1. We have recently used whole-genome yeast intergenic microarrays in PBM experiments to identify the DNA binding site specificities of yeast transcription factors (2). Microarrays spotted with coding regions are also expected to aid in identiffy-ing the sequence-specific binding properties of DNA binding proteins, even though it is currently thought that most in vivo functional regulatory sites will be located in noncoding regions. Since PBM experiments are an in vitro technology, as long as there is sufficient sequence space represented on the DNA microarrays, one can expect to be able to derive a good approximation of the DNA binding site motif. Along these lines, one need not utilize microarrays spotted with amplicons representing genomic regions from the same genome as the DNA binding protein of interest; one can rather use microarrays spotted with a different genome's sequence. Similarly, microarrays spotted with synthetic dsDNAs can also be used in PBMs. Likewise, the dsDNAs need not be made by PCR amplification, but rather can be made by other means, such as primer extension. We have successfully performed PBMs using microarrays spotted with PCR products whose lengths ranged from approx 60 to approx 1500 bp (2) and also using microarrays spotted with synthetic dsDNAs ranging from approx 35 to approx 50 bp (2,6).

2. Alternatively, PCR products may be precipitated with 1 M ammonium acetate and 2 vol of isopropanol, washed with 70% ethanol, dried overnight, and resuspended in 3X SSC printing buffer. The extra filtration provided by the MultiScreen® plates increases the purity of the double-stranded DNA.

3. Any remaining unused blank GAPS II or UltraGAPS slides from an opened package should be stored in a vacuum desiccator containing Drierite desiccant, as should all postprocessed, printed microarrays.

4. Baking the microarrays at 80°C for 2 h in a clean oven before UV-crosslinking may improve intraspot uniformity. However, baking may also result in a decreased shelf life of the microarrays.

5. Unless otherwise noted, buffers are prepared using distilled, deionized water (ddH2O).

6. Purified DNA binding proteins should be aliquoted before storing at -80°C, to avoid unnecessary freeze/thaw. Antibody should be stored according to the manufacturer's recommendations. For long-term storage of Alexa Fluor® 488-conjugated anti-GST polyclonal antibody (Molecular Probes), we recommend aliquoting and storing at -20°C, per the manufacturer's recommendations. However, we have observed no noticeable decrease in signal intensity after storing this antibody at 4°C for 1 yr.

7. ZnAc is necessary only when performing PBM experiments on zinc finger or other zinc-coordinating proteins.

8. Filtering of the 2% and 4% milk solutions serves two purposes: (1) sterilization; and (2) removal of fine particulates that may contribute to noise in the PBM data. We have found that the 4% milk solution readily clogs a 0.22-pm filter unit, so we recommend using a 0.45-pm filter unit for syringe-filtering of the 4% milk solution. The 2% milk solution can be syringe-filtered with a 0.22-pm filter. Alternatively, 0.45-pm filters can be used for sterilizing both the 2% and 4% milk.

9. Microarrays are incubated in a hydration chamber to prevent excessive evaporation of the reaction mixture under the cover slip. An empty pipet tip box works nicely. Lift out the tip rack, fill the bottom of the pipet tip box with about half an inch of sterile water, and replace the tip rack. Wipe off the inside of the lid and the tip rack with ethanol using a Kimwipe before every use and between reaction steps in the PBM experiments.

10. Note that SYBR Green I and Alexa Fluor® 488-conjugated anti-GST polyclonal antibodies are light sensitive, and so measures should be taken to avoid their photo-bleaching in the course of the SYBR Green I staining and PBM experiments. We recommend turning off all overhead and benchtop lighting when handling these reagents and also when handling the microarrays once they have been stained with either SYBR Green I or the fluorophore-conjugated antibody.

11. All wash steps in Coplin jars are performed by shaking at room temperature at approx 125 rpm on a platform shaker. Shaking at speeds faster than approx 125 rpm may cause the Coplin jar to tip over while shaking.

12. Drying the back and sides of the microarray with a Kimwipe helps to prevent solution from leaking out from the edges of the cover slip. Because of the hydro-phobicity of the active surface of the Corning® GAPS II and UltraGAPS glass slides, drying the front of the glass slide outside the perimeter of the DNA spots, using the etched grooves as a guide, can help to confine the protein or antibody solution, once dispensed onto the microarray, to the area of the microarray that contains the DNA spots. This must be done quickly so that the area containing the DNA spots does not dry.

13. Briefly centrifuge all reaction mixtures before applying to microarrays, to remove bubbles. When pipeting the reaction mixtures onto the microarrays, avoid pipeting any fine bubbles that may remain at the very top surface of the reaction mixtures. If nevertheless a few air bubbles become apparent once a reaction mixture has been applied to a microarray, very carefully attempt to remove the bubbles by pipetting, while avoiding removal of the reaction mixture. If some air bubbles still remain, they may be brought to the edge of the glass slide, and thus outside the spotted area, by gently rocking the cover slip as it is laid down on the microarray.

14. In applying reaction mixtures onto the microarrays, certain techniques can aid in spreading the mixture over the surface of the microarray and in increasing the homogeneity of the reaction mixture when it is applied to a prewet or washed microarray. The reaction mixture can be dispensed one droplet at a time, covering the entire surface where the DNA was spotted. The microarray can also be rocked back and forth to spread the reaction mixture uniformly across the spotted area. The use of LifterSlips™ cover slips helps to ensure a uniform distribution of the reaction mixture over the surface of the microarray.

15. For our ScanArray 5000 microarray scanner, we have found the PMT gain to be optimal between 70 and 80%. We typically fix the PMT gain setting and vary the laser power in increments of 10 to 15% (in terms of total laser power) such that there are no spots with saturated signal intensities in the lowest intensity scan.

16. It is important to ensure that each spot has enough DNA present to allow accurate quantification of its signal intensity, which is consequently used to estimate the degree of sequence-specific binding of a given DNA binding protein to that spot. If the DNA concentration at a particular spot is too low or if the DNA is spread non-uniformly throughout the pixels of a particular spot, accurate measurements are more difficult. For this reason, it is important to remove such error-prone spots from consideration. Since some spots may be noisy (i.e., spots with highly variable pixel signal intensities) even after the use of this filter, we also remove noisy spots from consideration.

17. We found empirically that the following three additional filtering criteria helped to eliminate 'false-positive' calls (i.e., spots with no identifiable binding sites being erroneously identified as bound): (1) DNA length greater than 1500 bp; (2) low SYBR Green I raw signal intensity; and (3) low DNA density (SYBR Green I/length). These three additional filters together removed 2.7% of spots from consideration in our PBM experiments using yeast whole-genome intergenic microarrays (2). Here we are not providing the actual values for the second and third filtering criteria because these values will vary somewhat among individual microarray scanners. We recommend that the user consider all three of these criteria as suggested guidelines to employ and adjust as may be appropriate.

18. This function is related to the probability of observing a data point greater than z standard deviations above the mean of a normal distribution. Strictly speaking, we are not calculating a true z-score, since here we do not calculate the p value relative to all the data, but rather just to the reflected left half of the distribution.

19. Other motif finders, such as AlignACE (5,13), MEME (14), and MDscan (15), can also be used to identify the DNA binding site motif. We chose BioProspector over other available motif finding programs because it proved to be the most inclusive in accepting the largest number of input sequences in construction of yeast transcription factor binding site motifs (2).

0 0

Post a comment