Improved NlaIII digestion of PAGE-purified 102 bp ditags by addition of a single purification step in both the SAGE and microSAGE protocols

Abstract

Despite the success of microarray technologies, serial analysis of gene expression (SAGE) still remains the only technique that allows an accurate quantitative and qualitative analysis of cell transcription in a variety of physiological and pathological conditions. Nevertheless, the efficiency of SAGE is limited by the numerous gel purification steps required and these increase the possibility of contamination and reduce or inhibit the activity of the enzymes used in the protocol. In order to eliminate this problem, we have modified the original protocol by adding a single purification step before NlaIII digestion of the ditags. This allows us to increase the yield of digested ditags without reducing the amount of DNA or affecting the subsequent concatemerization.

Serial analysis of gene expression (SAGE) is a widely used technology for determining gene expression profiles of tissues and cell lines under different extracellular stimuli or under different phenotypic changes derived from genetic mutations (1,2). This method relies on isolation of polyadenylated mRNAs with subsequent synthesis of biotinylated cDNAs. The biotinylated cDNAs are bound to streptavidin beads and then digested with a four-base cutter anchoring enzyme at the 3′-ends. After ligation with two different sets of linkers, namely linkers 1 A,B and 2 A,B, to the 3′ streptavidin-bound cDNAs, BsmFI, a type IIs restriction enzyme, is used to digest 19 bp downstream of the 3′-terminus of its recognition site encoded in the linker sets. This produces soluble linker sets (1 and 2), each with a 14 bp cDNA tag. The linker sets are blunted at their 3′-ends and are then ligated to form 102 bp ditags. The ditags are amplified by PCR, run on a 12% polyacrylamide gel, purified and further digested with the same anchoring enzyme.

Subsequently, the resulting 26mer ditags are subjected to 12% PAGE, purified and ligated to form concatemers, which are cloned into a pZero vector and transformed into DH10B bacteria. Sequencing the ditags within the concatemers enables cataloging and quantification of the frequency of the tags, which delineates identification and abundance of the transcripts for a given cell line or tissue.

For our work we used SAGE protocol version 1.0c (3) and the modified protocol of microSAGE (4) to construct libraries from NGF-treated and untreated pheochromocytoma clonal cells (PC12 cells). Although we have successfully constructed libraries following these protocols, nonetheless we have encountered intermittent failure of the anchoring enzyme NlaIII to adequately digest the polyacrylamide gel-purified 102 bp ditags as described in step 10b of the original protocol. Because of this, concatemerization of the 26mer ditags was not achievable due to their limiting amounts. Our failure to digest the 102 bp ditags was not caused by the inherent sensitivity of NlaIII to storage conditions, since different lots of NlaIII were unable to digest the gel-purified ditags on different occasions.

We reasoned that the polyacrylamide gel steps required for purification of 102 bp ditags can add soluble contaminants that inhibit NlaIII activity. To solve this problem, we developed two different single-step purifications for the SAGE and microSAGE protocols aimed at removing the contaminants prior to digestion of the 102 bp ditags.

The first procedure binds the 102 bp ditags to a fused silica membrane under high ionic strength conditions with subsequent washing. The ditags are then eluted with low ionic strength LoTE buffer. The second protocol removes the contaminants from the 102 bp ditags by centrifugation gel filtration.

For our first modification procedure we used the Qiaquick kit provided by Qiagen. Although this kit is designed for agarose gels (5), it has the advantage of purifying DNA ranging in size from 100 bp to 10 kb. We proceeded as follows. Large-scale PCR was carried out to produce 100 µg of 102 bp ditags. The ditags were separated from the 80 bp linker and primer bands and cut out of the gels. Extraction was performed by heating the gel slices to 65°C and spinning out the ditags through a SpinX column (Costar) by centrifugation. The ditags were precipitated as described in step 10b of the SAGE protocol and redissolved in 80 µl of LoTE buffer (3 mM Tris–HCl, 0.2 mM EDTA). Thereafter, 600 µl of high ionic strength QG buffer and 80 µl of isopropanol (to enhance binding of small fragments of DNA) were added. After mixing, the QG buffer solution containing the ditags was spun through a Qiaquick silica membrane column. The silica column with the bound ditags was washed with 700 µl of PE buffer containing ethanol and spun again without buffer to eliminate traces of ethanol. The Qiaquick-purified ditags were removed from the column with two succeeding spins with 40 µl of LoTE, the two 40 µl elutants were pooled and the ditags were then digested with NlaIII in 80 µl as described in the SAGE protocol. Figure 1A shows an example of a polyacrylamide gel with 102 bp ditags derived from a large-scale preparation, during construction of the untreated PC12 SAGE library, that were unable to be digested by NlaIII. The 102 bp ditags were extracted from the gel, purified through Qiaquick silica membrane columns and redigested with NlaIII in order to rescue the 26 bp ditags (Fig. 1B).

