People
 Platform

 Back to Hessner Lab

Today's date is:

Hessner Laboratory Project

Use of Fluoroscein-Labeled Probes as a Quality Control Tool for cDNA Microarrays

M.J. Hessner, K. Dowling, O. Kokanovic, L. Meyer, S.H. Nye, X. Wang, J. Waukau, S. Ghosh. The Max McGee National Research Center for Juvenile Diabetes, Department of Pediatrics, The Medical College of Wisconsin, Milwaukee, WI; Children’s Hospital of Wisconsin, Milwaukee, WI

ABSTRACT: DNA microarrays enable expression analysis of thousands of genes simultaneously making them a tool of growing popularity in the genetics research community for identification of candidate genes involved in human disease. Generation of quality arrays depends on a number of steps that are difficult to measure, especially when dealing with potentially thousands of cDNA amplicons. Recent reports have identified probe quality and quantity as critical control points. We have developed a means of quality control for probe amplification, array/element morphology and post-process probe retention, and potentially normalization of inter-element/slide variation by generation of probes with fluoroscein-labeled primers. This approach potentially eliminates the need to evaluate probe quality by gel electrophoresis as well as the need to visualize arrays with a DNA stain to evaluate probe retention and element morphology. Visualization of the entire array at any point from printing through hybridization is possible with no spectral interference when evaluating Cy3 or Cy5-labeled hybrids. The use of fluorosceinated probes for microarrays necessitates purification of amplified PCR products prior to printing in order to remove unincorporated dye-labeled oligonucleotide primers. We have evaluated high-throughput methods for probe purification and have found Promega’s MagniSilTM system an effective, high-recovery, automatable alternative to column-based purification products. By titrating purified labeled probes, we have demonstrated that a significant percentage (>75%) of probe printed in 50% water/50% DMSO is not retained by poly-L-lysine coated glass slides. Smaller array elements and greater probe retention is observed when printing is performed in 3X SSC/1.5 M trimethylglycine as reported by Diehl et al. [NAR, 29(7):e28]. We conclude that the use of labeled probes is affordable and greatly facilitates methods optimization and array quality control.

INTRODUCTION: DNA microarrays allow for mRNA expression to be evaluated for hundreds to thousands of genes in a single experiment and are a potentially powerful tool for studying disease pathogenesis. The construction and use of cDNA microarrays is an involved process comprised of multiple critical steps. First, there is construction of the array. Typically, hundreds to thousands of cDNA probes are amplified by PCR from bacterial culture. The products are purified and agarose gel electrophoresis, visualization, and manual scoring are used to laboriously assess PCR efficiency. The probes are normalized to the desired DNA concentration and printed onto prepared coated glass slides. The negatively charged DNA is electrostatically bound to the positively charged coating on the slide (typically poly-L-lysine). Slides are then fixed and blocked, and are finally competitively hybridized with two dye-labeled (typically, Cy3 and Cy5) cDNA targets derived from the two biological samples that are being compared for differential gene expression. After hybridization, the array is analyzed in a two-channel fluorescence scanner (Cy3 excitation 543 nm/emission 570 nm; Cy5 excitation 633 nm/emission 670 nm) and the relative amounts of an mRNA species in the original two samples is defined by the ratio between the two fluorophores at the homologous array element. However, full utilization this technology is currently limited by a number of problems including: standardization of intra-slide signal variation, poor correlation between microarray generated differential expression measurements versus other methods, and differences in ratio measurements and variance between replicate slides.
 
The generation of quality arrays, and therefore reliable gene expression data, depends on a number of steps that are difficult to measure, especially when dealing with potentially thousands of unique cDNA probes. It has been assumed that due to the competitive nature of two channel fluorescent hybridizations, the amount of probe DNA bound to the surface of the glass slide would have little effect on differential gene expression data. Yue et al., (2001) have shown that when an inadequate amount of DNA probe is bound to the slide an underestimation or complete failure to detect differential gene expression will occur. This highlights the importance of being able to track amounts of DNA printed and retained on the slide surface during array construction. It is our hypothesis that the quality of expression data generated can be improved by the ability to directly visualize the entire pre-hybridized array in order to qualitatively and quantitatively assess each array element prior to hybridization.

