The digestion of annealed homoduplex DNA alone with Surveyor nuclease produced no discrete products but did generate small amounts of nonspecific breakdown products shorter than full-length substrate (Figure 1, homoduplex lane). This material originates from cleavage at mismatches introduced by PCR errors and/or from nonspecific DNA endonuclease activity associated with Surveyor nuclease preparations (J.M. D'Alessio, unpublished observation). Reaction conditions used to generate Surveyor nuclease cleavage products were optimized for each separation platform to maximize signal and reduce this background as much as possible.Mapping Multiple Mutations in Large DNA Fragments with Surveyor Nuclease
Mutation detection by SSCP or DHPLC is generally limited to smaller DNA fragments (<1000 bp). It is also difficult to identify the number and location of multiple mutations in a fragment. Using Surveyor nuclease, the upper size range of mutation detection is extended considerably, and multiple mutations can be mapped. Figure 2 shows the Surveyor nuclease digestion products of a 2.95-kb fragment containing two single-base mismatches 1.4 and 1.8 kb from the 5′ end. Cleavage products 1.8, 1.55, 1.4, 1.15, and 0.4 kb long were easily resolved in this gel system. Single molecules cut only once by Surveyor nuclease at each of the mismatch sites will generate pairs of cleavage products, the sum of whose lengths total 2.95 kb (i.e., 1.4 + 1.55 kb and 1.8 + 1.15 kb) (Figure 2, left panel). Single molecules digested to completion (two cuts per molecule) will produce limit digest cleavage products 1.4, 1.15, and 0.4 kb long. Studies with a number of different DNA fragments containing two or three single-base mismatch sites show that cut substrate molecules are generally cleaved only once under the reaction conditions used with unpurified PCR fragments (data not shown). Efficient cutting at both mismatch sites of a single 2.95-kb fragment to yield substantial 0.4-kb product required post-PCR DNA cleanup before Surveyor nuclease digestion. In addition to DNA cleanup, increased incubation time and amounts of Surveyor nuclease enhance the complete digestion of heteroduplex DNA with multiple mismatches. Background from nonspecific DNA cleavage by Surveyor nuclease also increases with such changes.
Both the mismatch sites in the 2.95-kb fragment in Figure 2 produce GG and CC mismatches when mutant and reference DNA are annealed. The sequence context of each mismatch is different. After taking into consideration differences in length, there was greater than 2-fold more 1.4-kb product than 1.15-kb product at 20 min, which is consistent with the site 1.4 kb from the 5′ end being preferred by Surveyor nuclease over the site at 1.8 kb. This result suggests sequence context influences the Surveyor nuclease digestion rate.
Surveyor nuclease can be used to map the general location(s) of a mutation(s) based on the lengths of the digestion products. In an unlabeled DNA fragment, the orientation of the position of a mutation(s) with respect to the fragment ends can be mapped by altering digestion fragment length(s) through the use of an alternate PCR primer. Mapping of the locations of multiple mutations in a DNA fragment with respect to fragment ends can also be achieved by labeling the ends of a PCR product with primers containing two different fluorescent labels.The Sensitivity of Mutation Detection with Surveyor Nuclease
The CEL nuclease family of enzymes to which Surveyor nuclease belongs has been used extensively to screen for mutations in PCR product populations derived from pooled individual genomic DNAs (6,7,8). For example, an 8-fold pooling of individual DNAs (detection of 1 in 16 alleles) has been used in screening Arabidopsis mutants (6,7). To determine the sensitivity of mismatch detection using Surveyor nuclease, substrates containing different proportions of 632 bp (Figure 3A) or 2.23 kb (Figure 3B) single mismatch hetero-duplex and homoduplex DNA were digested using Surveyor nuclease and analyzed by agarose gel electrophoresis. The expected digestion products can be seen at 1 in 32 heteroduplex to homoduplex for the 632-bp PCR product, and clearly at 1 in 16 heteroduplex to homoduplex, and even at 1 in 32 for the 2.23-kb PCR product. Utilizing WAVE HSD HPLC and the ABI PRISM 3100 capillary electrophoresis with the 632-bp substrate mixture, the detection limits are 1 in 32 and 1 in 8, respectively. In other experiments performed with the WAVE HSD HPLC analysis of the products, similar levels of heteroduplex detection sensitivity were observed with PCR-amplified genomic DNA containing GA and TC, AA and TT, TG and CA, or GG and CC mismatched pairs all present in different sequence contexts (data not shown).
In conclusion, mutation detection with Surveyor nuclease provides a simple and versatile mutation detection method, based on a new class of mismatch recognition nuclease. We have shown its broad substrate specificity and high sensitivity with model substrates of different lengths.