously reported that all the t(11;22) breakpoints cluster in this AT-rich region (4,8). In the sequence of the cloned der(11) junction fragment, we have identified all of the primer sequences which Hill et al. used for amplification of the der(11) junction fragment in their experiments (Table 1). Their der(11) PCR data indicate that they generate a 544 bp product that includes the der(11) translocation breakpoint. However, based on the sequence of the cloned der(11) junction fragment, the actual distance between their PCR primers is 2.4 kb.
|
|
We performed the PCR experiment that they reported and obtained the same 544 bp PCR products for the der(11) not only from the cell line that they used in their study but also from the t(11;22) carrier that we used for junction fragment cloning (Fig. 2a). Normal controls did not yield PCR products, indicating that the products are specific for the t(11;22). Interestingly, we also amplified the same 544 bp PCR product from the 4.5 kb cloned junction fragment, even though the actual distance between the primers is significantly greater. This PCR product was directly sequenced. Analysis of the sequence of the 544 bp PCR product demonstrates that there is a deletion between the two Alu elements and that the real breakpoint is not present in the sequence (Fig. 1b). The two Alu elements belong to different Alu subfamilies, AluY and AluSq, and they have numerous nucleotide differences between them. The sequence of the PCR product demonstrates a mixture of sequences in the Alu element. Approximately half of the PCR products have the Alu sequence from chromosome 11 and the other half have the Alu sequence from chromosome 22 (Fig. 3). Taken together, these results lead us to conclude that the PCR product is an Alu-mediated PCR artifact.
|
|
During the elongation phase of PCR, DNA synthesis that begins from one primer sometimes stops before it reaches the second primer. During the next annealing cycle, the prematurely terminated PCR products usually anneal to the complementary strand of a full-length PCR product and DNA synthesis resumes again. If there are two Alu elements in the same orientation between the primers, different Alu elements may anneal to one another because of their high nucleotide identity (Fig. 4a). At the next elongation step, a truncated PCR product which deletes between the two Alu elements will be generated (9). Once the deletion occurs, a truncated PCR product is more likely to be amplified during the PCR because it is shorter than the authentic product. Furthermore, in the t(11;22) breakpoint region, the authentic PCR product would have a long AT-rich repeat between the Alu elements which might further inhibit the generation of authentic product. This increases the likelihood of amplification of the shorter artifact.
|
Hill et al. (5) isolated the putative der(11) junction fragment by inverse PCR using EcoRI-digested and self-ligated DNA from a t(11;22) balanced carrier as template. In their report the size of the inverse PCR product is 1.8 kb. However, according to our sequence data, the size of bona fide product should be 3.8 kb. We did not reproduce their inverse PCR method. However, we presume that the 2 kb region between the two Alu elements which includes the authentic breakpoint region is also deleted from the inverse PCR product. Thus, it too is likely to be a PCR artifact generated by a similar mechanism (Fig. 1a).
With regard to the der(22) junction fragment, the Hill et al. PCR reaction should be incapable of producing the desired result, because the primers used to amplify the junction fragment are not actually located on the der(22). The selection of these primers came from their comparison with the sequence from the inverse PCR-generated der(11) junction fragment to the existing normal genomic sequence in the databases. Thus, they are located on the der(11) within the region deleted from the der(11) PCR product (Fig. 1b). In addition, our analysis demonstrates that they are oriented facing away from one another, as they were perceived to be located on the der(22). Nonetheless, in our laboratory we also observed that these primers yield a 572 bp PCR product from the DNA of t(11;22) balanced carriers (Fig. 2b). Since normal controls do not yield the 572 bp product (Fig. 2b), this putative ‘der(22)’ PCR product appears to be translocation specific. Additional data shown in Figure 2b demonstrate that DNA from a somatic cell hybrid, which contains as its only relevant chromosome the der(22), as well as DNA from a patient with the supernumerary-der(22) syndrome does not yield a PCR product, while the cloned der(11) junction fragment used as template does yield this 572 bp PCR product. Therefore, as would be predicted from the sequence data, the putative ‘der(22)’ PCR product actually arises from the der(11). It is of interest that Hill et al. (5) did not report der(22) PCR results with the flow-sorted der(22) chromosome or the supernumerary der(22)- containing cell line GM06228. This is true despite the fact that they used the flow-sorted der(22) DNA and this cell line for PCR and Southern hybridization experiments while mapping the breakpoint. Such studies would have indicated that the 572 bp PCR product was not derived from the der(22).
Sequence analysis of the ‘der(22)’ PCR product yields results similar to those from the 544 bp PCR product indicating a mixture of Alu-derived products (data not shown). The regions immediately adjacent to both primers are also Alu elements, suggesting that the 572 bp product is also a PCR artifact that arises through this Alu-mediated mechanism. Thus, even when the PCR products are synthesized in opposite directions, after extending up to the Alu and terminating prematurely, they can anneal to one another in the next cycle and generate this type of PCR artifact (Fig. 4b). We do not know why normal samples or supernumerary der(22) samples do not yield such an artifact. In a sample that contains the der(11), these primers will anneal in opposite orientation but on a single der(11) molecule. Newly synthesized product will be located in close proximity in the PCR reaction solution and perhaps this facilitates generation of this type of artifact in a sample which contains the der(11).
