Cleaning Microsatellite PCR Products with Sephadex™ in 96-Well Filtration Plates Enhances Genotyping Quality

Automatic sequencing machines have greatly improved our ability to accumulate nucleotide sequence data. To enhance efficiency, sequencing reactions are routinely subjected to gel filtration through a porous particulate matrix prior to loading to remove salts, unincorporated nucleotides, and dye terminators that can interfere with electrophoretic dynamics and laser function. Such filtering is done either by column (1) or use of 96-well microplates for greater throughput (2).

Microsatellite genotyping uses the same automatic technology to assess polymorphism. Interestingly, however, there is no corresponding general tendency to clean microsatellite PCR products prior to genotyping, even though filtration would likely improve genotyping efficiency. Some workers do use Sephadex™ columns (Amersham Biosciences, Piscataway, NJ, USA) to clean PCR products before genotyping, especially when using fluorescently labeled dNTPs (e.g., http://www.mnh.si.edu/GeneticsLab/TechnicalPage/Protocols/Microsat/MicroWorkshop.pdf). Unfortunately, this technique is not widely known, it does not apply to fluorescently labeled primers, and does not favor the high-throughput systems in use today.

Our own efforts to run unfiltered multiplexed microsatellite PCR products on a BaseStation™ DNA Fragment Analyzer (MJ Research, Reno, NV, USA) are often frustrating. Many gels exhibit an inward bending of lanes, especially in the outer regions, due to excess salts interfering with electrophoresis. This makes it difficult to identify lanes, thus increasing the likelihood of improperly assigning fragments to neighboring individuals. Furthermore, the analysis software is often unable to accurately detect and/or size fragments due to interference from excess signal associated with unincorporated fluorescent dyes (Figure 1A). We therefore tested whether inclusion of a filtration step, analogous to what is routinely used in automatic sequencing, would improve our ability to analyze microsatellite genotypes without sacrificing throughput.

Figure 1. Effect of Sephadex filtration on the quality of gel-based analysis of microsatellite PCR products.

Shown are two 96-lane gels of the same microsatellite reactions (A) without filtration and (B) following Sephadex filtration. Each lane represents a multiplex of six microsatellite loci. Amplifications were performed as described in Reference (3) using DNA from collared lizards (Crotaphytus collaris).

We subjected our PCR products to a slightly modified version of a Sephadex filtration protocol (www.millipore.com/publications.nsf/docs/tn053) originally designed to remove salts and unicorporated dyes from sequencing products using MultiScreen® 96-well filtration system (Millipore, Billerica, MA, USA). Filters were prepared by pouring dry Sephadex G-50 Fine DNA Grade powder (Amersham Biosciences) onto a MultiScreen Column Loader (Millipore). A MultiScreen HV plate (Millipore) was placed upside-down on the column loader, and both were inverted to transfer the Sephadex to the wells of the MultiScreen HV plate with tapping. We added 300 µL of distilled water to each well and allowed the slurry to incubate at room temperature for 1 h (the original protocol specifies a 3-h incubation). The MultiScreen HV plate was then placed on top of a standard 96-well microplate and spun at 550× g for 2 min (the original protocol specifies 910× g for 5 min) on a Marathon 21000 centrifuge (Fisher Scientific, Hampton, NH, USA) using a two-tray rotor to remove excess water from the columns. The excess water was discarded.

Multiple PCR products were run simultaneously by combining 10 µL from six separate PCRs, each of which targeted a distinct microsatellite locus. We loaded 20 µL of the multiplexed mixture directly into the center of each Sephadex well. The MultiScreen HV plate was placed on top of a clean 96-well V-bottom microplate (USA Scientific, Ocala, FL, USA) and spun at 550× g for 2 min to move the products through the Sephadex. This filtered supernatant was loaded directly onto the BaseStation.

The results were dramatic (Figure 1). We observed dozens of such filtrations, and in each case, the lanes ran straight with little background to interfere with the accurate detection of fragments. The resulting gels were much easier to score since lanes were easily identified, all bands were properly assigned to the appropriate individual, and the laser was better able to accurately identify and size fragments. Examination of the traces showed that while the brightest products on the uncleaned gels were not necessarily improved by filtration, the signal-to-noise ratio improved for all other products—most by a factor of from two to five. Because microsatellite amplifications are often analyzed by capillary electrophoresis, we also examined the effect of our protocol on traces generated using an ABI Prism® 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). DNA from two individuals was multiplexed for three microsatellite loci with cleaned and uncleaned aliquots subjected to capillary analysis. The resulting traces showed signal-to-noise ratios for cleaned traces as much as 3.5 times higher than the uncleaned traces, suggesting the clean-up procedure works equally well on gel-based and capillary-based sequencing machines.

This protocol represents a relatively minimal investment in time and money, yet the benefit of being able to more reliably read data is invaluable. The filtration takes less than 2 h, and aside from an initial investment of approximately $240 for the MultiScreen Column Loader and the MultiScreen HV plate, the technique cost less than 10 cents per lane. We suggest this procedure be used as routinely with microsatellite genotyping as it is for nucleotide sequencing.

Competing Interests Statement

The authors declare no competing interests.

Acknowledgments

This work was performed in the laboratory of Alan R. Templeton at Washington University in St. Louis during a sabbatical leave from Whitman College for D.H.W. Financial support was provided by a National Science Foundation (NSF) Research Opportunity Award (no. DEB-0108874-002) and a grant from the Salzman-Medica Foundation to D.W.H. We also thank Keith Crandall, Rebecca Scholl, and Jennifer Neuwald for help with the capillary analysis.

Additional information

Funding

This work was performed in the laboratory of Alan R. Templeton at Washington University in St. Louis during a sabbatical leave from Whitman College for D.H.W. Financial support was provided by a National Science Foundation (NSF) Research Opportunity Award (no. DEB-0108874-002) and a grant from the Salzman-Medica Foundation to D.W.H. We also thank Keith Crandall, Rebecca Scholl, and Jennifer Neuwald for help with the capillary analysis.