Figure 1.

Polyacrylamide gels (12% PAGE in A–C, 8% PAGE in D), stained with SYBR Green I, that show the SAGE steps towards concatemerization. Using the original protocol leads to unsuccessful NlaIII ditag digestion in several cases, as depicted in (A) for naïve PC12 cells. Purification of ditags through a silica membrane rescues 102 bp ditags from the gel and increases the yield of 26 bp ditags, as shown in (B) for naïve PC12 cells and (C) for NGF-primed PC12 cells. This additional step combined with the heating step (6) allows formation of concatemers, as shown in (D) for NGF-treated PC12 cells. M1, 20 bp DNA ladder; M2, 100 bp DNA ladder.

Ditags from NGF-primed PC12 cells underwent the same treatment with silica membrame columns, which led to complete recovery of the 102 bp ditags and efficient NlaIII digestion.

Utilization of the Qiaquick kit does not adversely affect the ability of the 26 bp ditags of the NGF-treated and untreated libraries to concatemerize, as in Figure 1D, which shows concatemers resulting from ligation of previously NlaIII-digested ditags (Fig. 1C). Moreover, the yield of large-scale prepared ditags resulting from purification with the Qiaquick column was unaffected, as determined by the ethidium bromide–DNA dot assay described in the SAGE protocol.

The second modification protocol is based on the use of Clontech spe10 spin gel filtration columns. These columns have a gel pore size that excludes 100 bp fragments of DNA but retains small molecules such as nucleotides and other contaminants. A large-scale preparation of polyacrylamide gel-purified ditags was ethanol precipitated, redissolved in 80 µl of LoTE buffer and divided into two 40 µl aliquots. Each aliquot was applied to a spin spe10 column pre-equilibrated with LoTE by three successive spins with 100 µl of LoTE buffer at 3000 g for 5 min in an Eppendorf centrifuge. The 102 bp ditags were spun out of the columns by a single centrifugation as described above. Figure 2A shows the indigestible ditags obtained from a second large-scale PCR amplification in the microSAGE protocol with untreated PC12 cells. Therefore, after cutting out and re-extracting the ditags and then passing the ditags through the spe10 centrifugation column, the 102 bp ditags were able to be digested by the same lot of NlaIII, resulting in formation of the 26 bp ditags (Fig. 2B). Like the former protocol, there was no detectable loss of material and concatemerization was successful (Fig. 2C).

Figure 2.

SYBR Green I stained polyacrylamide gels (12% PAGE in A and B, 8% PAGE in C) showing the efficacy of the purification step in the microSAGE procedure for untreated PC12 cells. Prior to the use of Clontech spe10 spin gel filtration columns, 26 bp ditags are not detectable (A). As shown above in Figure 1 for SAGE, DNA purification through the columns remarkably improves NlaIII digestion (B) and facilitates formation of concatemers (C). M1, 20 bp DNA ladder; M2, 100 bp DNA ladder; M3, 1 kb DNA ladder.

In summary, we have developed two modifications, silica membrane and centrifugation gel filter purification, to be used for step 10b in version 1.0c of the SAGE protocol, which remove the contaminants found in the polyacrylamide gels and consequently permit reliable, extensive digestion of the 102 bp ditags by NlaIII. Significantly, both these methods resulted in a minimum reduction in yield of the ditags and reliably increased the yield of 26 bp ditags for both the SAGE and microSAGE protocols, as shown in Table 1. The time increase in step 10b to carry out either method is ~15–20 min. Integrating our modifications with the heating modification step previously described (6) will give robust yields of 26mer ditags for concatemers that contain up to 67 tags per clone. Although our purification methods were designed to be used in SAGE analysis, either can be used to enable complete digestion of DNA purified by PAGE with other restriction endonucleases.

Table 1. Summary of the attempts at digesting 102 bp ditags with NlaIII restriction enzyme.

	102 bp ditag NlaIII digestion
	No purification step			Purification step
	–	+	++	–	+	++
SAGE	1	2				4
MicroSAGE	3					5

The digestion reactions are grouped according to the use of a previous purification step (performed using either a Qiaquick Kit or Clontech spe10 spin columns): –, no digestion; +, 25–50% of 102 bp ditags were cut; ++, 80–90% of 102 bp ditags digested.