This experience should serve as a warning for those attempting analysis of an unknown sequence region by PCR. In the case presented here, although the Alu elements share a maximum of only 84% identity within 100 bp, such an artifact was easily generated. Numerous cycles of nested PCR may further contribute to artifact formation. Our experience was similar in that unexpectedly short PCR products were generated when we analyzed the t(11;22) breakpoint region by nested PCR using primers designed in our experiments (8). Additional analysis proved that they were truncated PCR products which had a deletion within the AT-rich region. Even when both of the sequences flanking the primers are correct, the PCR product may contain an artifact. Thus, PCR-based analysis of non-human primates, whose genomes also harbor Alu elements, would pose similar problems. Likewise, this would be true for any other organism whose DNA contains numerous interspersed repeat elements. These data suggest that care should be exercised when analyzing PCR-based data, particularly when analysis is carried out in an unknown region. This is especially noteworthy when the PCR product contains any type of repeat structure and where there exists a specific, difficult-to-amplify DNA sequence. This is certainly the case for the PATRRs involved in the t(11;22).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTS AND DISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The 4.5 kb HindIII der(11) junction fragment was obtained from the genomic library made from a t(11;22) balanced carrier (4). The cloned insert was completely sequenced by primer walking. Sequence was analyzed using RepeatMasker2 (http://ftp.genome.washington.edu/cgi-bin/RepeatMasker ) and ClustalW (http://www2.ebi.ac.uk/clustalw/ ).
The primers for nested PCR were generated according to a recent publication (5). PCR conditions for the first PCR reaction were 35 cycles of 94°C for 30 s, 66°C [for der(11)] or 62°C [for the putative ‘der(22)’] for 30 s and 72°C for 1 min. The conditions for the second PCR reaction were 35 cycles of 94°C for 30 s, 64°C for 30 s and 72°C for 1 min. PCR products were purified with the aid of the QIAquick PCR purification kit (Qiagen, Valencia, CA) and then sequenced directly utilizing the PCR primers on both strands.
The t(11;22) carriers and the somatic cell hybrid with the der(22) used for breakpoint analysis have been previously described (4,8). GM03372, GM04403 and GM06229 were obtained from the Coriell Mutant Cell Repositories (Camden, NJ).
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank Prescott L. Deininger for helpful discussions and Stephanie St Pierre and the t(11;22) Together Network (http://www.nt.net/~a815/index.html ) for assistance in obtaining patient samples for this study. These studies were supported in part by CA39926 (BSE), DC02027 (BSE) and HD26979 from the NIH.
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +1 215 590 3856; Fax: +1 215 590 3764; Email: beverly@mail.med.upenn.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTS AND DISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Triglia, T., Peterson, M.G. and Kemp, D.J. (1988) A procedure for in vitro amplification of DNA segments that lie outside the boundaries of known sequences. Nucleic Acids Res., 16, 8186.
2 Riley, J., Butler, R., Ogilvie, D., Finniear, R., Jenner, D., Powell, S., Anand, R., Smith, J.C. and Markham, A.F. (1990) A novel, rapid method for the isolation of terminal sequences from yeast artificial chromosome (YAC) clones. Nucleic Acids Res., 18, 2887–2890.
3 Jones, D.H. and Winistorfer, S.C. (1992) Sequence specific generation of a DNA panhandle permits PCR amplification of unknown flanking DNA. Nucleic Acids Res., 20, 595–600.
4 Kurahashi, H., Shaikh, T.H., Hu, P., Roe, B.A., Emanuel, B.S. and Budarf, M.L. (2000) Regions of genomic instability on 22q11 and 11q23 as the etiology for the recurrent constitutional t(11;22). Hum. Mol. Genet., 9, 1665–1670.
5 Hill, A.S., Foot, N.J., Chaplin, T.L. and Young, B.D. (2000) The most frequent constitutional translocation in humans, the t(11;22)(q23;q11) is due to a highly specific Alu-mediated recombination. Hum. Mol. Genet., 9, 1525–1532.
6 Deininger, P.L. and Batzer, M.A. (1999) Alu repeats and human disease. Mol. Genet. Metab., 67, 183–193.[ISI][Medline]
7 Shen, M.R. and Deininger P.L. (1992) An in vivo assay for measuring the recombination potential between DNA sequences in mammalian cells. Anal. Biochem., 205, 83–89.[ISI][Medline]
8 Kurahashi, H., Shaikh, T.H., Zackai, E.H., Celle, L., Driscoll, D.A., Budarf, M.L. and Emanuel, B.S. (2000) Tightly clustered 11q23 and 22q11 breakpoints permit PCR based detection of the recurrent constitutional t(11;22). Am. J. Hum. Genet., 67, 763–767.[ISI][Medline]
9 Ji, W., Zhang, X.Y., Warshamana, G.S., Qu, G.Z. and Ehrlich, M. (1994) Effect of internal direct and inverted Alu repeat sequences on PCR. PCR Methods Appl., 4, 109–116.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Nakayama and T. Ishida Alu-mediated 100-kb deletion in the primate genome: The loss of the agouti signaling protein gene in the lesser apes. Genome Res., April 1, 2006; 16(4): 485 - 490. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||