Full Text (PDF)
Evolutionary biology has been revolutionized by the application of molecular biology techniques, many of which require the extraction of DNA from biological specimens. Typical procedures for DNA extraction depend on chemical or physical treatments to release DNA from cells, followed by a clean-up step to purify the DNA from unwanted cellular components prior to any downstream analysis. However, since current DNA polymerases used in PCR are so robust, S. Shadi, G. Singer, and M. Hajibabaei at the University of Guelph (Ontario, Canada) hypothesized that it might be possible to obtain PCR-amplifiable DNA from the preservative solution—usually ethanolic— surrounding biological specimens. If true, this would bypass the need for standard DNA extraction, which would be particularly useful for non-invasive analyses of rare or type specimens, or in situations where the specimen has been completely consumed in previous analyses and only the preservative medium remains. To test their hypothesis, the authors initially applied their idea to mescal, the Mexican liquor which often contains a “worm” (actually the caterpillar of the agave butterfly, Hypopta agavis). The mescal was evaporated and its residue dissolved in water which was then purified on a spin column to remove PCR inhibitors. PCR was carried out to amplify 130-bp and 658-bp fragments from the cytochrome C oxidase 1 (CO1) gene, which were then sequenced. The 130-bp CO1 fragment was successfully amplified from mescal and its sequence matched that of the amplicon from DNA directly extracted from the caterpillar tissue. The authors found this result encouraging and also somewhat surprising since mescal is a less-than-ideal preservative medium. The larger 658-bp fragment, however, did not amplify from mescal. Using the same technique—but without spin column purification—on various freshly collected and archival specimens of plants and insects stored at room temperature in 95% ethanol, the authors successfully PCR-amplified and sequenced additional gene fragments from nearly all samples, even for larger amplicons such as the 658-bp CO1 fragment. Their results demonstrate that sufficient DNA is released from specimens into ethanol storage medium for PCR amplification without the need for invasive DNA extraction from a specimen's tissue.
See “Direct PCR amplification and sequencing of specimens' DNA from preservative ethanol”.
Man or Mouse?Human cancer xenografts maintained in immunodeficient mice greatly facilitate the genetic analysis of human tumors. During the propagation of these xenografts in mice, however, mouse stromal cells tend to replace human stromal cells, resulting in mixed-species samples. This can complicate genetic analysis of the xenografts, particularly efforts to perform deep sequencing of human tumor genomes. In this issue, M.-T. Lin, J. Eshleman, and colleagues at The Johns Hopkins University School of Medicine (Baltimore, MD) describe their simple and robust molecular approach to quantifying the relative contributions of mouse and human DNA in mixed DNA samples obtained from mouse xenografts. Their assay uses PCR to detect length variations between particular homologous human and mouse genes (i.e., orthologs) that are due to insertions and/or deletions. By using the appropriate primer pair designed to bind conserved regions flanking the insertion/deletion, two PCR amplicons with species-specific lengths are generated and then subjected to capillary electrophoresis, which can sensitively and accurately resolve even small size differences. The percentage of mouse DNA can then be calculated from the relative peak heights of the mouse- and human-specific PCR products in the electropherogram. Initially, the authors tested primer pairs designed to bind 12 loci that have species-specific length differences on a mixture of 20% mouse DNA and found that only three of these pairs provided accurate peak height ratios. These three pairs were next tested on mixed DNA samples containing differing amounts of mouse DNA, demonstrating very accurate measurement of the mouse DNA ratio. When further tested on a set of 93 human pancreatic cancer xenografts, the percentage of mouse DNA determined by each of the three primer pairs for each xenograft was very similar in the vast majority of cases. The technique was also found to be useful in monitoring mouse cell contamination during the establishment of human pancreatic cell lines from explanted mouse xenografts. The authors also identified another 370 potential loci that should be compatible with their assay.
Feeding RNAiRNAi is frequently used in reverse genetics studies in multiple cell types and organisms. In Caenorhabditis elegans, the model in which RNAi was initially uncovered, inhibition of gene expression can be accomplished by delivering double-stranded RNA into any part of the worm, since RNA is transported across cell boundaries in this organism. Thus RNAi can be accomplished in worms by feeding them bacteria expressing the dsRNA, soaking them in solutions containing dsRNA, or injecting them with RNA directly. Feeding is the preferred method for single gene RNAi since it is convenient and inexpensive. But knock-down of multiple genes using the feeding approaches can produce poor results; worms fed a mixture of two types of bacteria—each expressing dsRNA targeting a different gene—often show incomplete phenotypes. In this issue of BioTechniques, K. Min, J. Kang, and J. Lee at Seoul National University (Seoul, Korea) present data showing that dsRNA from separate bacteria fed to worms at the same time results in a dilution effect that reduces the manifestation of the expected phenotype for some genes. Their simple solution to this problem was to prepare bacteria expressing dsRNA for both genes of interest using a single vector, rather than feeding two different bacteria, thus allowing the same amount of dsRNA for both genes to be delivered to the animal. Using this method, the authors were able to successfully knock down endogenous and episomal genes with improved efficiency compared to experiments where two different bacteria were fed to the worms. Even using RNAi constructs for three sequences, the percentage of worms showing complete phenotypes for all three genes was greater than experiments using separate bacteria. The orientation and order of the inserts did not impact the efficiency of inhibition. This new approach should serve as a useful alternative when performing RNAi of more than one gene in C. elegans.
Array Variation
Several options are available for analyzing miRNA expression including high-throughput DNA sequencing–based digital gene expression (DGE) profiling, real-time qPCR, and microarrays. Each method has its advantages and disadvantages, yet no consensus exists regarding which technology produces the most accurate results. It was recently shown that DGE profiling is biased toward certain small RNAs (Linsen et al. 2009. Nat. Methods 6:474–476), indicating that the method is appropriate for differential expression analysis, but not absolute quantification. Biases were also seen in quantitative reverse transcription PCR and shown to be dependent on library preparation. Microarray techniques are cost-effective and simple to use, but are generally considered to be less accurate and quantitative than the other platforms. S. Pradervand and colleagues at the University of Lausanne (Lausanne, Switzerland) compared miRNA expression patterns in brain and heart samples using Affymetrix, Agilent, and Illumina array platforms, and qPCR and DGE. The authors found that expression results from DGE and qPCR were highly correlated and consistent, but the number of miRNAs identified as differentially expressed, as well as the fold change levels determined, varied when using the microarray platforms. Agilent microarrays had the best correlation with qPCR results, while Illumina and Affymetrix varied considerably. Illumina microarrays showed a strong fold change compression—most likely caused by the PCR amplification step needed for target preparation, according to the authors—while Affymetrix arrays returned false negative data for several miRNAs for which they lacked sensitivity. To investigate why the miRNAs were not detected by the Affymetrix microarrays, the authors compared the GC content of the false negative miRNAs with those that were identified as differentially expressed on the array, DGE, and qPCR. They found that for the Affymetrix platform, the GC content of false negatives was significantly lower than the GC content of the true positives. Agilent and Illumina array probes are designed to balance the melting temperature by adding a guanine nucleotide to the 5′ end of each probe or shortening probe length. Hierarchical clustering analysis showed more dissimilarity among the three microarray platforms than among DGE, qPCR, and the Agilent microarray. The results indicate that factors such as microarray probe design, target preparation, or hybridization stringency are more responsible for variation than the type of technology used.
