Protein arrays are becoming ubiquitous in research laboratories worldwide, as they provide a simple and high-throughput solution for proteomics research. Although commercial arrays are now becoming available, many scientists still prefer to construct custom arrays in order to have more control and flexibility in their experimental design. In any array system, spot identification is an important issue, but it is even more so with custom systems due to the additional variables introduced by their small-scale production. In an attempt to get a handle on some of the issues the may affect spot detection, Smith et al. describe the use of a median filter algorithm to distinguish spot features from the background using a titration method. This filter has previously proven to be useful for removing noise from images; it works by replacing any one pixel's value with a median measurement derived from surround pixels. Innate to the mathematical process is the ability to determine how much the spot blends with its surroundings (i.e., the background), making it appropriate for determining threshold detection levels. The median filter method for classifying spots as either present or absent, when compared with a human worker, showed equal or better performance while at the same time being compatible with high-throughput analysis. -Page 74
Chew on ThisExonucleases may be the bane of many experiments, but their nucleic acid nibbling does have a useful side. Last year, Wang et al. described an exonuclease protection assay for antibody-free measurement of NF-κB. The assay detected protein levels by gauging the extent to which a FRET-dsDNA probe containing an NF-κB binding site was protected from exonuclease III by specific binding of the transcription factor. Although elegant, the method was relatively expensive and did not lend itself to scale-up. As a result of these concerns, the group has devised an appealing new strategy for assessing DNA-protein binding. The procedure first involves coupling a DIG-labeled dsDNA probe to the surface of a plate. Then, the nuclear extract to be assayed for the transcription factor of interest is incubated with the DNA probe. After a suitable period for binding, unbound factors are washed away and exonuclease III is added. Following digestion and further wash steps, colorimetric detection of the DIG label occurs. The intensity of the signal is proportional to the amount of the DNA-binding protein being studied, as the presence of the protein prevents degradation of the DIG-containing probe. The assay, which was also tested with AP1, offers the potential for parallel analysis of multiple transcription factors in a dsDNA microarray format. -Page 79
RNA silencing strategies have proved to be a potent means to accumulate functional genomics data. First-generation shRNA expression vectors often use the U6 snRNA promoter, driven by pol III. Though effective, pol III promoters offer little flexibility and have not yielded particularly promising results in transgenic animal studies. More recently, the miR-30 microRNA has been used as a platform for shRNA expression—these transcripts are also driven by pol II and can mediate high-efficiency knockdown. The fact that natural miRNA transcripts can contain multiple hairpins inspired two studies published in this issue. Both Xia et al. and Sun et al. investigated the feasibility of simultaneous gene knockdown by miRNA-based transcripts bearing multiple shRNAs. In the Xia et al. study, a vector driven by a tetracycline-regulatable ubiqui-tin C promoter was employed. The authors demonstrate that they can knock down up to three target genes from their doxycycline-triggered multi-hairpin construct. Another appealing approach is deftly demonstrated by Sun et al. This group used the ever-popular CMV promoter in a lentiviral vector context. Multimers of a given shRNA enhance silencing; moreover, expression of two distinct shRNAs achieves extremely efficient knockdown of separate target genes. Together, these strategies should permit coordinated silencing of multiple targets, and promise to simplify the tasks of deciphering signaling pathways and uncovering synergistic or antagonistic crosstalk among proteins under investigation. -Pages 59 and 64
Maximum Mobility
Gene regulation, through the binding of transcription factors to specific noncoding regions that flank the coding sequences, allows for exquisite control of all manner of cellular functions. Since completion of the human genome sequencing project, there has been a concerted effort to identify transcription factor binding sequences, as well as the specific proteins that bind to them. The electrophoretic mobility shift assay (EMSA; also known colloquially as a gel shift assay), is just one technique developed to identify these sequences. It is based on the simple observation that DNA to which protein is bound moves more slowly through a gel than the same DNA sans protein, and has been developed into a powerful and informative technique. In a further development of this methodology, Chernov et al. describe the use of their so-called 2D-EMSA. In this two-step procedure, a regular EMSA gel is first run under nondenaturing conditions. Following localization by autoradiography, the portion of the gel containing the radioactively labeled DNA is excised, and the DNA:protein complexes disassociated by SDS treatment. The resultant gel containing only the DNA fragments is then run in the second dimension. Since DNA not bound by protein will have the same mobility in both dimensions, they will form a diagonal line, while DNA that formed complexes retarded in the first dimension will be off the diagonal and can be easily identified and cloned for further analysis. -Page 90
The Return of the NativeSome of the most widely used ORF resource clones dispense with native termination codons. This makes perfect sense if the goal is to permit rapid exchange between various expression vectors, since the initiation and termination codons can be provided by the plasmid backbone in a wide variety of contexts—including N-terminal and/or C-terminal fusions. However, this flexibility comes at the cost of being unable to express the protein in its true native form. This limitation can be a showstopper; for instance, C-terminal ends of receptor tyrosine kinases cannot have extraneous nonnative residues if accurate signaling analysis is to be performed. Itoh et al. describe now how to have the best of both worlds. In their ORFeome vector strategy, the native termination codon is present but easily removable. The trick involves the use of a type-IIS restriction enzyme: thanks to strategic placement of the recognition site, the termination codon can be excised by a digestion/self-liga-tion procedure. Because the strategy does not rely upon PCR, expensive and time-consuming resequencing is minimized. Importantly, users of this modified entry clone sacrifice nothing in convenience, since the vector provides the choice of standard recombinational cloning or homing endonuclease-mediated recombination for vector exchange reactions. -Page 44
QC with TdT
Quality control of microarrays is essential for users to have confidence in the data produced. Artifacts created during microarray fabrication (for example, holes, doughnuts, and speckles) can cause multiple miscalls and can confound results. Therefore, determining spot quality before investing time, expense, and often precious samples in a full experiment, is both reasonable and prudent. Microarray QC is usually carried out using short, random-labeled oligonucleotide probes or by nonspecific labeling of targets using, for example, SYBR® Green. Each technique has its drawbacks, including high background often seen with the nonspecific dyes, and limitations of both techniques when used on short probe arrays. Guerra has come up with an alternate method that utilizes terminal deoxynucleo-tidyl transferase (TdT), the same enzyme used in the popular TUNEL apoptosis assay. TdT is adept at the addition of fluor-labeled nucleotides to the 3′ end of nucleic acid strands, but is sensitive to both the nucleo-tide added and its attached fluorescent molecule. The author determined that optimum labeling activity was achieved using Cy™5-UTP (incorporation of NTPs, as for ddNTPs, terminates further elongation activity), and that terminal end labeling with TdT was a robust and effective predictor of spot failure. The technique is adaptable to any DNA/RNA microarray in which the target molecules are coupled to the substrate via their 5′ ends, and can be performed even after the actual hybridization experiment so that valuable arrays do not have to be used just for QC. -Page 53
