Serial microanalysis of renal transcriptomes

Abstract

Large-scale gene expression studies can now be routinely performed on macroamounts of cells, but it is unclear to which extent current methods are valuable for analyzing complex tissues. In the present study, we used the method of serial analysis of gene expression (SAGE) for quantitative mRNA profiling in the mouse kidney. We first performed SAGE at the whole-kidney level by sequencing 12,000 mRNA tags. Most abundant tags corresponded to transcripts widely distributed or enriched in the predominant kidney epithelial cells (proximal tubular cells), whereas transcripts specific for minor cell types were barely evidenced. To better explore such cells, we set up a SAGE adaptation for downsized extracts, enabling a 1, 000-fold reduction of the amount of starting material. The potential of this approach was evaluated by studying gene expression in microdissected kidney tubules (50,000 cells). Specific gene expression profiles were obtained, and known markers (e.g., uromodulin in the thick ascending limb of Henle's loop and aquaporin-2 in the collecting duct) were found appropriately enriched. In addition, several enriched tags had no databank match, suggesting that they correspond to unknown or poorly characterized transcripts with specific tissue distribution. It is concluded that SAGE adaptation for downsized extracts makes possible large-scale quantitative gene expression measurements in small biological samples and will help to study the tissue expression and function of genes not evidenced with other high-throughput methods.

Figure 1

Overview of differences between SAGE and SADE. (Left) Typical yields of different steps used for construction of a SAGE library. Total RNAs are extracted by using the acid guanidinium thiocyanate-phenol-chloroform method (8), and poly(A) RNAs are isolated by oligo(dT)-cellulose chromatography. Synthesis of cDNA is initiated with a biotinylated oligo(dT) primer. The cDNA then is cleaved with the SAGE anchoring enzyme (Sau3AI), and its 3′-end is isolated by binding to streptavidin beads. After recovery and ligation of cDNA tags to form ditags, PCR generates the expected fragments (110 bp), together with shorter parasitic products. (Right) mRNAs are directly isolated from the tissue lysate by binding to oligo(dT) covalently bound to magnetic beads. Synthesis and cleavage of cDNA then are performed on beads. The high yield of the procedure makes it possible to generate sufficient ditags for predominant amplification of the expected product (see text for details).

Figure 2

Comparative analysis of gene expression by SAGE and RT-PCR in the kidney (K), MTAL (M), and OMCD (O). RT-PCR (25 PCR cycles) was performed in the presence of [α-³²P]dCTP by using RNA amounts corresponding to 10² cells. The DNA fragments were electrophoresed through a 2% agarose gel and detected by autoradiography. The expected size of the predominant PCR product was as follows: 60S ribosomal protein (60S RP), 392 bp; type II Na-Pi transporter (Na-Pi 2), 272 bp; Na,K-ATPase γ-subunit (Na-K γ), 413 bp; ClC-K1, 330 bp; type II 11β-hydroxysteroid dehydrogenase (11β-HSD 2), 461 bp. Abundance of SAGE tags (normalized to 7,500 total tags) in each library is shown above the gel. The RT-PCR experiment is representative of three that gave similar results.