1. USE FLUOROSCEIN LABELED PROBES TO OPTIMIZE PRINTING AND POST PROCESS
METHODS

A. THROUGH DIRECT EVALUATION OF ARRAY MORPHOLOGY
B. THROUGH ABILITY TO QUANTIFY PROBE RETENTION ON SLIDE

2. EVALUATE UTILITY OF MEASURING PREHYBRIDIZATION ARRAY QUALITY

METHODS: Sequence-verified human cDNA clone inserts (Research Genetics, Huntsville Alabama) were amplified directly from 0.5ul bacterial culture by PCR using vector-specific forward (5’fluor-CTG CAA GGC GAT-fluor TAA GTT GGG TAA C-3’) and reverse (5’ fluor-GTG AGC GGA T-fluorAA CAA TTT CAC ACA GGA AAC AGC-3’) oligonucleotide primers at a concentration of 0.26uM each in a PCR master mix consisting of 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 3.0 mM MgCl2, 0.2 mM dNTPs, 1.0 M betaine, 0.25U Taq in a total reaction volume of 20 µl. Products were purified to remove unincorporated primer and PCR reaction components using either the Wizard® MagnesilTM PCR Clean-Up System (Promega, Madison, WI) or Multiscreen®-PCR Plates (Millipore, Bedford, MA), quantified using the picogreen assay (Molecular Probes, Eugene Oregon), dried down, and resuspended at the desired concentration in printing solution. Slides were printed with a GeneMachines Omni Grid printer (San Carlos, CA) using SMP3 split pins (Telechem International, Sunnyvale, CA) onto poly-L-lysine coated slides. The methods for slide preparation, post-print fixing/blocking, mRNA samples, labeling and hybridization were conducted as previously described (Schena, 1996; Stanford University http://cmgm.stanford.edu/pbrown). After hybridization, the slides were scanned with a GSI Lumonics ScanArray 5000 (Billerica, MA). An alternative fix block procedure was evaluated and adopted (Diehl et al., 2001). Imaging was accomplished with the Gene Pix Pro 3.0 software package (Axon Instruments, Foster City, CA). Analysis was performed with the Matarray software package (Wang et al., 2001).

Conclusions:
1. Generation of fluoroscein labeled probes enables direct visualization of array immediately after printing and again after blocking, allowing for quantification of bound probe.
2. Methods optimization is simplified since printing and blocking are isolated from hybridization. No spectral interference is observed when analyzing Cy3- and Cy5-labeled hybrids.
3 The use of labeled probes has made possible the rapid evaluation of printing buffers for optimal element morphology (size and shape) and post-process probe retention on the slide.
4. Method has enabled observation of sequence-dependent probe binding efficiency. Among buffers evaluated, this effect is eliminated only when probes are printed in solutions possessing betaine and DMSO.
5. Post-blocking image quality is predictive of hybridization performance.

Next Steps:
1. Establish minimum bound probe threshold values (ie what pg amount of probe needs to be bound, in general, for detection of differential expression?)
2. Use minimum bound probe threshold values as a means data filtering, and evaluate fluoroscein signal as a tool for interslide normalization.
3 Establish an overall prehybridization slide quality score (based on probe retention, background, etc) so that poor slides can be avoided in key experiments, therefore improving reproducibility, potentially reducing number of replicates necessary, and ultimately reducing costs.
4. Evaluate the use of post-print image of array to directly assess initial probe amplification efficiency to reduce time consuming quality control of probes through gel electrophoresis.
5. Integration of pre- and post-hybridized image quality filters into final data analysis.

References:
Diehl F, Grahlmann S, Beier M, Hoheisel JD. 2001. Nucleic Acids Research 29:e38.
Schena M, Shalon D, Heller R, Chai A, Brown PO, Davis RW. 1996. Proc. Natl. Acad. Sci. USA 93, 10614-10619.
 Wang X, Ghosh S, Guo S-W. 2001. Nucleic Acids Research. 29:e75
 Yue H, Eastman PS, Wang BB, Minor J, et al., 2001. Nucleic Acids Research 29: e41