Since its introduction in the late 1500s, microscopy has been an essential tool for examining cell morphology and biomarker distribution. Microscopy, similar to the trend among many other techniques, has advanced towards high-throughput capabilities, with current microscopes capable of collecting thousands of high-resolution images per day. Along with these throughput advances have come the challenges associated with analyzing such large volumes of data. High content screening enables data collection from multiple cellular parameters—such as size, cell cycle stage, and intensity and localization of fluorescently labeled components—and is increasing in popularity for therapeutic drug discovery and functional genomics applications. Until now, methods available for interpreting the data acquired from these images relied upon parametric statistical tools such as principal component analysis and clustering, all of which assume a normal data distribution—a rarity in the cellular world. When incorrect, this assumption can lead to inadequate interpretation of the data. Reporting in this issue, J. Gorenstein and colleagues at Merck Research Laboratories describe a method using Kolmogorov-Smirnov (KS) nonparametric statistical analysis to enable unbiased interpretation of high-content image data. KS analysis determines the maximal difference for a single parameter between two populations without requiring assumptions regarding distribution of the data. This results in equally weighted dimensionless scores for each parameter that can be combined to generate a cumulative score. The use of cumulative scores allows for simple comparisons of multidimensional cellular changes between test and control conditions. The authors demonstrate their method by scoring phenotypic changes induced during a small molecule screen and with a subgenome-wide siRNA screen, showing that this method of analysis enables quantitative screening for mechanistic changes in cell phenotypes which can assist in target identification and prioritization when applied to compound screens.
Of Mice and Men
Inflammatory bowel disease (IBD), including ulcerative colitis and Crohn's disease, is characterized by chronic swelling and tissue damage in the gastrointestinal tract. Treatments exist to ease symptoms of the disease, but currently there is no cure and the precise cause of the disease is unknown. In order to develop new treatments, the proper tools must be made available to researchers. Mouse models where colitis is induced through the use of trinitrobenzene sulfonic acid (TNBS) or oxazolone haptens have proven useful for the study of IBD. TNBS and oxazolone haptens are believed to induce colitis by binding endogenous proteins in the colonic mucosa and inducing a local immune response through activation of macrophages and T cells. However, researchers have been unable to confirm this since the proteinbound chemical compounds are indistinguishable from free chemicals in the intestine. In this issue, K. Ishiguro and colleagues at Nagoya University (Aichi, Japan) describe a new mouse model for investigating the mechanisms responsible for IBD symptoms and evaluating potential therapeutics. The authors induced colitis by intrarectal administration of 4-chloro-7-nitro-2,1,3-benzoxadiazole (NBD-Cl), a compound that fluoresces upon binding to endogenous proteins, thereby allowing observation of hapten-protein formations. Visualization confirmed that NBD-protein conjugates formed and were by macrophages, resulting in inflammatory cell infiltration and crypt-epithelium destruction within 2 days of NBD-Cl delivery. T-cell involvement was also confirmed in controlled experiments using BALB/c wild-type and thymus-deficient nude mice. With the ability to visualize the NBD-proteins in the colonic mucosa, researchers will now be able to investigate inflammatory reactions and use this resource for screening experimental interventions for colitis when hapten-protein formation is confirmed.
Microbial mats are exemplary model systems for metagenomic analysis of complex interactions among diverse microbes. But unlike soil or sediment samples, the efficient extraction of DNA from hypersaline microbial mats is hindered by high levels of salts and extracellular polymeric substances (EPSs). While current methods for DNA extraction from these samples are effective, they are also harsh, resulting in extracted DNA that is too small for large-insert cloning and which may not accurately represent the diversity of microbes in the sample. In this issue, R.S. Norman and colleagues at the University of South Carolina, (Columbia, SC) describe their modified DNA extraction method from microbes in hypersaline mats that overcomes the shortcomings of other techniques. The authors compared their modified freeze-thaw/polyethylene glycol (PEG) precipitation method with three other standard methods: (i) freeze-thaw/lysozyme (LYS/FT), (ii) freeze-thaw (FT), and (iii) bead beating (BB) using a commercial kit. The critical first step for all of these methods, even prior to DNA extraction, is to separate microbes from the matrix of the mats, which is the major source of EPS contamination that hinders downstream applications such as PCR. The PEG method yielded both the purest and highest amount of DNA, while BB was the poorest for both of those criteria. High molecular weight DNA was obtained from the PEG, LYS/FT, and FT methods, in contrast to BB, which produced lower molecular weight, sheared DNA. For all methods, extracted DNAs were of high enough quality for PCR amplification of bacterial 16S rRNA genes. 16S rRNA gene analysis using denaturing gradient gel electrophoresis demonstrated that DNA extracted using the PEG method, when compared to DNA extracted directly from the remaining cells in the sediment left after cell separation, faithfully represented the microbial diversity originally present in the mat. Based on sequencing of pooled archaeal and bacterial 16S rRNA gene amplicons, the taxonomic diversity of DNA isolated using the PEG method was higher than that from the BB method. Therefore, this improved method for DNA extraction from microbial mats provides high quantities of pure, high molecular weight DNA that accurately represents the taxonomic diversity of the sample and is suitable for large-insert cloning and metagenomic analysis. And for those working on microbial communities from other locations, authors suggest their method should be applicable to other similarly recalcitrant samples.
Incorporation of nonstandard amino acids into proteins is necessary for structural analysis using different spectroscopic techniques. Both the incorporation of the unusual amino acid and the overall yield of the target protein have to be maximized, which is often difficult to achieve. Typically, the unusual amino acid is incorporated through overexpression of the protein in the corresponding auxotrophic Escherichia coli strain grown in minimal medium supplemented with the amino acid analog, but this can result in low protein yields. Addition of the natural amino acid to increase protein yield decreases incorporation of the analog. One approach to introduce aromatic amino acid analogs into proteins expressed in E. coli that doesn't require auxotrophic strains is through the use of N-(phosphonomethyl)glycine (glyphosate), an herbicide that inhibits biosynthesis of aromatic amino acids. Cell growth in the presence of glyphosate requires media supplemented with all three aromatic amino acids (phenylalanine, tyrosine, and tryptophan). Substitution with the desired aromatic amino acid analog leads to its incorporation into the overexpressed protein. However, when this method was used to introduce fluorotryptophan into proteins for NMR studies, the yields were modest and incorporation only reached ~80% because tryptophan had to be added to allow adequate cell growth. To greatly increase incorporation of fluorinelabeled amino acids and protein yield, M. Neerathilingam and J. Markley at the University of Wisconsin-Madison have combined the glyphosate method with the use of auto-induction medium. This type of medium enables E. coli growth to high density followed by automatic induction of protein expression from lac promoter constructs without the need for IPTG. The use of glyphosate with auto-induction medium enabled the substitution of 6-fluorotryptophan at all eight tryptophan residues in maltose binding protein with 99.3% incorporation and a high protein yield.